AGILISYS LIMITED AGAINST CGI IT UK LIMITED AND CGI IT UK LIMITED AGAINST AGILISYS LIMITED [2018] ScotCS CSOH_112 (04 December 2018)

BAILII [Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]

Scottish Court of Session Decisions


You are here: BAILII >> Databases >> Scottish Court of Session Decisions >> AGILISYS LIMITED AGAINST CGI IT UK LIMITED AND CGI IT UK LIMITED AGAINST AGILISYS LIMITED [2018] ScotCS CSOH_112 (04 December 2018)
URL: http://www.bailii.org/scot/cases/ScotCS/2018/[2018]_CSOH_112.html
Cite as: 2019 GWD 3-40, [2018] CSOH 112, [2018] ScotCS CSOH_112

[New search] [View without highlighting] [Printable PDF version] [Help]


Page 1 ⇓
OUTER HOUSE, COURT OF SESSION
[2018] CSOH 112
CA54/17 & CA55/17
OPINION OF LORD BANNATYNE
in the cause
AGILISYS LIMITED
against
CGI IT UK LIMITED
and
Pursuer
Defender
CGI IT UK LIMITED
Pursuer
against
AGILISYS LIMITED
Defender
Pursuer: Cormack QC, J Young; Pinsent Masons LLP
Defender: Sandison QC, G Reid; Brodies LLP
4 December 2018
Introduction
[1]       This matter came before me for proof before answer in the commercial court. The
general background to the dispute is a contract between the parties (“the Subcontract”). The
subject matter of the proof is issues arising from the service by Agilisys on CGI of Relief
Page 2 ⇓
2
Notices in terms of the Subcontract. The scope of the proof is as set out in a joint list of
questions submitted by parties for determination by the court.
The questions to be answered by the court
Relief Notice 002 3rd Party APIs and Supplier Portals
1. What were the respective obligations of the parties (including as to the Milestone
Dates in the Implementation Plans and timing generally) in relation to the
procurement, installation and provision of access to a development, test and
production instance of the Northgate API and the test data?
2. Did CGI breach its obligations by failing to provide access to a development, test and
production instance of the Northgate API and test data?
3. If the answer to question 2 is yes, was CGI’s breach caused or affected by a breach by
Agilisys of its obligations?
4. If the answer to question 2 is yes, did that breach cause, or was it reasonably likely to
cause, Agilisys to fail to achieve the Drop 1 Milestone MS-OSC 1.1 in the I
Implementation Plan by the effective date?
Relief Notices 004 & 006 Biztalk
5. What were the respective obligations of the parties (including as to the Milestone
Dates in the Implementation Plans and timing generally) in relation to the provision
of experienced Biztalk development resources and/or access to the Biztalk
development environment?
6. Did CGI breach its obligations by failing to provide access to Biztalk development
resources and/or access to the Biztalk development environment?
Page 3 ⇓
3
7. If the answer to question 6 is yes, was CGI’s breach or breaches caused or affected by
a breach or breaches by Agilisys of its obligations?
8. If the answer to question 6 is yes, did any breaches cause, or were they reasonably
likely to cause, Agilisys to fail to achieve the Milestone MS-2.5 in the I
Implementation Plan by the effective date?
Relief Notices 005 and 011 Payroll Information
9. What were the respective obligations of the parties (including as to the Milestone
Dates in the Implementation Plans and timing generally) in relation to procuring that
CGI fulfilled its role under the Subcontract of providing business information,
knowledge and decisions to enable the I payroll solution to be configured?
10. Did CGI breach its obligations by failing to procure that CEC provided business
information, knowledge and decisions to enable the I payroll solution to be
configured?
11. If the answer to question 10 is yes, was CGI’s breach caused or affected by a breach
by Agilisys of its obligations?
12. If the answer to question 10 is yes, did that breach cause, or was it reasonably likely
to cause, Agilisys to fail to achieve the Milestone MS-2.2 in the I Implementation Plan
by the effective date?
Relief Notices 007 and 016 Data Migration
13. What were the respective obligations of the parties (including as to the Milestone
Dates in the Implementation Plans and timing generally) in relation to the provision
of data extracts from I’s existing Finance, Human Resources and Payroll systems?
Page 4 ⇓
4
14. Did CGI breach its obligations by failing to provide data extracts from CEC’s existing
Finance, Human Resources and Payroll systems?
15. If the answer to question 14 is yes, was CGI’s breach caused or affected by a breach
by Agilisys of its obligations?
16. If the answer to question 14 is yes, did that breach cause, or was it reasonably likely
to cause, Agilisys to fail to achieve the Milestone MS-2.5 in the I Implementation Plan
by the effective date?
Relief Notice 008 Interface Specifications
17. What were the respective obligations of the parties (including as to the Milestone
Dates in the Implementation Plans and timing generally) in relation to the provision
of a complete set of Agresso technical interface specifications?
18. Did CGI breach its obligations by failing to provide a complete set of Agresso
technical interface specifications?
19. If the answer to question 18 is yes, was CGI’s breach caused or affected by a breach
by Agilisys of its obligations?
20. If the answer to question 18 is yes, did that breach cause, or was it reasonably likely
to cause, Agilisys to fail to achieve the Milestone MS-2.5 in the I Implementation Plan
by the effective date?
Relief Notices 009, 014 and 015 I Environments
21. What were the respective obligations of the parties (including as to the Milestone
Dates in the Implementation Plans and timing generally) in relation to the provision
Page 5 ⇓
5
of fully functional environments and related infrastructure, including suitable
environment management services, in relation to the I Project?
22. Did CGI breach its obligations by failing to provide fully functional environments
and related infrastructure, including suitable environment management services, on
the I Project?
23. If the answer to question 22 is yes, was CGI’s breach or breaches caused or affected
by a breach by Agilisys of its obligations?
24. If the answer to question 22 is yes, did any breaches cause, or were they reasonably
likely to cause, Agilisys to fail to achieve the Milestones from MS2.5 onwards in the I
Implementation Plan by the effective dates?
Relief Notice 012 I Environments
25. What were the respective obligations of the parties (including as to the Milestone
Dates in the Implementation Plans and timing generally) in relation to the provision
of access to fully functional I environments?
26. Did CGI breach its obligations by failing to provide access to fully functional I
environments?
27. If the answer to question 26 is yes, was CGI’s breach or breaches caused or affected
by a breach by Agilisys of its obligations?
28. If the answer to question 26 is yes, did any breaches cause, or were they reasonably
likely to cause, Agilisys to fail to achieve the Milestone MS-SC1.1 in the I
Implementation Plan by the effective date?
I would intend to divide my opinion into a number of sections:
Page 6 ⇓
6
I
Background
II The tactical burden of proof
III Objection
IV Preliminary legal issue: whether CGI has compromised any part of its
claim?
V Preliminary legal issue: the validity of purported changes to the
Subcontract in the absence of the use of change control
VI Approach to evidence
VII Assessment of witnesses
VIII Approach to the construction of the Subcontract
IX The roles of transformation director, application architect and solutions
architect
X The nature and scope of the management and leadership role of
Agilisys
XI The Relief Notices
I Background
[2]       Agilisys carries on business as a software and business service provider and is
engaged in business in the United Kingdom. CGI is an information technology consulting,
systems integration, outsourcing and solutions provider engaged in business in the United
Kingdom and elsewhere worldwide.
[3]       On 23 August 2015 CGI entered into a contract with the City of Edinburgh Council
(“CEC”) for the provision to CEC of ICT (ie information, communication and technology)
and Transformation Services (the “Prime Contract”). Prior to the Prime Contract, CEC’s ICT
services were provided by an external supplier, namely: British Telecom (“BT”).
Page 7 ⇓
7
[4]       In terms of the Prime Contract CGI undertook to deliver a variety of IT functions to
CEC.
[5]       On 11 September 2015 CGI and Agilisys entered into the Subcontract under which
certain transformation services were delegated to Agilisys, mainly consisting of the elements
of Enterprise Resource Planning (“ERP”) and Enterprise Integration (“EI”).
The ERP project
[6]       The ERP project correlated to Outline Business Solution (“OBS”) OBS 1, 2, 3, 30, 47,
and 51 and the Implementation Plan called “Finance and HR/Payroll Transformation” (see:
1119-1120 of JB 1).
[7]       In broad terms, they were concerned with the replacement and consolidation of three
legacy computer systems: (1) the debtors (or accounts receivable) system called Northgate
PPSL, per OBS 1, 2, 30, and 47; (2) the core financial services, reporting, and procurement
system called Oracle e-business, per OBS 2, and 47; and (3) the HR and Payroll system
called Midland iTrent, per OBS 3 and 47. These systems were to be replaced by a single ERP
system called Business World (formally known as Agresso) (the “BW solution”).
[8]       At the beginning of the Subcontract, per the agreed Implementation Plan (contained
in Schedule Part 6.1 of the Subcontract), work was to commence on 14 September 2015 and
complete work with the BW solution going live on 3 October 2016.
The EI project
[9]       The EI project correlated to OBS 13 and 21, and the Implementation Plan called
“Enterprise Integration” (see page 1124 of JB 1).
Page 8 ⇓
8
[10]       In broad terms, the EI project was about allowing citizens to access CEC’s services
through an online platform in order to perform digital transactions. The online platform
was proprietary software belonging to Agilisys known as Agilisys Digital (“AD”).
[11]       At the beginning of the Subcontract, per the agreed Implementation Plan, work was
to commence on 7 September 2015 and was not due to complete until 3 April 2018.
Milestone MS-OSC 1.1 was scheduled for 1 April 2016.
The Subcontract
Services and charges
[12]       In terms of the Subcontract the rights and responsibilities in respect to services and
charges are:
[13]       By Clause 5.1 of the Subcontract, Agilisys agreed to provide the Services to CGI. The
Services comprised the Transition Services and the Transformation Services, which were
both further defined by the Subcontract. The Transformation Services were defined as the
Services which Agilisys was obliged to provide in order to implement and complete the
Transformation Plan (hereinafter referred to for consistency as the “Implementation Plan”)
(which can be found at Schedule Part 6.1).
[14]       By Clause 1.1 and Schedule Part 1 Definitions of the Subcontract, Charges were
defined as the charges for the provision of the Services set out in, or otherwise calculated in
accordance with, Schedule Part 7.1, including any Milestone Payment or Service Charge. By
Paragraph 1 of Part B of Schedule Part 7.1 of the Subcontract, on Achievement of a Milestone
Agilisys is entitled to invoice CGI for the Milestone Payment associated with the Milestone
(Milestones are further referred to below).
Page 9 ⇓
9
[15]       By Clause 10.1 of the Subcontract, Agilisys was to invoice CGI for Services and CGI
(subject to its right reasonably to dispute the invoice) was to pay the Charges to Agilisys in
respect of the Services within 30 days of receipt of a valid invoice under Paragraph 2.1, Part
E of Schedule Part 7.1. Paragraph 1.3 of that Part provided for Agilisys to first submit a draft
invoice which was to be agreed by CGI before the final invoice was submitted.
Times and milestones
[16]       In respect to times and milestones the rights and responsibilities of parties are:
[17]       By Clause 6.7 of the Subcontract, Agilisys was obliged to comply with the
Implementation Plan and ensure that each Milestone was achieved on or before its
Milestone Date. Milestones were particular events or tasks described in the Implementation
Plan and, where applicable, the relevant Milestone Date was the date set out against that
Milestone in the Implementation Plan by which the Milestone was required to be achieved.
If a Key Milestone was not achieved by the relevant Milestone Date, then Agilisys was
potentially exposed to liability to make Delay Payments to CGI.
[18]       The Relief Notices, the subject matter of the proof, are governed by clause 31 of the
Subcontract which provides reading short, as follows: a mechanism for dealing with delay
caused by CGI under the Subcontract. In terms of that clause, Agilisys must issue a Relief
Notice notifying CGI of the CGI cause that has caused, or is reasonably likely to cause, a
Supplier Non-Performance (see: page 49-50 in JB 1). CGI must then, as soon as is
reasonably practicable, consider the Relief Notice and whether it agrees with the Agilisys
assessment.
[19]       The term “CGI cause” is defined as “any material breach by CGI of any of the CGI
Responsibilities, except to the extent that such breach is: (a) the result of any act or omission
Page 10 ⇓
10
by CGI which the Supplier (the Supplier in terms of the Subcontract is Agilisys) has given its
prior consent; or (b) caused by the Supplier, any Sub-contractor or any Supplier Personnel;
(c) the subject of Pre-Agreed Resolution” (see: page 68 of JB 1).
[20]       The division of responsibilities between the parties was agreed in three places in the
contract as follows: (a) in the column dividing responsibility in the OBS; (b) in the Rules
and Responsibilities document in section C of Schedule Part 4.1; and (c) the obligations on
CGI to procure various things from CEC inserted in Annex 2 to Schedule Part 6.1.
The Relief Notices
[21]       The Relief Notices issued by Agilisys in this case can helpfully be grouped in the
following manner:
ERP Environments: the Relief Notices in this group were: RN 001 (1 February
2016), RN 009 (18 March 2016), RN 013 (3 June 2016), RN 014 (26 July 2016) and
RN 015 (13 October 2016). RNs 001 and 013 are not in issue in the sense that they
were accepted by CGI and relief was granted. RNs 009, 012, 014 and 015 are within
the scope of the Proof.
ERP Data Migration: the Relief Notices in this group were: RN 007 (18 March
2016) and RN 016 (13 October 2016). Both of these RNs are within the scope of the
Proof.
ERP Biztalk: the Relief Notices in this group were: RN 004 (10 February 2016) and
RN 006 (18 March 2016). Both of these RNs are within the scope of the Proof.
Page 11 ⇓
11
ERP Interfaces: the Relief Notice in this group was RN 008 (18 March 2016), which
is within the scope of the Proof.
EI Environments and Interfaces: the Relief Notices in this group were RN 002
(29 January 2016), RN 003 (28 January 2016), and RN 010 (6 April 2016), and RN012
(12 May 2016). Only RN002 is within the scope of the Proof the other two were
accepted and relief was in fact granted by CGI.
Payroll information the Relief Notices in this group were: RN 005 (11 March 2016)
and RN011 (6 April 2016). Both of these RNs are within the scope of the Proof.
The dispute
[22]       In the course of its written submissions Agilisys sought to define broadly the issues
in the dispute as outlined in parties’ pleadings. I believe that it is a fair representation of the
broad outlines of the dispute between the parties and I consider it is helpful in order to set
the questions posed for determination by the court in context to set this out:
[23]       Agilisys aver that they issued a number of Relief Notices between January and
October 2016, and that CGI failed to adequately respond to them: Article 20 of the Agilisys
summons. Agilisys, generally, relies on CGI’s failure to adequately respond to the Relief
Notices (and related breaches such as failing to approve associated change requests) as an
integral part of the breaches entitling it to rescind the Subcontract in March 2017:
articles 20-25 of Condescendence.
[24]       CGI aver that the delays in the ERP and EI projects were properly attributable to
Agilisys’ failure to competently manage the ERP and EI projects: see article 9 of CGI
Page 12 ⇓
12
summons. In particular, in relation to ERP, they aver that the “operative cause of Delay to the
Transformation Plan (Implementation Plan)” was Agilisys’ failure to manage the ERP project.
Stripped to its essentials the allegations seem to be that there were failures as follows: (a) to
produce clear plans and strategies for implementation of the ERP project, to document
project meetings and management clearly, establish a clear governance structure and
communicate that to CGI, or to provide a business transformation director: article 32.1 of
condescendence; (b) to provide a business transformation director or application architect:
article 32.2 of CGI summons; (c) to provide a suitable technical design for an archive
solution: article 32.3 in CGI summons; (d) to provide a system architecture design:
article 32.4 in CGI summons; (e) to provide standard hardware (exacerbated by a refusal to
accept CEC laptops); (f) to manage Unit 4; (g) delaying in designing and building BW
solution: article 32.7 in CGI summons; (h) to design and manage the data migration
process, including provision of templates, or to provide a data migration strategy:
article 32.8.
[25]       CGI disputed Agilisys’ entitlement to rescind the Subcontract. CGI, following
receipt of the notice rescinding the contract for Agilisys, served on Agilisys a document
terminating the Subcontract in terms of clause 33.1.2 on the basis that a supplier termination
event had occurred, namely: that Agilisys had committed an irremediable material default
(article 40 CGI summons).
[26]       Accordingly in light of the above the ultimate issues in the case: were the validity of
Agilisys’ recission and the validity of CGI’s termination.
[27]       Resulting from the above, Agilisys sought payment in terms of the Subcontract and
also advanced a damages claim for its loss of opportunity (see: articles 29-37 of Agilisys’
Page 13 ⇓
13
summons). CGI sought damages for certain losses as set out in articles 43 and 44 of CGI’s
summons.
Submissions
[28]       Counsel prepared very lengthy and extremely detailed written submissions which
they supplemented by oral submissions. They were of considerable assistance to me. When
dealing with the various sections of my discussion I will outline the submissions of each
party.
Discussion
II The tactical burden of proof
[29]       It is not controversial that the tactical burden of proof is this: it is for Agilisys in the
first place to establish that a CGI cause under the contract existed. It is also accepted by
Mr Sandison that in many instances CGI do not take serious issue with that first point. In
respect to the next question: were any such breaches on the part of CGI caused or affected
by a breach by Agilisys of its obligations then the tactical burden of proof shifted to CGI.
III Objection
[30]       During the course of oral submissions, and thereafter in terms of a note, an objection
was taken by CGI to various productions referred to in the written submissions of Agilisys
as not having been spoken to by a witness (see: Appendix 4). Agilisys responded to that
objection in the form of a note (see: Appendix 5) accepting that certain productions had not
been referred to and deleting reference thereto in its submissions; detailing where certain
Page 14 ⇓
14
documents had been referred to; and altering certain references. So far as I could see these
responses answered fully the points made on behalf of CGI in the first part of its note.
[31]       In the second part of its note CGI referred to documentary productions which were
referred to by Dr Hunt in the course of her report but which had not been spoken to by a
witness to fact. Dr Hunt was the expert witness for Agilisys. It was accepted on behalf of
Agilisys that it was correct that many of the documents on CGI’s list had not been referred
to by a witness to fact. Agilisys, however, pointed out that this was equally true of the
report prepared by Mr Coyne, CGI’s expert, and it produced a list of these documents.
[32]       Agilisys argued that it could identify no issue in respect to partiesexperts referring
to such documents. It was Agilisys’ position that both experts referred to documents which
they had found as a result of their investigations. The experts were entitled to give evidence
in this respect and the references to the relevant documents by the experts were sufficient for
those documents to be in evidence.
[33]       I am satisfied that the above analysis by Mr Cormack is clearly correct, that
accordingly such documents form part of the evidence and that the experts were entitled to
make use of these documents in reaching their conclusions.
[34]       CGI’s position seemed to develop in response to Agilisys’ said position and became
this: it was for the court to decide how much weight should be attached to Dr Hunt’s
evidence which proceeds by reference to her own interpretation of documents after the fact,
rather than factual evidence of those involved at the time and thus little weight should be
placed on passages in her evidence which rely on such documents.
[35]       I agree it is for the court to assess the reasoning of experts in relation to the opinions
which they express and to consider the weight which should be placed on any conclusions
the expert reaches having regard to the reasoning including any documents to which the
Page 15 ⇓
15
expert refers. Looking to the evidence of Dr Hunt where she does refer to such
documentation I believe it is fair to characterise her position as this: she gives an accurate
summary of what is patent from the terms of the document itself and goes no further than
that. In particular I observe (as referred to by Agilisys) that I do not recall that she was
challenged in the course of cross-examination that her approach to such documents had
been anything other than as I have described above. I am not persuaded that only little
weight should be attached to her evidence which relies on such documents.
IV Preliminary legal issue: whether CGI has compromised any part of its claim?
[36]       In the course of his submissions Mr Cormack raised a short and discrete legal point
and it is I believe appropriate to deal with this submission before turning to a detailed
consideration of the evidence and other submissions.
[37]       The argument put forward by Mr Cormack in summary was this: where CGI
accepted CGI cause following the service of a Relief Notice by Agilisys and granted relief in
the form of extension of milestone dates, it is not open to it to revisit those determinations.
That he submitted is consistent with the general principle that where an extension has been
granted it operates as a complete defence to liquidated damages: see Keating on
Construction Contracts (10th ed) at para 10-038. Otherwise, it would have the absurd
commercial consequence that long after the event Agilisys would suddenly become liable
for delay payments. The effect of CGI accepting the Relief Notices that it did has the effect
that, up until around July 2016, it had effectively compromised any claims it wished to make
in relation to delay on the part of Agilisys in that period.
Page 16 ⇓
16
[38]       In support of his submissions Mr Cormack directed my attention to first De Beers UK
Ltd v ATOS Origin IT Services UK Ltd [2010] EWHC 3276 per Mr Justice Edwards-Stewart at
paragraph 98 where he observes:
“Fortunately, it is not necessary for me to consider which of these causes of delay
may have been dominant or critical because, as I will explain shortly, in early April
2008 DB agreed to a revision of the programme which put back the dates for delivery
of the software. Although there was no express agreement to this effect, I find that
this agreement effectively compromised any mutual claims for delay up to about the
middle of March 2008.”
[39]       Secondly he generally referred me to David MacBrayne Ltd v ATOS IT Services (UK)
Ltd [2018] CSOH 32, Lord Doherty at paragraphs 18 and 66.
[40]       Mr Sandison’s response to this argument was in summary: the granting of relief in
the course of the execution of the contract does not prevent CGI revisiting these Relief
Notices where it granted relief where the question arises in a different context, namely: the
entitlement of Agilisys to repudiate the contract. He accepted he was not entitled to
withdraw the fact that relief was granted. However, CGI was entitled to put forward a
position as to the whole circumstances in which the issue complained of occurred and the
consequences of its occurrence. The exercise in which parties are presently engaged is not
directed as to whether there was delay but whether what occurred justified the repudiation
of the contract by Agilisys.
[41]       I am persuaded by the argument put forward by Mr Sandison. What was
compromised by CGI accepting Relief Notices is this: it cannot seek, in respect to the subject
matter of those Relief Notices it accepted, a Delay Payment in terms of Schedule Part 7.1
Part C which sets out a liquidated damages clause which operated if a Key Milestone had
not been achieved by Agilisys on or before the relevant Milestone Date. This prevents
Agilisys, as argued by Mr Cormack, from long after the event suddenly becoming liable for
Delay Payments. I agree with Mr Cormack that that would be a commercially absurd result.
Page 17 ⇓
17
But CGI are not seeking Delay Payments, they are seeking to examine the Relief Notices in
an entirely different context as outlined by Mr Sandison. I believe the view which I have
arrived at is consistent with the passage to which I was directed in Keating. It also I think is
consistent with the observations in De Beers Ltd which clearly envisages a compromise for
“mutual claims for delay” in terms of the operative provisions of the contract. For the
foregoing reasons I reject Mr Cormack’s submissions on this issue.
IV The second preliminary legal issue
[42]       Mr Sandison advanced the following argument in his written submissions:
“36. The Subcontract contained the (Implementation) Plan in Schedule Part 6.1. This
stipulated a number of ‘milestones’ by which key events in the Subcontract
were to be achieved. As a part of the Subcontract, the (Implementation) Plan
could only be amended by utilising the Subcontract’s change control
procedure.
37. In addition to the (Implementation) Plan, a considerable number of (Project)
Plans were produced during the course of the Subcontract by Agilisys. These
plans were considerably more detailed than the (Implementation) Plan. They
contained many more dates by which items were to be provided (in order to
achieve the more general milestone dates set out in the (Implementation) Plan.
Anna Vale’s evidence was that these plans were intended to ‘sit behind’ the
contractual (Implementation) Plan.
38. Not all (Project) Plans were released to CGI (or to CEC): of more than one
hundred which were created, only six ‘key’ versions were provided, with other
versions ‘released for review’.
39. These (Project) Plans, therefore, although created by Agilisys during the
operation of the Subcontract, were not contractual in nature or effect. They
represented the plans which either the parties between themselves or (in the
case of the plans not released to CGI) Agilisys individually had arrived at in
order to fulfil the contractual dates. Moreover, the (Project) Plans could not, in
the absence of the operation of the contractual change control procedure,
constitute formal alterations to the Subcontract and, accordingly, to the
(Implementation) Plan. During the course of the Subcontract, delays occurred
with the result that many dates in the (Implementation) Plan were not
achieved. Replans of the dates took place using the (Project) Plan. These,
however, did not have contractual effect.”
Page 18 ⇓
18
[43]       In reply to this argument Mr Cormack submitted that it was going too far to say that
you must have change control every time you make a change to a plan. Mr Cormack in
elaboration of this general position advanced, first, the following argument:
[44]       There is an oddity around the fact that consequential changes required to the
Implementation Plan are, at different points in Clause 31, apparently under the control of
CGI or require to go through the Change Control Procedure: see Clause 31.1(c)(ii) and 31.7.
He submitted, however, that these provisions are, in reality, just a form of dispute resolution
that do not affect the substantive provisions. As was said in De Beers, “If it were otherwise,
[a party] could effectively stifle any claim for breach of contract by the simple expedient of
refusing to agree it within the Change Control Procedure.: Para 199. In fact, that is
precisely what CGI attempted to do, because it repeatedly granted relief and then refused to
operate the Change Control Procedure. This point has particular relevance in this case to the
claims that the Implementation Plan was never formally amended.
[45]       He submitted that it would be absurd, and entirely destructive of the practical effect
of Clause 31, if CGI could exercise a discretionary right to reject changes under the Change
Control Procedure where such a change was necessary to implement an entitlement of
Agilisys to relief under Clause 31. That cannot have been the intention of parties. Any
tension or conflict in the Subcontract, it is submitted, can be resolved under reference to the
order of precedence in Clause 1.5. As one would expect, this gives paramount precedence to
the express Clauses of the Subcontract over the Schedules. There is no reason in principle
that an amendment to the Implementation Plan could not be agreed through the Relief
Notice, and responses, procedure. In particular, the form of a Change Request in Schedule
Part 8.2 only required to be used unless both parties agreed otherwise: see paras. 2.2.3 of
Schedule Part 8.2.
Page 19 ⇓
19
[46]       In considering the competing submissions I believe the starting point must be Rock
Advertising Ltd v MWB Business Exchange Centres Ltd [2018] UKSC 24 which Mr Sandison in a
somewhat oblique fashion referred to in his oral submissions.
[47]       In the above case the Supreme Court considered the issue of the validity of an oral
modification to a contract where there was a “No oral modification clause” in the contract.
[48]       Lord Sumption gave the judgement, with which the other Justices agreed,
Lord Briggs dissenting as to reasoning but not as to the result. Lord Sumption having
considered the common law jurisprudence at paragraphs 8 and 9 opined at paragraph 10:
“In my opinion the law should and does give effect to a contractual provision
requiring specific formalities to be observed for a variation.
[49]       On a reading of paragraphs 10 to 15 of his judgement, it is clear that the principle
enunciated is applicable not only to NOM clauses but to all clauses which lay down a
specified procedure for making changes to a contract. Thus where in the present case
specific formalities for making changes to the Subcontract namely: Change Control have to
be followed a valid variation cannot be made in the absence of such formalities being
complied with.
[50]       I believe that the authority relied on by Mr Cormack is readily distinguishable. It is
evident that the clause being considered in De Beers is not one where “Change Control
operates as a specific formality required to amend the contract. Rather “Change Control” in
the context of De Beers relates to the issue of agreement of “loss or extra cost” (properly
understood a mere form of dispute resolution procedure) (see: paragraph 197). Thus the
observation at paragraph 199 relied upon by Mr Cormack provides no assistance to him.
[51]       Secondly, if CGI do not operate change control in an appropriate manner then
Agilisys would not be prevented from proceeding to litigation. It could proceed to litigation
on the basis of breach of contract by CGI. I do not think that the practical effect of clause 31
Page 20 ⇓
20
is entirely destroyed by concluding that for an amendment to the Implementation Plan to be
validly achieved Change Control must be used. Agilisys still has a remedy.
[52]       For the foregoing reasons I reject Mr Cormack’s first argument.
[53]       The only exception to the application of the principle enunciated by Lord Sumption
would be in circumstances where the concept of waiver is applicable. No case of waiver was
plead in this case or argued on behalf of Agilisys. As I have set out Agilisys’ argument in
respect to this point was one based on the proper construction of the Subcontract.
[54]       Mr Cormack did however advance a secondary argument in relation to this issue.
He submitted that the various dates in Project Plans which had not gone through change
control could have relevance. The argument was advanced most clearly at paragraphs 4.4.1
and 4.4.2 of his written submissions and was as follows: in considering whether a
dependency had been delivered within the time required by the Implementation Plan,
where no express time limit was given regard should be had to the various dates in the
Implementation Plan together with dates in the various Project Plans. Whilst they are not
set in stone and whilst the court must ultimately bear in mind the contractor will have built
some contingency into plans, a Project Plan such as those prepared and agreed by both
parties in this case can provide the court with a useful guide or starting point as to what is
reasonable in this context: see Hudson’s Building and Engineering Contract (13th edition) at
Para 6-054, if only because it was plainly judged as reasonable by the party at the time.
Indeed, as Anna Vale agreed in the course of her cross-examination, once a party has agreed
to plans they are expected to comply with them: see day 10, page 43, of the transcript.
[55]       It seems to me that the above argument put forward by Mr Cormack is correct. I
believe the court is entitled to have regard to such dates for the purpose contended by
Mr Cormack. I believe that such evidence can give guidance as to what is reasonable in this
Page 21 ⇓
21
context for the reasons advanced by Mr Cormack. By using the dates in this way the
Milestone Dates or any other term in the Subcontract are not altered and accordingly it does
not breach the principle laid down by the Supreme Court.
[56]       Accordingly, for the above reasons when considering the various Relief Notices
which are the subject of the proof, the Milestone Dates to which I have had regard are those
set out in the Implementation Plan in Schedule Part 6.1. Other dates in the Project Plans I
have had regard to only to the limited extent above set out.
V Approach to evidence
[57]       The issue of the use and value of contemporaneous documentation in assessing the
evidence was, in light of the way that the evidence developed and the submissions of
parties, a significant issue. In short, Mr Cormack’s position was this: given the
circumstances of this case such documentation was the best evidence. Mr Sandison’s
position was this: he urged that considerable caution should be exercised in retrospectively
attributing significance to the precise terms of contemporaneous documentation.
[58]       I am of the view that in the circumstances of this case contemporaneous
documentation is of significance. It appears to me that it is of particular importance in
reaching a conclusion as to the true position regarding the substantive issues. I believe it is
of significance in that:
Witnesses were being asked about events which occurred some time ago.
Questioning related to events which occurred between approximately September
2015 and February 2017. They were being asked questions about a time when
they were very busy.
Witnesses, in many cases, were giving evidence on a substantial range of issues.
Page 22 ⇓
22
It was apparent that many of the witnesses, not surprisingly, did not have a full
and clear recollection regarding many of the matters about which they were
questioned.
Mr Cormack in respect to critical parts of the evidence of CGI witnesses, namely:
that certain documents, information or assistance which was essential to CGI
performing its obligations was not provided by Agilisys, advanced the following
position: it has been developed with the benefit of posterity and does not truly
reflect the position of CGI at the time of the implementation of the projects. This
was shown clearly, he argued, by the complete absence of written requests by
CGI to Agilisys to provide, eg a data migration strategy, or specifications for
environments and no written complaints relating to these. The best CGI has been
able to do at all is: (1) to advance a completely strained reading of the response to
RN002, which seeks to read the response out of context; and (2) to rely on
Action 25 in the Partnership Board Minutes, which when read carefully, in its
context and with the email correspondence in the chains at JB 893 and 908 that
relate to it, appears to only relate to Biztalk and in any event was not actually put
to any of the witnesses involved.
There is a root and branch attack advanced by Agilisys in relation to the
reliability of nearly all of CGI’s witnesses and in relation to certain of these
witnesses the attack goes further and relates to their credibility.
[59]       Against that background, I consider that contemporaneous documentation is likely
to provide the most accurate version of what occurred as there is little chance of it having
been composed for the benefit of posterity. Accordingly it is the best evidence.
Page 23 ⇓
23
[60]       In approaching the evidence regarding contemporaneous documentation I have:
(1) not taken individual documents in isolation but have sought to see if a pattern is formed
within the documentation; (2) examined the context in which documentation was produced
and in particular in so doing I have considered whether documents formed part of ordinary
contractual correspondence or whether the document falls outwith such a category and was
one which when written would be given materially more consideration by the writer and
recipient and therefore to which greater significance can properly be attached;
(3) considered witnesses’ explanations regarding the terms of contemporaneous documents
or why contemporaneous documentation was not available; and (4) in considering the
documentation, had regard to the issue of whether the lack of contemporaneous
documentation regarding certain issues is of significance.
[61]       Moreover, in reaching my views as to what can be taken from contemporaneous
documentation I have had regard to a series of detailed written submissions made by
Mr Sandison.
[62]       The first point he made was this: the court required to have regard to the way
according to the CGI witnesses the Subcontract was operated. This evidence he argued
emphasised a partnership between CGI and Agilisys; CGI not seeking to attribute and
document fault; and CGI being interested in moving ahead with the contract.
[63]       As I understand it the court is asked to take from this evidence that it explained any
lack of contemporaneous documentation on the part of CGI raising issues with Agilisys
which CGI at proof said were of significance at the time of the performance of the contract.
[64]       This type of evidence was undoubtedly a mantra in the evidence of witnesses for
CGI and in particular in the evidence of senior members of the CGI team. I do not find this a
Page 24 ⇓
24
plausible explanation for lack of such contemporaneous documentation. It is wholly
unconvincing and I do not accept it.
[65]       The position put forward by CGI witnesses in the course of the proof was: there
were a significant number of very serious, longstanding and repeated failures by Agilisys to
meet their obligations in terms of the Subcontract. If that in fact were the case I would
expect there to be contemporaneous documentation emanating from CGI clearly, repeatedly,
consistently and with increasing force and urgency calling on Agilisys to deal with these
issues.
[66]       Because these witnesses said that they regarded the relationship between Agilisys and
CGI as a partnership, they wanted to make progress and therefore did not want to attribute
blame, does not explain the lack of such contemporaneous documentation. Such an attitude, if
in fact held by CGI witnesses, would not have stopped them raising such issues with Agilisys.
If CGI were genuinely being hampered by failures on the part of Agilisys to fulfil its obligations
I would have expected to see a consistent picture of such issues being raised with Agilisys in the
contemporaneous documentation. I believe that is particularly so given the very serious effect
which CGI said these failures were having on the implementation of the two projects. To do
otherwise would be to act in a wholly uncommercial fashion. CGI is a very large commercial
organisation and I cannot accept it would have acted in the way advanced by its witnesses. If
there were such issues they would be apparent in contemporaneous documentation where
Agilisys was called on to conform to its contractual obligations. There would be a consistent
picture of the various issues now relied on by CGI being raised with Agilisys in the
contemporaneous documentation. I found no such pattern in relation to certain issues and I
will comment in detail on this matter when looking at the evidence of certain of CGI’s
witnesses.
Page 25 ⇓
25
[67]       Secondly it was argued by Mr Sandison that a change of approach by CGI following the
issuing of RN 014 was unsurprising, as it was the first Relief Notice which sought
compensation. I do not accept this submission and in the course of dealing with the evidence
on certain issues I will make more detailed findings regarding this. However, as a matter of
generality, it appears to me that the response to RN014 was so out of alignment with what had
occurred to that point as to allow me to hold that it was a surprising change in position. The
position advanced by CGI from that point on was I believe entirely inconsistent with the
position it had adopted to that point. I will look at this matter in more detail later in this
opinion.
[68]       The third point made by Mr Sandison was this: the approach advanced by Agilisys was
irrelevant to the task of analysis which required to be performed by the court in respect of each
of the Relief Notices. In elaboration of this he submitted that the court was being asked to
consider whether obligations were breached in respect of the subject matter of each Relief
Notice. The correct question, accordingly, was not whether the parties at the time believed, or
did not believe, that there had been a breach of obligations in relation to the Relief Notice, but
rather whether there had, in fact, been such a breach.
[69]       I accept what is said by Mr Sandison as to what is the correct question for the court.
However, I am in no doubt in the circumstances of the present case, that the position of the parties
at the time throws, for reasons which I will elaborate upon later in this opinion, significant light on
the issues that the court is required to consider.
[70]       Mr Cormack advanced a second general argument in respect to the evidence led on
behalf of CGI, namely: the vagueness of the allegations. No two witnesses from CGI were
clear on what they meant by various terms such as “environmental specification or design”.
No two witnesses identified precisely the same issue. A number of witnesses had to be
Page 26 ⇓
26
prompted by questions from the bench as to what it actually was that they said they needed
from Agilisys. And no two witnesses gave a consistent explanation of precisely what that
was or what Agilisys was supposed to have done that they failed to do. Further, they bore
absolutely no relation to issues that actually happened. By the end of the proof, it is
submitted that nobody was any clearer. For example: (a) Paul Hamilton said in his witness
statement (WS) that Agilisys should have gone as far as providing High Level and Low
Level Designs: para 2.7 of his WS; (b) in evidence, he was asked repeatedly what was
required. He gave very long answers in which about the only concrete issue was, in
essence, details about versions of software to be installed on the software and unspecified
security and connectivity aspects: see Day 11, page 75-77, and again at page 100;
(c) Niall SheriffsWS contradicts him insofar as security aspects are concerned and says that
CGI would not expect Agilisys’s designs would encompass security policies etc… but they should
have provided enough detail about the computing power required, storage, connectivity, supporting
software etc…”: para 3.129 of Niall SheriffsWS. But, again, nobody suggested that
computing power was a problem, or that there were storage issues, or that there was an
issue with supporting software (other than possibly the Microsoft Office matter). Further,
no explanation was given to explain why the details contained in the Unit 4 Questionnaire
on software versions, computing power and storage were insufficient: see JB 938-939, and
particularly at pp 11933 (software and versions), 11937 (number of users), p 11940
(processor requirements), 11969 (levels of data), 11978 (Memory requirements); and
(d) Kevin Bryce just referred generically to a design: see, for example, para 2.3 of his WS.
[71]       I consider both submissions, namely the lack of contemporaneous documentation
and lack of clarity and consistency fundamentally undermine CGI’s position. I will discuss
this in detail later in this opinion.
Page 27 ⇓
27
Witnesses
[72]       The following witnesses gave evidence on behalf of Agilisys:
Robert Price:
Claire Conaghan:
Peter Hall:
Ravinder Johal:
Kevin Sander:
David Buchanan:
Louise Ah-Wong:
Dr Gill Hunt:
His role in respect to the project was principally bid lead and he
also dealt with the Relief Notices;
ICT Consultant with Agilisys and the Application Architect as
provided for in the Subcontract;
Managing Consultant with Agilisys;
Principal Consultant Agilisys, the Solution Architect as
provided for in the Subcontract;
EPR Programme Director for Agilisys;
Project Manager and worked on EI Project;
Managing Director Business Infrastructure Unit, the Business
Transformation Director as provided for in the Subcontract;
Expert Witness.
In addition the evidence of Andrew Mindenhall and Steven Beard as contained in their
statements was admitted.
[73]       The following witnesses gave evidence on behalf of CGI:
Niall Sheriffs:
Anna Vale:
Bridget Fox:
Andy Cleaver:
Paul Hamilton:
Kevin Bryce
David Legge:
Paul Neary:
Project Delivery Director from January 2015 to December 2016;
Vice President (including) Utilities and Telecoms;
Role mainly involves negotiating public sector contracts and
flow clause from Prime Contracts;
Commercial Manager;
Brought in to get development and test environments up and
running;
Involved in development and test environment, from about
February 2016;
Between October 2015 until June 2016 worked on data
migration;
Joined CGI in July 2016 and worked on the EI Project;
Page 28 ⇓
28
Stephen Waters:
Joined CGI in April 2016, worked a short time on data
migration, moved to the EI workstream, returned to ERP
workstream from September 2016;
Richard Salmond: Solutions Architect and described role with EI and ERP Projects
as peripheral (4.3 his WS);
Tom Fulda:
Directly worked on the Projects from November 2016 and also
prepared two audit reports in respect to the projects at an
earlier stage;
Kenny McAllister: Supporting the existing platform and to do the migration to the
new platform for EI;
Andrew Couch: Project Manager EI Project from September 2016;
Cliff Graham:
Partnership Director CGI from April 2016 for the contract with
CEC;
Colin Brown:
Originally his role with CGI in respect to the Prime Contract
was as Alliances Director. He became project manager ERP in
July 2016;
Jason Coyne;
Expert Witness.
In considering the evidence I did not have regard to certain sections which parties agreed for
various reasons should not form part of my consideration.
[74]       As is normal practice in the commercial court in advance of the proof each party
lodged signed witness statements of the witnesses it proposed to lead. The witnesses who
were called to give evidence adopted their statements, which in large part constituted their
evidence-in-chief.
VI Assessment of witnesses
Overview
[75]       In respect to disputed issues, I preferred the evidence of the witnesses for Agilisys to
those of CGI.
Page 29 ⇓
29
[76]       Generally I found the CGI witnesses to be defensive and very much inclined to
minimise or overlook points unfavourable to CGI and overly ready to place the entire blame
on Agilisys.
[77]       I did not detect the same attitude in the evidence of the witnesses led on behalf of
Agilisys.
[78]       Particularly in the evidence of the more senior members of the CGI staff at the time
of the Subcontract I formed the impression of an attempt to rewrite history in the period up
to about July 2016 and the issuing of RN014 and to say that Agilisys were to blame for
everything in that period despite having granted relief to Agilisys prior to that date and
despite certain matters which they raised from that point onwards and which they relied on
at proof not having appeared to be significant issues, if issues at all, up to that time. I will
elaborate on this matter when looking at the evidence of particular CGI witnesses.
[79]       I found their explanations for the granting of relief during the period up to about
July 2016 and how this fitted in with the picture that they were at proof seeking to present
wholly unsatisfactory and unpersuasive and I later will deal with specific aspects of this
when dealing with material issues which arose in relation to various Relief Notices and the
evidence of individual CGI witnesses.
[80]       To an extent I felt that the evidence of CGI’s witnesses was founded on a
misunderstanding of the roles and responsibilities of Agilisys and CGI in terms of the
Subcontract.
Detailed views on credibility and reliability
[81]       In the course of his written submissions Mr Cormack made very detailed criticisms
of almost all of the CGI witnesses (see: section 3.7) and in summary submitted that I should
Page 30 ⇓
30
not accept their evidence on disputed matters. In contrast Mr Sandison did not make any
detailed submissions in relation to this aspect of the case. He largely confined himself to
submitting that the court without difficulty could accept as credible and reliable the
evidence of each of the witnesses led on behalf of CGI.
[82]       He made one specific point in respect to the credibility and reliability of these
witnesses. He highlighted that many of the witnesses who had given evidence on behalf of
CGI no longer had any connection with CGI. He submitted that therefore in no sense were
they beholden to CGI and their evidence could be said to be independent of CGI. This
supported its essential credibility and reliability. He also made a number of comments
about the evidence of Mr Coyne and I will look at these later.
[83]       I first turn to the evidence of Niall Sheriffs, I find him an unsatisfactory and
unimpressive witness. His evidence is a paradigm example of the general observations in
respect to CGI witnesses which I have made in my overview of the evidence. Over and
above that his answers relative: (1) to the granting of relief to Agilisys in respect to certain
Relief Notices during the course of the project; (2) the reply to the Ah-Wong email (which I
will detail later); (3) his position regarding what properly could be taken from the first
report of Mr Fulda; and (4) his position regarding the signing off of SB03 I did not find
acceptable for reasons I will detail later when dealing with evidence in the context of
considering specific Relief Notices.
[84]       Moreover, I believe the further criticisms made by Mr Cormack at paragraph 3.7.3.1
of his written submissions regarding this witness are justified. Overall I believe his evidence
could not be accepted unless independently supported.
Page 31 ⇓
31
[85]       As regards Cliff Graham he looked most uncomfortable throughout his evidence. He
answered questions in a hesitant manner and often had to have what were straightforward
questions repeated. He did not make a good impression.
[86]       In regard to one particular matter of considerable significance I believe that his
change of position during the course of his cross-examination when taken with other
matters regarding his evidence caused me to believe that I could not hold this witness to be
entirely credible. I will elaborate on this point later when looking at his evidence on certain
issues. I do not find him a satisfactory witness. I am not willing to accept his evidence
unless independently supported.
[87]       In respect to Mr Fulda I believe he was unable to explain inconsistencies between his
first and second reports which he had prepared and in particular he could not explain why,
given his position in evidence, certain issues were not raised in his first report. I will
consider this matter further later in this opinion. Further criticism of his evidence made by
Mr Cormack in his written submissions was I believe justified. He was not a satisfactory
witness. His evidence could only be accepted where independently supported.
[88]       As regards Colin Brown I agree with the criticisms made by Mr Cormack in the
course of his written submissions. I gained the impression that this witness had an agenda.
In addition, despite the leading role he played in respect to the Subcontract, he constantly
downplayed his involvement and took the position that others would have to be asked
about matters, where one would have expected him to have knowledge. I agree with the
position advanced by Mr Cormack relative to the issue of “relations” and Mr Brown’s
evidence relative to this. I do not find him an impressive witness. I do not think his
evidence could be relied on unless independently supported.
Page 32 ⇓
32
[89]       Turning to Paul Hamilton, he was not involved in the project for any great length of
time (four weeks). For reasons I detail later I do not find him reliable.
[90]       Broadly in respect to Mr Hamilton I agree with this comment made by Mr Cormack,
namely: that he came across as someone who had come into a project that was already
delayed, who was not provided with all of the facts, and was understandably simply
frustrated by the process.
[91]       Regarding Kevin Bryce, I could not hold him reliable. I believe, in particular in
respect to his evidence, it was given in a somewhat one-sided manner, as argued by
Mr Cormack. He appeared as someone who was very diligent and someone who was very
much wedded to CGI’s culture, which came across as a somewhat bureaucratic one in which
matters were put down in writing and issues raised in a structured way in writing. Given
his character and his attachment to the CGI ethos, his evidence was undermined by his
inability to point to where issues he claimed were of significance at the time of the
implementation of the projects were raised in the contemporaneous documents. I will look
at this issue in detail later in this opinion. For these reasons I could not hold him reliable. In
order to accept his evidence it needed to be independently supported.
[92]       Regarding Anna Vale and Bridget Fox I agree with the criticisms made by
Mr Cormack in his written submissions. In particular so far as the day to day working in
relation to the ERP and EI Projects they seemed to have a somewhat peripheral role and did
not appear to have any significant direct knowledge. I did not find them reliable witnesses
and only where independently supported would I be prepared to accept their evidence.
[93]       As regards Andy Cleaver, broadly for the reasons advanced by Mr Cormack which I
find compelling, I reject his evidence except where independently confirmed.
Page 33 ⇓
33
[94]       In respect to Paul Neary, broadly for the reasons advanced by Mr Cormack, I reject
his evidence except where independently confirmed.
[95]       Turning to Stephen Waters, for reasons which I detail later, I did not find him
reliable.
[96]       In respect to Richard Salmond I will deal with his evidence in detail later. However,
I did not find him to be a reliable witness.
[97]       There were two further witnesses, a Mr McAllister and a Mr Couch, I do not
comment on their evidence as I did not think they added anything of significance.
[98]       The last witness for CGI was David Legge. I found this witness wholly
straightforward in the way he gave his evidence. He gave his evidence in a careful,
thoughtful and balanced manner. He was the only CGI witness whose evidence I am
prepared to accept in its entirety.
[99]       For all of the foregoing reasons I reject the evidence of the CGI witnesses unless
clearly and independently supported, other than that given by Mr Legge.
[100]       I now turn to my views as regards the witnesses for Agilisys.
[101]       My general view was this: in contrast to the witnesses led on behalf of CGI I found
them to be satisfactory witnesses.
[102]       The first was Robert Price, I felt he gave his evidence in a straightforward manner.
He answered fully and properly the questions put to him and he was not caught out in
cross-examination. I thought he was both credible and reliable.
[103]       As regards Claire Conaghan I formed the same view. Again she seemed an entirely
straightforward witness, I had no difficulty in holding her credible and reliable.
[104]       Peter Hall, I formed a favourable view of this witness and thought he was credible
and reliable.
Page 34 ⇓
34
[105]       As regards Ravinder Johal I found her a somewhat prickly witness and of all the
Agilisyswitnesses I thought she was the most defensive when giving her evidence.
However, overall I thought her evidence could be accepted as credible and reliable.
[106]       Turning to Kevin Sander, I found him an acceptable witness and thought him
credible and reliable.
[107]       As regards David Buchanan I had no concerns about this witness and believed him
credible and reliable.
[108]       Lastly there was Louise Ah-Wong, she appeared nervous throughout her evidence.
She was not a confident witness. Her evidence I thought somewhat halting and from time to
time she appeared to get slightly mixed up and was unsure in giving her answers to
questions. She was prone to giving somewhat rambling answers. I had no difficulty in
holding her credible. I considered carefully the issue of the reliability of her evidence and
came to the view that she could be relied upon. I felt that the matters to which I have
referred arose as she was someone wholly unused to appearing in court and who found it
particularly challenging when giving evidence.
The expert evidence
[109]       Turning to the expert evidence in the case, the expert for CGI was Mr Coyne. I have
come to the view that his evidence was one-sided. His approach was I believe not balanced.
In addition for various other reasons I believe his evidence was not acceptable. In arriving
at this view I considered a number of specific matters in relation to his evidence.
[110]       First I did not understand it was a contentious matter that Agilisys depended in
material respects in carrying out its obligations in terms of the Subcontract on others
fulfilling their obligations (ie party one was dependent on another party fulfilling its
Page 35 ⇓
35
obligations in order for party one to fulfil its obligations) including CGI and CEC. Thus
there were what was described throughout the proof as dependencies.
[111]       However, when Mr Coyne was asked about this, he answered in this way:
Q. But did you understand that in carrying out its work on the projects, Agilisys
depended, as you understood it, in material respects on other people, including
CGI; would that be right?
A. I’m not sure whether it would be right to characterise it in that way. For the
particular projects that Agilisys were involved in, CGI were providing services
into Agilisys.
Q. What do you mean by ‘CGI were providing services into Agilisys’?
A. Well, essentially when it comes to providing environments, the provision of the
environments was to allow Agilisys to satisfy their wider obligations in EI and
ERP.
Q. Yes. And Agilisys had a dependency on CGI providing the environments; that
would be right, wouldn’t it?
A. Yes, and they had the responsibility for managing that dependency.
Q. And would it also be right that there were dependencies on the council doing
certain things?
A. Yes, there were dependencies on the council, and, again, Agilisys were
responsible for managing those dependencies.”
(see: pages 35 and 36 of the transcript of his evidence)
[112]       It is noticeable that in this passage of the evidence Mr Coyne cannot answer the first
question he is asked with a straightforward yes, which was all that was necessary, but
uses the rather odd phrase “providing services into Agilisys”. When asked what “providing
services into Agilisys” means he goes on and gives an answer still not using the word
dependency. Eventually he accepts there is a dependency but even then, although, not
asked about the issue of management (a critical part of CGI’s case at the proof) he has to put
in that Agilisys was responsible for managing those dependencies. He gives this answer
when asked about both CGI and CEC. Thus on this matter he seems somewhat unwilling to
Page 36 ⇓
36
accept the uncontentious and straightforward proposition that there were dependencies and
even when he eventually accepts that there were dependencies he has to put a gloss on that
answer which favours CGI’s position.
[113]       Secondly, at pages 55 and 56 of the transcript of his evidence Mr Coyne is asked
about sections 3.19 to 3.34 of his report which sets out the obligations of Agilisys. He is
asked whether he has quoted any of CGI’s obligations and he accepts he has not quoted any
of them. He is asked why not. He then appears to suggest that it is Agilisys’ obligations
which are relevant not CGI’s. He is then asked specifically does he think that CGI’s
obligations are relevant and he backtracks on his previous answer and says this:
“they are relevant, I don’t believe CGI’s obligations have been called into question
with regard to the delaying and failed aspects of this project.”
[114]       He is then asked about this answer:
“Q. But you know that this whole case is about, among other things, Agilisys
serving relief notices on CGI --
A. Yes.
Q. -- indicating respects in which Agilisys thinks that CGI hasn’t complied with its
obligation?
A. Yes.”
[115]       For Mr Coyne to say that: “I don’t believe CGI’s obligations have been called into
question” I consider is an extraordinary answer given the nature of the dispute between the
parties. Fundamentally that is the case presented by Agilisys. This answer I believe
suggests a very one-sided approach by Mr Coyne, namely: he is not considering whether
CGI has breached any of its obligations.
[116]       Mr Sandison sought to deal with this issue in re-examination at pages 112 and 113 of
the transcript of the evidence. Mr Coyne began by accepting that CGI’s responsibilities were
not the subject of much express discussion in his report.
Page 37 ⇓
37
[117]       Mr Sandison then asked this:
“Q. Are you aware, having carried out the review of the contractual documentation
that you described this morning, of what the extent of those responsibilities are,
at least to have formed a view about what they are?
A. Yes, I understand that CGI have wider obligations on other aspects of different
projects with the same client.”
[118]       It is noteworthy that he does not refer to any obligations owed by CGI in respect to
Agilisys. Again it appears from this answer that Mr Coyne did not believe that CGI had any
obligations in relation to the ERP and EI projects in respect to Agilisys. This answer is again
suggestive of a lack of balance in Mr Coyne’s evidence. He again appears to be ignoring
obligations and responsibilities CGI had in respect of Agilisys in terms of the two projects.
[119]       Mr Sandison then asked:
“Q. Okay. And insofar as we don’t see detailed description and analysis of those
obligations or responsibilities, is that the product of some deliberate decision to
exclude them from your report or otherwise not to deal with them?
A. No, there’s been no deliberate decision to leave anything out of the report.
Everything that was considered as relevant whether to help inform or support
my positions or adverse to my positions were included in the report.”
[120]       That answer appears to me to be incorrect, Mr Coyne has not put in and apparently
considered, as I have just identified, what appears to be of critical importance in this case,
namely: what are the obligations of CGI in respect to the two projects. It would appear to
me to be essential in considering the various issues in this case to have regard to both the
obligations of CGI and the obligations of Agilisys. Thus in this respect his report does
appear to be essentially lacking in balance.
[121]       Alternatively, or perhaps in addition, the above answers appear to suggest a
misunderstanding on Mr Coyne’s part as to the relevance and significance of CGI’s
obligations.
Page 38 ⇓
38
[122]       Thirdly, at page 60 to 61 of the transcript of his evidence, Mr Coyne is asked about a
particular provision and whether he gave that consideration and in particular he is asked
whether he understood that it was a CGI responsibility and he said yes.
[123]       He is then asked what his position is as regards this matter and at pages 60 to 61 he
gives a lengthy explanation of his position regarding this provision.
[124]       If that was his position regarding this particular matter it ought to have been set out
in his report.
[125]       Fourthly, it appears to me to be of some significance in considering Mr Coyne’s
report as a whole that the foundation of his conclusions (see pages 61 and 62 of his
cross-examination) is his analysis of the meaning of the contract. However, his conclusion
on the meaning of the contract is arrived at, having regard to the passages of evidence to
which I have already referred, on a consideration of Agilisys’ obligations but without proper
reference to and analysis of the obligations incumbent on CGI. This I consider undermines
the conclusions which he reaches. Moreover, I am of the view that his analysis of the
meaning of the contract is fundamentally wrong and I will further detail this later in the
opinion.
[126]       Fifthly, at pages 70 and 71 of the transcript of his evidence Mr Coyne is asked about
the issue of body shopping. The questioner is careful to ensure that he and the witness
when discussing body shopping, both have the same understanding of this term. To begin
with Mr Coyne accepted that Mr Legge was body shopped from CGI to Agilisys. Following
his acceptance of that position this question is asked:
Q. Yes. So if that’s right, then … and Mr Legge was body shopped from CGI to
Agilisys, why would CGI have any interest in a data migration strategy or
templates because Agilisys was doing the activity using Mr Legge’s labour on
that analysis?
Page 39 ⇓
39
A. Well, I think that’s perhaps where… where the body shopping takes it… takes
it too far. Whatwhat essentially should have happened with the data
migration is the leadership and management of the process was Agilisys, the
setting of the strategy and the creation of the strategy documents was Agilisys.
Once that was boiled down to specific tasks which need to be undertaken, that
came down to CGI to use their technical skills to effect and provide the data
back to Agilisys. So that was the process.
Q. So is that the same thing as saying that Mr Legge was not body shopped from
CGI to Agilisys for his work on data extraction?
A. Perhaps we shouldn’t have gone down the road of trying to define body
shopping, really. Perhaps that’s the mistake.
[127]       It is quite clear at this point in his evidence that Mr Coyne realises he has been
caught out and that by his answer regarding body shopping he has undermined a position
he has developed at other points in his evidence that there had been a failure on behalf of
Agilisys to provide an adequate data migration strategy or adequate templates. When he
realises this he immediately, although with some considerable difficulty, reverses his
position. This is a short passage of evidence, however, it was in respect to an important part
of the case. CGI’s position in respect to data migration was to a large extent founded on the
inadequacy of the data migration strategy and templates provided by Agilisys. The way
that Mr Coyne dealt with this material issue, namely: switching from one position to
another was highly unimpressive in the context of someone who is being offered as giving
expert evidence.
[128]       Sixthly, at page 101 Mr Coyne is asked about Relief Notices and recognises that:
“Well, I see from some of the relief note responses that CGI do…I don’t know if their
words are ‘accept responsibility’ but they do acknowledge responsibility for delay
and sometimes give relief for that delay.”
[129]       However, as I understand it, he simply put that evidence to one side and does not
consider it to be of any relevance. I believe that if his position is that this evidence was of no
relevance he has to explain why. He has to look at the situation on the ground and explain
Page 40 ⇓
40
how his view on causation of delay fits in with what was the situation which appeared to be
accepted at the time. I do not believe that the situation on the ground can simply be ignored
as I think, having regard to the totality of his evidence, he has done.
[130]       Equally at pages 101 and 102 of the transcript he says that the Ah-Wong email which
I will set out in full later and the response thereto is of no relevance. This is simply not
discussed in his report. At page 102 he says this:
A. -- I could narrate the entire history, but I don’t believe it would add any value
to what my opinion is, is that there’s a failure to deliver an environments
design or specification.”
Q. And can you direct us to any discussion at all of 22 April conditions (set out in
the Ah-Wong email) in the section of your report dealing with the provision of
environments?
A. No, I don’t believe it’s in there.”
[131]       Again in my view this is a material failure not to consider this. I believe that this is a
relevant document. It is a document which is dealing with partiespositions regarding the
delivery of environments.
[132]       He is then asked as to how long “do you say that the failure by Agilisys that you
identify in relation to an environment design continued?”
“A. Well, I don’t believe that a design specification was ever delivered, but what
was achieved but it was through an iterative process and CGI’s delivery of
environments, testing by Agilisys, being told things are not… are wrong or
missing and things like that and then delivering another iteration of it, that the
actual then need for a design was moot then because the system or the
system… the environments had been created”
[133]       At the bottom of page 103 this matter is followed up by the cross-examiner:
“Q. So by whatever means, an iterative process or otherwise, are you able to
identify a point in time where the issue about the provision of an adequate
design for the environments was resolved or was it never resolved in your
evidence?
A. Well, it was never resolved, because a design was never provided.”
Page 41 ⇓
41
[134]       Again this seems an entirely wrong approach. It is to take the approach that there is
no specification provided at the outset and that everything else that is provided by way of a
specification following on therefrom which eventually results in environments being created
can be ignored, as somehow being of no relevance. This is again to ignore the situation on
the ground in the project. It is to ignore what was actually happening.
[135]       This I believe is an example of the criticism made by Mr Cormack of Mr Coyne that
he did not make a logical link between the criticisms he made of Agilisys and what was in
fact happening in the course of the projects.
[136]       Seventhly, an important part in Mr Coyne’s opinion is to the effect that there had
been repeated material failures in management on the part of Agilisys. His view of what
management amounts to is given in a passage of his evidence between pages 36 and 38 and
in summary comes to this: Agilisys were “accountable for ensuring that those dependencies
(the dependencies of CGI and CEC) are fulfilled” (emphasis added). He comes very close to
accepting that what he is saying in relation to the issue of management amounts to more or
less a “guarantee” although he will not use that word. However, in essence his position is
that Agilisys project management responsibilities in respect to the delivery of dependencies
by parties such as CGI or CEC amounts to making sure that the party does deliver the
dependencies (a virtual guarantee). His final answer at page 37 in respect to this issue is to
repeat that “they are accountable for making sure it gets … done”. That view on what
management in the context of the Subcontract amounts to must I believe on a sound
construction of the Subcontract be wrong. It is wrong first as it empties CGI’s
responsibilities of any content. It in effect amounts to CGI doing no more than providing
bodies, a view Mr Coyne eventually did not accept. He accepted, in the passage to which I
earlier referred, that what was happening was not body shopping. It allows CGI to argue
Page 42 ⇓
42
this: no matter for what reason delivery of a dependency on its part failed it was due to the
fault of Agilisys in that, to use the words of Mr Coyne, Agilisys is “accountable for ensuring
that those dependencies are fulfilled”. Such a wide construction of management or
leadership makes no commercial sense and goes well beyond what on a plain reading either
the word “leadership” or “management” would be understood. This view of management
is, however, consistent with Mr Coyne’s view, to which I have earlier drawn attention, as to
the lack of relevance or significance of any obligations incumbent on CGI as any failure by
CGI is turned into a failure by Agilisys to manage. Much of Mr Coyne’s opinion is based on
this understanding of the management responsibility of Agilisys. I believe his view
regarding the scope of the management responsibility of Agilisys is wrong. Thus I consider
the opinions he expresses in his report are fundamentally undermined.
[137]       Beyond the above I am also of the view that as argued by Mr Cormack at various
points in his opinion, Mr Coyne expresses ipse dixit sweeping opinions.
[138]       Lastly Mr Coyne refers only briefly to the factual matrix in respect to how matters in
terms of the projects were proceeding. Given the issues upon which he was commenting I
believe a fuller consideration of such matters was an important element.
[139]       Mr Sandison submitted that I should accept the evidence of Mr Coyne. In particular
he submitted that Mr Coyne had taken an objective and balanced approach. When looked at
fairly and carefully, his report proceeded on a tacit acceptance that CGI had responsibilities.
His report focused, having tacitly accepted the above, on the failures by Agilisys. It
concentrates on the primary failure by Agilisys which prevented CGI fulfilling its
obligations. For the reasons I have given above I do not accept this submission.
Page 43 ⇓
43
[140]       Moreover, certain further criticisms of Mr Coyne are made by Mr Cormack in his
written submissions and I believe they are soundly made criticisms. I will look further at the
evidence of Mr Coyne later in this opinion.
[141]       Overall I did not form a favourable view of this witness and I am not prepared to
accept his evidence on disputed issues in preference to that of Dr Hunt. It did not seem to
me for the above reasons that I could place reliance on any of his views.
[142]       Turning to the evidence of Dr Hunt, I found her generally to be a most impressive
witness who gave her evidence in a careful and measured fashion.
[143]       My overall impression was that the terms of her report and her evidence were
balanced in the way that they approached the issues.
[144]       I believe the contrast between her and Mr Coyne in respect to the issue of balance
was marked.
[145]       When looking at the issue of responsibilities she looked at the responsibilities of both
CGI and Agilisys. Examples of this are at 4.2.4, 4.2.5 and 6.2.1 of her report.
[146]       In addition she is prepared to make criticisms of Agilisys (see: example
paragraph 6.5.4 and 6.8.6 of her report).
[147]       The above I believe are examples of the essential balance in her approach to her
evidence, in marked contrast to Mr Coyne.
[148]       When considering issues she goes into considerable detail and provides substantial
reasoning for the conclusions she reaches. This is again in marked contrast to Mr Coyne.
Overall I found her reports and evidence, detailed, thorough, well-reasoned and balanced.
[149]       A good example of the above approach by Dr Hunt is her treatment of the issue of
project management with which she deals at section 5 of her report. In this section by
reference to documentation she clearly defines what management means and then seeks to
Page 44 ⇓
44
examine each of the aspects of management which she identifies and how these various
aspects were managed by Agilisys. In her consideration of project management she puts
this issue within its proper context by carefully considering the provisions of the
Subcontract and the relevant background.
[150]       She sets out in considerable detail her views on this aspect of the case. She combines
this with detailed reasoning and reference to supporting documents when necessary in
order to properly support her views. This can be contrasted with Mr Coyne’s approach
which starts from a misunderstanding of what management properly means, and does not
in detail define management and does not seek to put management into the context of the
Subcontract and thereafter in detail say how any failures he identifies contributed to actual
delay.
[151]       I believe that the comments I have made regarding Dr Hunt’s approach to the section
of her report dealing with management apply to all of the sections of her report.
[152]       Another aspect of her report which is of importance in considering whether to accept
or not the views she advances is this: she has proper regard to what was happening as
regards the implementation of the Subcontract on the ground.
[153]       A particularly good example of this is in the section dealing with data migration.
She starts this section (see: 8.4) by setting out the relevant pleadings on behalf of CGI. She
accepts that what is said there is a valid approach to data migration. Once more I think the
acceptance that what is being put forward by CGI is a valid approach shows the essential
balance in her report. She then discusses the approach which was actually taken by Agilisys
explaining why it had to be taken given the timetable, agreed by CGI and CEC. She then
explains how that approach fitted in with Agilisystemplate approach and points out that
CGI and CEC were aware of this approach and at the time raised no issues with it. She then
Page 45 ⇓
45
deals with the acceptability of that approach. Mr Coyne I believe in expressing his views
takes no account of matters such as the situation on the ground regarding timing in respect
to the performance of data migration which I believe he ought to have regard to for his
report to be given any weight.
[154]       Overall the report of Dr Hunt is an impressive piece of work and I believe that
reliance can properly be placed upon the views which she expresses therein. I consider that
in respect to all aspects of the dispute for the foregoing reasons her opinions were entitled to
be given substantial weight. I preferred her evidence on the disputed issues to that given by
Mr Coyne.
VII Approach to construction of the Subcontract
[155]       It was as I understand it not a matter of contention between the parties that the
proper approach to the construction of the Subcontract could be summarised in the
following way, namely: as set out in the well-known passage from the judgment of
Lord Neuberger in Arnold v Britton 2015 AC 1619 at paragraph [15]:
“When interpreting a written contract, the court is concerned to identify the intention
of parties by reference to ‘what a reasonable person having all the background
knowledge which would have been available to the parties would have understood
them to be using the language in the contract to mean’, to quote Lord Hoffmann in
Chartbrook Ltd v Persimmon Homes Ltd, [2009] AC 1101, para 14. And it does so by
focussing on the meaning of the relevant words … in their documentary, factual and
commercial context. That meaning has to be assessed in the light of (i) the natural
and ordinary meaning of the clause, (ii) any other relevant provisions of the lease,
(iii) the overall purpose of the clause and the lease, (iv) the facts and circumstances
known or assumed by the parties at the time the document was executed, and
(v) commercial common sense, but (vi) disregarding subjective evidence of any
party’s intentions.”
[156]       There was certain discussion about the relevance of intention in pre-contractual
discussions to the issue of contractual construction. It seems to me that the proper approach
Page 46 ⇓
46
to such is set out by Lord Hodge in Patersons of Greenoakhill v Biffa Waste Services 2013
SLT 729 in the following observations:
[17] Not everything that the parties knew when negotiating an agreement can be
considered when the court construes the contract. For reasons both of relevancy and
also of pragmatism the law has set its face against the consideration of parties'
statements of intention in the negotiations leading to the contract. There is recent
authority for this (Chartbrook Ltd, Lord Hoffmann at pp.11151121, paras 2742;
Luminar Lava Ignite Ltd v Mama Group Plc at 2010 S.C., pp.319321; 2010 S.L.T.,
pp.153154, paras 3945). Again there is also older authority in this jurisdiction
(Buttery & Co Inglis, Lord Gifford (dissenting) at (1877) 5 R., pp.6970; Inglis v
Buttery & Co, Lord Blackburn at (1878) 5 R. (H.L.), pp.102103).
[18] The rule excluding statements of intention in pre-contractual negotiations has its
limits. In Chartbrook Ltd Lord Hoffmann stated (at p.1121, para.42): The rule excludes
evidence of what was said or done during the course of negotiating the agreement
for the purpose of drawing inferences about what the contract meant. It does not
exclude the use of such evidence for other purposes: for example, to establish that a
fact which may be relevant as background was known to the parties, or to support a
claim for rectification or estoppel. These are not exceptions to the rule. They operate
outside it.’”
[157]       Accordingly I have only had regard to statements of intention in pre-contractual
negotiations for the limited purpose identified by Lord Hodge.
VIII The roles of transformation director, application architect and solution architect
[158]       Before turning to look at the Relief Notices and the specific questions posed by
parties in respect thereto there is a further general issue I require to deal with and this is a
submission made by Mr Sandison. It was his submission that the Subcontract provided for
three key roles in the implementation of the ERP and EI projects, namely: transformation
director; application architect; and solution architect.
[159]       It was his position that these roles reflected Agilisysleadership role in relation to
these two projects. Moreover, the roles themselves contained important responsibilities
which required to be properly discharged in order for the Subcontract to be properly
operated.
Page 47 ⇓
47
[160]       The first of these roles was business transformation director. This role is defined in
Annex 4 to Schedule Part 8.1 of the Subcontract which provided:
“The business transformation director will be a Supplier employee and will represent
the interests of both the Supplier and CGI in dealings with the Authority. The
purpose of the role is to develop new business opportunities with the Authority
under the Business Case Development Services, and opportunities for projects with
PSPs. The person will work closely with the CGI team to:
Define, agree and initiate transformational work or new business
opportunities under the Business Case Development Services.
Manage the quality of the OBC and Business Case Development Services for
CGI and the Supplier, commissioning support from the relevant Party and
managing the delivery of the agreed outputs to the Authority.
Act as a trusted advisor to senior and executive Authority staff.”
The Supplier in this context is Agilisys and the Authority is CEC.
[161]       Business transformation activities were widely defined in terms of the Subcontract
and included the responsibility of leading CEC through the change process:
“Consultancy activities to support a client in changing the way it carries out its
business processes or organises its workforce to deliver its business services to
enable the implementation of Supplier IPR, Products, Knowhow. Such consultancy
includes the following to the extent they are specific to and depended upon detailed
knowledge of the Supplier IPR, Products, Knowhow: business case development,
options appraisal, organisational design, change management, business process
analysis and design, application selection and project management of application
implementations.” (See: Annex A to Schedule Part 4.1, JB 1101.)
The person designated to hold this role was Louise Ah-Wong of Agilisys.
[162]       On a sound construction I think that this role was in large part a forward facing one
and the extent to which it was engaged in the implementation of the ERP and EI projects
was limited. I do not consider that it was a key role in the implementation of the ERP and EI
projects.
[163]       I observe that it is of critical importance in understanding this role to note its
purpose:
“The purpose of the role is to develop new business opportunities with the Authority
under the Business Case Development Services, and opportunities for projects with
PSPs.”
Page 48 ⇓
48
It is clearly a development role in respect to “new business opportunities” with the
Authority and “opportunities for projects with PSPs” (emphasis added). The role was
forward looking in this sense: it was one which related to the future and was not intended
given its purpose to play any significant role in the implementation of the ERP and EI
projects which were current projects.
[164]       The bullet points which follow on in the description of the role must be read in the
context of the purpose of that role.
[165]       I do not of course say that it has no relevance to the ERP and EI projects, however, it
cannot be said to be a major role. I will where necessary comment further on the role in
terms of the specific parts of the projects considered in terms of the questions posed to the
court.
[166]       Secondly there was the role of application architect (the role was held by
Claire Conaghan), the responsibilities of which are as follows:
“that the Application Architect was to ‘work with clients to define an overall
enterprise solution/service architecture that supports the delivery of existing and
new services.’ The responsibilities of the Application Architect included:
Understand the role and configuration of systems that support business
process with specialism across local government application landscapes.
Define information systems architecture to support clients’ business strategy
or short-term objectives and ensure that technology objectives are aligned to
business goals.
Closely engage with customer stakeholders to determine the systems and
application requirements to support business plans and activities;
recommend technology opportunities and innovations to support business
improvements.
Build and sustain a relationship with key customer stakeholders
Support programmes and projects with architectural support and
requirements analysis
Page 49 ⇓
49
Provide guidance and direction on the procurement of technologies;
developing cost effective solutions to business requirements, optimising
technologies already used and identifying suitable replacements and
advancements
[…]
Document and communicate the architecture and its benefit and impact on
business processes to key stakeholders
Work with the project teams, infrastructure architects, third party suppliers
and other stakeholders to assemble and manage technical project plans and
deliverables including time and cost estimations to deliver the solution.
Evaluate and select suitable products from 3rd party vendors to deliver
solutions
[…]
Work with technical architects to ensure that non-functional requirements
such as security, resilience, performance, on-going maintenance and other
considerations are accounted for in the overall architecture
Provide Technical leadership and guidance to solution architects, application
service delivery teams as required.” (See: Annex 4 to Schedule 8.1 of the
Subcontract)
[167]       This role is, given the definition, a fundamentally client facing role. The role is “to
define an overall enterprise solution/service architecture…” (emphasis added). It is clearly a
very high level and strategic role. The various responsibilities have to be read in that
context. On a plain reading there are clearly responsibilities which have relevance to the
implementation of the ERP and EI projects and I will consider these further where this role
is referred to in respect to the various questions posed for determination by the court.
[168]       The third key role was said to be that of solution architect, which role was held by
Ravinder Johal of Agilisys.
[169]       In terms of the PID the solution architect was to:
(a) co-ordinate all activities in relation to system environment set up;
(b) co-ordinate and manage all activities related to bespoke developments;
Page 50 ⇓
50
(c) provide advice and guidance on enhancing system performance;
(d) liaise with existing ICT supplier, BT, to obtain any necessary data.
[170]       On a plain reading there are responsibilities which have relevance to the
implementation of the ERP and EI project and I will consider the significance of these in
respect to the various questions posed for determination by the court.
IX The nature and scope of the leadership/management role of Agilisys
[171]       It was a recurring theme in the submissions of CGI that Agilisys had breached its
obligations to lead and manage the two projects and that this had been causative of delay. I
have touched on this to a limited extent when considering the evidence of Mr Coyne.
[172]       There was no dispute between the parties that in respect to the ERP and EI projects,
Agilisys was the party with the specific experience in dealing with such projects in the
context of local government. Nor was it a matter of contention that in general terms Agilisys
was to lead or manage the implementation of these two projects.
[173]       The issue between the parties was the scope and extent of that role.
[174]       Mr Cormack in advancing his position on this issue argued that the Subcontract
clearly envisaged and recognised that a high degree of co-operation, mutual partnership,
and interdependence was going to be necessary. However, when regard is had to the
structure and context overall, it emphasised the junior or subordinate position of Agilisys as
a party that was ultimately, reliant on and under the control of CGI as the prime contractor.
[175]       In support of these propositions Mr Cormack relied in summary on the following
aspects of the contract:
The express governance arrangements contained in clause 11 and Schedule
Part 8.1 of the Subcontract;
Page 51 ⇓
51
Agilisys’ reliance on CGI’s PMO function (roles and responsibilities document in
Schedule Part 4.1);
The partnership charter (Annex 2 to schedule 8.1);
CGI’s responsibility for management of CEC (clause 5.17 and Schedule Part 3);
CGI cause and Relief Notices (clause 31);
The length of the Subcontract and business case development and PSP
opportunities (clause 4.12, 5.2 and Schedule Part 4.1);
[176]       Against that background he submitted: that sweeping statements such as the
following, it is “apparent from the contractual documentation that [Agilisys] was required to
be the programme manager for the delivery of the associated projects”, are misleading and
utterly simplistic. The quote was from the report of Mr Coyne at paragraph 3.3.
[177]       He submitted that what CGI was seeking to put forward was a position that the
leadership role of Agilisys extended to managing, and taking contractual responsibility, for
the way in which CGI performed its roles and responsibilities or for a failure by CEC to do
the things that CGI had undertaken to procure it would do.
[178]       The above he submitted would fundamentally invert the careful division of
responsibilities and the associated careful arrangement of charges for provision of services
in Schedule Part 7.1 that was agreed between the parties in the Subcontract. By design, he
submitted, Agilisys had absolutely no direct contractual relationship with CEC and no
method of forcing them to do anything other than relying on CGI to exercise contractual
powers.
[179]       In development of this argument he submitted that the construction and approach
put forward by CGI effectively emptied CGI’s responsibilities of any practical content other
than the provision of bodies. Indeed, at various points in their evidence CGI appeared to be
Page 52 ⇓
52
making just such an extreme case in their references to “body shopping”. He submitted that
there were two obvious objections to that construction of the Subcontract.
[180]       First, if that was what parties had really intended, it was not easy to understand the
reasons why they would have gone to such lengths in the Subcontract to demarcate various
areas of responsibility or would have altered the financial model so as to simply remove any
bodies from Agilisys for the relevant tasks. If that were truly the arrangement, Agilisys
would simply have taken on the entire responsibility for everything on the project together
with all of the profit and would have entered a separate agreement that CGI provide bodies
at a lower cost.
[181]       Secondly, and closely linked to the first point, it made absolutely no commercial
common sense. This construction involved the absurd suggestion that Agilisys agreed to
keep all of the commercial and contractual risks for the areas of responsibility that CGI took
back from them in the bid phase but lost all of the profit or direct control over delivery. That
would require the clearest of language in the Subcontract to achieve such an uncommercial
and obviously ill-advised arrangement. He submitted that there was nothing in the
Subcontract that achieved that degree of clarity. Further, the wider context made it clear
that this was not simply a body shopping exercise. As Anna Vale accepted in
cross-examination, the whole language of dependencies was inconsistent with a body
shopping exercise and further Mr Coyne did not accept any suggestion of body shopping.
None of CGI’s witnesses, apart from possibly Niall Sheriffs, genuinely supported the
argument that the true nature of the arrangement between the parties was one of body
shopping.
[182]       Thus he submitted that on the basis of some general leadership concept CGI in
respect to data migration sought to categorise its own failure as one of mismanagement by
Page 53 ⇓
53
Agilisys of CGI and that the failure to build environments became a failure by Agilisys to
tell CGI how to do it.
[183]       In relation to this issue Mr Sandison in his written submissions said this:
“16. The Subcontract related to the ERP and EI projects which formed part of the
far wider scope of the Prime Contract between CGI and CEC. CGI had the
relationship of Prime Contractor with CEC. CGI was also responsible for the
programme-level relationship with CEC, which extended beyond the two projects
with which Agilisys was involved. Throughout the Subcontract, Agilisys is referred
to as the ‘Supplier’.
17. Notwithstanding Agilisys’ position as a subcontractor, however, Agilisys had
specific responsibilities of leadership in relation to the EI and ERP Projects. The
Subcontract was not operated in a manner whereby Agilisys, as subcontractor, was
subordinate to CGI. As Robert Price recognised in his evidence, this was a contract
in which both sides required seats ‘at the top table’. The ‘whole approach from the
outset was one of partnership and collaboration across the board’.
18. Whilst CGI led at a programme level, therefore, Agilisys was responsible for
leadership of EI and ERP at a project level. Specific aspects of this have been noted
in relation to individual Relief Notices, below. In particular, the Responsibilities
Document provided in relation to project roles for both ERP (which was divided into
a Finance and Procurement element and an HR and Payroll element) and EI:
‘Supplier Leading delivery of this project,’
19. These contractual responsibilities were reflected in the practical operation of
the Subcontract. In his evidence, Robert Price acknowledged that the ERP and EI
projects were led by Agilisys. Agilisys was the overall project manager for ERP.
This reflected Agilisys’ expertise in implementation (and, indeed, much of the
day-to-day’ work by Agilisys did not require CGI’s input). EI was a project which
Agilisys was ‘managing and delivering’. A component of the role of project manager
in that project, in Mr Price’s evidence, was to ‘seek to find out’ why a participant in
the project has not produced a deliverable, and, having done so, to attempt to assist.
Further, it was important that the leading role played by Agilisys encompassed not
only technical leadership but also change management (something which was
reflected in the specific division of roles and responsibilities in the Subcontract). As
it was expressed by Cliff Graham,
Agilisys were responsible for successfully delivering the ERP project and ultimately
ensuring that CEC was receptive to change was part of that. Agilisys couldn’t
decouple that from their responsibility to ensure that the project progressed to the
plan.
20. The leadership role to be played by Agilisys reflected the respective expertise
of the parties. Louise Ah-Wong’s evidence was that ninety per-cent of Agilisys’
business and experience was in local government. This included both ‘one of the
Page 54 ⇓
54
largest Unit4 ERP practices in the UK’ and, in relation to EI, the AD product which
was to be used for the EI Project. Anna Vale’s evidence was that the decision was
taken to work with Agilisys because Agilisys were ‘promoting themselves as experts
in local government’ and ‘as a result of that expertise’.
21. It is submitted that a key theme in relation to the operation of the Subcontract
and the subject matter of the Relief Notices was Agilisys’ failure adequately to
discharge its obligations of leadership. Agilisys was required to provide leadership
both in relation to CGI (in respect of the delivery of the EI and ERP Projects) and also
in relation to CEC (who required to be led through the process of transformation).
Specific aspects of this are discussed in relation to individual Relief Notices. At a
general level, as a result of Agilisys’ failure to provide the expected level of
management, CGI were required to put in place additional management for the ERP
and EI projects.”
[184]       In respect to this issue I prefer the position advanced by Mr Cormack. When dealing
with the evidence of Mr Coyne I have already given certain reasons why I consider the
position taken by Mr Cormack is the correct one. In elaboration of those reasons I would
make the following observations.
[185]       First, to state the obvious Agilisys was a subcontractor of CGI. This as a matter of
generality placed it in a subordinate position to CGI. It had no contractual relationship with
CEC. I believe this is significant in considering the scope and extent of its leadership and
management role in respect to CEC and has to be considered in the context of the
responsibility which lay on CGI “to procure that CEC fulfilled the roles and responsibilities
allocated to them in the Subcontract” (emphasis added) which I will consider further below
(see: Schedule Part 3 Clause 2.1.3).
[186]       Secondly, Mr Cormack directed my attention to the express governance
arrangements. In terms of this the parties had agreed how the services to be provided under
the Subcontract were to be managed and governed, the structure included: (a) nomination
of suitable representatives with authority to act on behalf of either party; (b) a series of
mutual obligations about a provision of suitable resources and staff; (c) proactive
management of risks attributed to each party; and (d) ensuring both parties provide
Page 55 ⇓
55
representatives to attend the various boards. It is noteworthy that the chairperson for each
board was given the responsibility for scheduling, preparing agenda, minuting, and
facilitating giving effect to decisions (see: paragraph 5.2 of Schedule Part 8.1). Agilisys was
not the chairperson for any of the boards. The programme management review board was
the only one exclusive to Agilisys and CGI. It was to be chaired by a CGI representative.
Thus the governance arrangements in the Subcontract, as argued by Mr Cormack, reflect the
common sense position that Agilisys was a subcontractor and junior partner that was
obliged to fit within CGI’s wider governance arrangements. Accordingly, it was to remain
subject to CGI’s ultimate direction and control.
[187]       Turning to the PMO function (this is a reference to the Project Management Office)
whilst, as was accepted, Agilisys had to lead the ERP and EI projects, it was expressly
acknowledged that in doing so Agilisys would “rely on CGI PMO for support. CGI to
support with the Authority governance and resources.” This once more reflects the general
contractual position that Agilisys were subcontractors with no contractual relationship with
CEC and no means of compelling them to do anything. The party with such power was CGI
who had the direct contractual relationship with CEC and had in addition the responsibility
to leverage its commercial powers to ensure CEC complied with its obligations (see:
schedule 4.1 section C of the Subcontract, JB 1, p 1080).
[188]       Thus overall having regard to the above I consider it was for Agilisys to work within
the wider project management arrangements operated by CGI, it was not its responsibility
to set up and operate a different approach to governance.
[189]       In respect to the issue of CGI management of CEC I observe it was a responsibility of
CGI to procure that CEC fulfilled their roles and responsibilities allocated to them in the
Subcontract (para 2.13 in Schedule Part 3 and Annex 2 to Schedule Part 6.1.)
Page 56 ⇓
56
[190]       As to the meaning of the word “procure” I found assistance in arriving at a true
construction in the observations of Mr Justice Leveson as he then was in Barnicoat & Another
v Knights & Others [2004] EWHC 330 (Ch).
[191]       The clause that was being considered by Mr Justice Leveson was in the following
terms:
“The purchasers shall procure the repayment of the director’s loans upon the
following terms:
3.63 The purchasers shall procure the company to pay interest on the principal
amount outstanding being not greater than…”
[192]       At paragraph 23 he holds that on a true construction “ ’procuremeans more than
simply command. Its ordinary meaning is to achieve a result. In my judgment clause 3.6
has the same legal effect as a guarantee.”
[193]       The precise context in which Mr Justice Leveson was considering the meaning of the
word procure is different from the present case. However, what I believe is important is
that what he has regard to is the ordinary meaning of the word procure. In the
circumstances of the present case I am unable to identify any reason why I should apply any
other meaning to the word “procure” than its ordinary meaning. Accordingly I am of the
view that the obligation incumbent on CGI was on a plain reading of the provision to
guarantee CEC’s fulfilment of its responsibilities.
[194]       I have also found some assistance in arriving at this conclusion in the observations of
the Mr Justice Peter Smith in Nearfield Limited v Lincoln Nominees Ltd and another
[2006] EWHC 2421 (Ch) at paragraphs 96 and 97 who reached the same conclusion as to the
meaning of the word “procure”.
Page 57 ⇓
57
[195]       It appears to me that the critical responsibility in relation to management of CEC is
placed on CGI. Ultimately it is for CGI to ensure that CEC complies with its responsibilities;
it has guaranteed that CEC will do this.
[196]       Beyond the above I consider that the mechanism of CGI cause and Relief Notices also
implicitly recognise the extent of Agilisysdependence on CGI fulfilling its responsibilities.
This mechanism is difficult to understand if that is not correct.
[197]       I consider that this general leadership/management role which is summed up in the
provision referred to by Mr Sandison “supplier leading delivery of this project” cannot be
turned into a management role whereby Agilisys takes contractual responsibility, for the
way in which CGI performed its roles and responsibilities or for a failure by CEC to do the
things that CGI had undertaken to procure it would do. As argued by Mr Cormack, such an
approach is not consistent with the division of responsibilities in the Subcontract. Beyond
that, it makes no commercial sense for the reasons advanced by Mr Cormack.
[198]       Further the above analysis I also observe fits in with the analysis of Dr Hunt in her
initial report in section 5. She analyses what management means in terms of the
Subcontract. She identifies four detailed areas which it encompasses, namely: planning;
progress; risk, issue and action tracking; and escalation. Her analysis is clear and
compelling as to the scope and nature of the management role and how it required to be
carried out within the context of a much larger contract where Agilisys was the
subcontractor. Thereafter, Dr Hunt between paragraphs 5.2.4 and 5.5.4 of her initial report
carefully analyses and comments upon Agilisys’ performance of its management role: first
within the context of the wider project management system operated by CGI; and secondly
having regard to practical difficulties it encountered such as the high turnover of CGI
personnel in the course of the implementation of the projects (see: 5.3.16) and finds its
Page 58 ⇓
58
performance satisfactory. I believe her analysis and conclusions regarding the scope of the
general management role of Agilisys and regarding its performance of that role should be
accepted.
[199]       Lastly, the above analysis of the management role fits with the evidence of the wider
context which came from Ms Vale and Mr Coyne, namely: their rejection of the concept of a
body shopping exercise. I think that it would only be in the context of such an extreme
position of body shopping that Agilisys management role could have had the scope
contended for by CGI. Ultimately Agilisys relied on and were under the control of CGI.
[200]       For all of the above reasons I reject the overarching position developed by
Mr Sandison across all aspects of the project that there was a fundamental failure by Agilisys
to discharge its responsibilities of leadership and management of the projects. I will further
refer to this issue as I consider the individual Relief Notices.
The Relief Notices
ERP data migration, RN 007 and RN 016
[201]       RN007 concerned delay of provision of data extracts. Agilisys asserted “a CGI cause
is reasonably likely to cause Agilisys to fail to achieve a Milestone Date System Acceptance
Testing... CGI have not provided a data extracts from existing Finance, HR and Payroll
systems and have not confirmed a date by which these will be available.” RN016 also
concerned delay of provision of data extracts. Agilisys asserted a CGI cause is reasonably
likely to cause Agilisys to fail to achieve a Milestone Date User Acceptance Testing (UAT).
An earlier Relief Notice, covering exactly the same cause, was raised on 18 March 2016
(RN007) but CGI have still not provided data extracts from existing Finance, HR and Payroll
systems and have not confirmed a date by which these will be available.
Page 59 ⇓
59
[202]       RN007 was based on a failure to provide extracts by either the original deadline
agreed between the parties in version 9 of the Project Plan of 22 December 2015 or by the
later deadline in version 12 of the Project Plan of 3 February 2016. RN016 was based on the
continuing failure to provide extracts by the revised deadline that had been agreed between
the parties of 17 August 2016 in version 13 and 29 September 2016 in version 16.6 of the
Projects Plan. These deadlines were agreed between the parties in the context of the
conditions in the Ah-Wong email (which I will detail later in this opinion) and the replan
and relief which was granted at the time.
[203]       Data Extraction forms part of the data migration process and involves the process of
retrieving data from an old system (the legacy system), typically in a neutral format such as
a CSI file. Data migration is the process of moving data from one system to another, most
commonly so that a new system can make use of data contained in the legacy system. In the
present case CEC data had been held on a legacy system operated by BT and was being
moved to a new system (the target system).
[204]       The first question in relation to this group of Relief Notices is this: What were the
respective obligations of the parties (including as to the Milestone Dates in the
Implementation Plans and timing generally) in relation to the provision of data extracts from
CEC’s existing Finance, Human Resources and Payroll systems?
[205]       Agilisys submitted in its written submissions that the Subcontract allocates the
following responsibilities to CGI (either directly or by virtue of its obligation to procure that
CEC perform its responsibilities).
Section
Extract
OBS 1, OBS 2, OBS 3,
OBS 30
Provide technical resource to develop all data migration scripts
from the legacy system…and manage the delivery of any 3rd
parties such as BT to complete.”
Page 60 ⇓
60
“Address data quality issues that will impact the integrity of the
system.” (via CEC)
Manage the contractual relationship with BT enabling CGI to
work as [CEC’s] agent.” (via CEC)
OBS 47
Undertake the data migration out of PPSL, Oracle e-business
and iTrent to enable Agilisys to import to Unit 4 Business
World.”
“Address data quality issues that will impact the integrity of the
system.”
Roles &
Responsibilities
Document, Section C
of Schedule Part 4.1
Manage the contractual relationship with BT enabling CGI to
work as [CEC’s] agent.” (via CEC)
Project Roles:
“CGI providing data migration and integration services.”
[Council] is responsible for Data cleansing”
CGI Responsibilities & Deliverables:
Technical resource to define the technical specifications of the
data migration extraction files, create the data extraction scripts
and undertake all testing and documentation to complete the data
extract process, working to the Agresso solutions architect. The
data will be extracted from the legacy systems Oracle e-business,
and PPSL..”
Annex 2 to the
Implementation Plans
(p 1125 of JB)
CGI shall procure that the Authority will comply with the duties
listed below, as required to deliver the Authority functions listed
in Schedule 2.1 (Service Description):
Request the Incumbent Supplier to provide data extracts in
accordance with specifications provided by the Supplier.
It further argued that the Subcontract allocates the following responsibilities to Agilisys:
Page 61 ⇓
61
Section
Extract
OBS 1, OBS 2, OBS 3,
OBS 30, and OBS 47
Define and agree the solution design with CEC based on Agilisys
local government templates and agreed through workshops with
CEC.”
“Working with CEC, define the technical specification of the data
to be migrated from the [relevant legacy system]
“Undertake the data migration into Unit 4 Business World using
the data extracted by CGI teams.” (NB. In OBS 47 additional
words ‘ERP Solution’ are inserted after ‘Business World’ but
otherwise is identical)
Identify any data issues that will impact the integrity of the
system.”
Roles &
Responsibilities
document, Section C
of Schedule Part 4.1
Project Roles:
(9) [Agilisys] to provide templates and specifications for data
migration requirements.
Supplier Responsibilities & Deliverables:
Solutions architect resource which will generate…data migration
strategy…Each of these will produce a document deliverable for
sign-off with the Authority. The solution architect will manage
the technical resource provided by CGI.”
“Lead on the creation of the technical specifications, specifying the
data translation rules. CGI will lead on the lower level
specification elements such as coding standards, error trappings
etc..
“Technical resource to undertake migration into Agresso, using
the data extracted by the CGI technical data migration resource.”
[206]       CGI referred to three other responsibilities incumbent upon Agilisys, I do not
understand that these were disputed by Agilisys. They were:
Facilitate design workshops to agree the implementation design of the Agresso
system this will cover all business areas and include interfaces, data migration
and reporting. Deliverables will include the workshop materials and the overall
solution design document.
Page 62 ⇓
62
Change management resource to lead the Authority through the change
programme. Deliverables will include change management strategy,
communications plan and materials for relevant workshops and
communications
The archive solution will be configured by Agilisys and Agilisys will import the
archive data provided by CGI that has been exported from the legacy systems.
[207]       CGI emphasised that the above responsibilities were incumbent upon Agilisys in
addition to the general obligations of leadership and management which I have already
discussed and in respect to which I have set out my views as to the scope of this
responsibility. Lastly Mr Sandison stressed that the solutions architect was to produce a
data migration strategy and to manage the technical resource provided by CGI.
[208]       I consider that the obligations referred to above by each side in respect to these Relief
Notices were incumbent on the parties and were relevant to the discussion of the various
issues relative to these Relief Notices.
[209]       At paragraph 5.2.4 of its written submissions Agilisys argued that on a proper
construction of these contractual provisions the scope of the responsibilities on a proper
reading of the foregoing provisions is as follows:
“Agilisys were to work with the Council to produce a documented and signed-off
design of the Business World solution (including data migration strategy). Secondly,
in implement of that design, Agilisys were to provide suitable templates to CGI.
Thirdly, CGI had to use the templates to undertake all of the work to extract the data
from the legacy systems (including, managing BT and providing all of the technical
programming such as scripts or specifications of the extraction files). Fourthly, CGI
had to give those extracts to Agilisys. They, also, had an absolute obligation to
procure that CEC would request the Incumbent Supplier to provide such data
extracts to Agilisys. Fifthly, Agilisys had to load that data into Business World.”
[210]       As to a sound construction of the various provisions I observe first that it is implicit
in the whole structure and nature of the contract and the nature of the data migration work
itself that Agilisys was dependent on CGI to perform its obligations in relation to extraction,
in order for Agilisys to load the data (for which Agilisys is responsible). Data migration
Page 63 ⇓
63
involves three principle steps performed sequentially, namely: (a) extract; (b) transform;
and (c) load. Secondly, I am persuaded that on a plain reading of the Subcontract the
submission put forward by Mr Cormack is broadly correct as to what is a sound
construction of the parties’ respective responsibilities as regards data extraction subject to
this: the general management role of Agilisys as I have earlier defined it and the express
obligation that “the solution architect will manage the technical resource provided by CGI”.
For the reasons which I have already given the foregoing cannot be turned into a
management role whereby Agilisys takes contractual responsibility for the way in which
CGI performed its roles and responsibilities or for a failure by CEC to do the things that CGI
had undertaken to procure it would do. The proposition that CGI was doing no more than
providing a body to be managed by Agilisys (body shopping) and nothing more is not for
the reasons I have earlier set out consistent with the terms of the Subcontract when looked at
as a whole. In this context I believe management of the resource amounted to this: giving
assistance and advice where necessary or requested. Agilisys was not responsible in respect
to the data migration process in the way submitted on behalf of CGI.
[211]       In respect to issues of timing and whether CGI had failed to provide the data extracts
within the time necessarily required by the Implementation Plan, Agilisys submitted that the
obligation required to be construed in light of the sequence of events in the Implementation
Plan. Thus for the reasons I have set out above, Agilisys was dependent on CGI performing
its obligations first: Agilisys could not load the data until CGI had extracted it. In the
Implementation Plan in the Subcontract, data migration, building and testing was agreed as
being a lengthy activity commencing on 1 December 2015 and concluding on 2 September
2016 (see: Schedule Part 6.1, JB 1, p 1119).
Page 64 ⇓
64
[212]       As set out earlier I have held that agreement of certain dates by parties can be had
regard to in judging what was a reasonable time in terms of the Implementation Plan as to
when dependencies required to be produced. Thus I have had regard to the dates set out in
the Project Plans and the Implementation Plan set out above when considering this question.
[213]       It is clear from the chronology of events, which I will set out when considering the
second question relative to these Relief Notices, that the data extracts were not provided
within the time necessarily required by the Implementation Plan. CGI accordingly breached
its obligations in respect of data migration. Therefore in respect of the question did CGI
breach its obligations by failing to provide data extracts from CEC’s existing Finance,
Human Resources and Payrolls systems? I accordingly answer yes. In any event when
CGI’s response in its written submissions is considered it is in the following terms:
151. Certain data extracts were provided by David Legge between 19 and 31 May
2016, but these could not be validated by Agilisys. In August 2016, data migration
entered ‘special measures’. At that stage, sixteen of the required twenty-seven data
extracts had been provided. Following the entry of data migration into special
measures, however, it became apparent that there were significant difficulties in
defining the scope of live data migration (in addition to historic data, which is
discussed in answer to question 16, below). These issues had not been resolved when
Agilisys left the project in March 2017.
152. Accordingly, CGI did not entirely meet its obligation to provide full data
extracts.
[214]       The next question is this: if the answer to question 14 is yes, was CGI’s breach
caused or affected by a breach by Agilisys of its obligations?
[215]       As above set out CGI accepted it had not entirely met its obligations to provide data
extracts. However it went on to submit this:
“It is submitted, however, that this failure was a result of the failure of Agilisys to
comply with its obligations under the Subcontract in respect of data migration.
CGI’s position was set out in its response to RN016, issued on 13 October 2016.”
Page 65 ⇓
65
[216]       Given CGI’s position regarding the second question in relation to this group of Relief
Notices, the central question in regard to data migration becomes this: was CGI’s breach
caused or affected by a breach of Agilisys of its obligations?
[217]       As regards the period up to the email JB 716 referred to in the next paragraph, the
chronology of events in respect to data migration was as set out in an Appendix produced
by Agilisys along with its written submissions. I believe this correctly sets out the timeline
during that period and the causes in that period for extracts not being produced. Moreover,
I consider that the appendix correctly sets out the timeline for the whole of data migration
including data archiving and the causes for extracts not being produced. This is Appendix 1
to this opinion.
[218]       The principal difficulty in relation to implementation of data extraction was
obtaining access to the legacy system. This issue is concisely captured in (1) David Legge’s
email of 7.1.16 (JB 716): “I still don’t have access to ANY of the legacy systems. It’s getting
beyond a joke now.”; and (2) an email (JB 738) on 21 January 2016: “Getting access to
HR/Payroll and Finance still seems to be problematic. Looks like the issue is around the fact
that they don’t want us working with personal data and they do not have test instances that
have been anonymised.” At the beginning of February 2016, it was confirmed that
David Legge had access to Oracle and iTrent. But the access granted to him by BT was such
that he was not actually allowed to do any extracts and was read only. His access was
conditional on not accessing live data other than to anonymise it: see emails at JB 790,
and 796.
[219]       By approximately May 2016 when Mr Legge left CGI he still did not have full access
to the legacy systems. And he had not passed any data extracts to Agilisys: see para 2.38 of
David Legge’s WS. As he said in evidence: “We didn’t get to the point of being able to
Page 66 ⇓
66
transfer any data to Agilisys.” See: page 65, line 6 of the transcript. Accordingly, very little
actual progress was made with data extraction. The following events between March and
May 2016 I believe are noteworthy:
On 18 March 2016, Agilisys issued RN007 relating to Data Migration delays:
Peter Hall’s WS 8.43, and see RN007 at JB 16. It was issued then because CGI’s
delay in providing full data extracts to Agilisys it was asserted was going to
affect their ability to meet the System Acceptance Testing Milestone. By this
stage, CGI was almost 6 weeks late in delivering full data extracts in accordance
with the Implementation Plan v 12 deadline of 3 February 2016. This was
ultimately accepted insofar as Relief was granted as a result of the ERP re-plan
agreed with the Council at the end of April 2016: see response to RN007 among
others at JB 30.
By early April 2016, Kevin Bryce, Stephen Waters, and David Legge were
meeting with CEC to try to persuade them to let Agilisys have access to live data:
see JB 993. Following this meeting it was identified by Derek Brown of CGI (who
bears to have been involved in data security), and having discussed it with CEC,
that a statement of why access was required needed to be put together to obtain
formal approval. Kevin Bryce acknowledged in an email that CGI had missed
the data extract deadline in February 2016, and wanted to set a new date of
18 April 2016, which would still be unlikely to hit but would mean they could
push David Legge: see Peter Hall WS 8.42.3, and JB 1001-1002). This meeting was
followed by a meeting between Ravinder Johal, Stephen Waters, and Derek
Brown to discuss the necessary issues around access to live data, and
Derek Brown then puts the document together: see references in JB 1037, 1038
Page 67 ⇓
67
(WS of Ravinder Johal at paragraph 4.15.6 and Steven Waters pages 30-31 of the
transcript of his evidence). This document was ultimately rejected and a Privacy
Impact Assessment had to be submitted: see JB 1144. It is in this context that
Peter Hall was asked to get the consent of the Information Asset Owners within
the Council on 26 April 2016, and ultimately only obtained full consent on
19 May 2016: see JB 1241.
In the course of April 2016, Stephen Waters and David Legge bear to have started
to try to make more progress. In particular, it is in this context that there are
discussions about provision of further templates to David Legge, which
ultimately were sent on 22 April 2016: see emails in JB 1030, and JB 1141.
There is the sequence that was explored in cross-examination with
Stephen Waters and Colin Brown in relation to the preparation of the plan
provided for in the 22 April Conditions: see JB 1241. This plan is notable because
it failed to meet the deadlines of a first extract cycle being completed by 7 June
2016, and a second cycle by 8 July 2016. It is also notable in that it identified no
dependencies on Agilisys for a data migration strategy or anything else. As I will
discuss further later, there is a general lack of reference in the contemporaneous
documents of any issues regarding a lack of such a strategy or adequate
templates.
In cross-examination, David Legge accepted that the data migration document he
had created in May 2016 was a combination substantially of material taken from
the SDDs joined together particularly with some content written by himself about
anonymisation: see page 27, lines 1-7 of the transcript.
Page 68 ⇓
68
[220]       In June 2016: this period I observe is notable in the evidence for the, almost,
complete absence of any evidence of CGI doing anything in relation to data migration. The
evidence tends to support the view that CGI struggled to obtain resource to replace
David Legge after he left and nothing much seems to have happened in this period: see
Ravinder Johal’s WS at para 4.15.6. It is not until Graeme King’s involvement in mid-July
2016 that there is anything obviously concrete that happens.
[221]       Between July and August 2016 data extraction was placed in special measures. In
summary, the sequence of events established in evidence was that, by the time David Legge
left at the end of May 2016, not even initial extracts had been provided to Agilisys and not
all of the systems had even been able to be accessed. Almost nothing appears to have
happened in the immediate period after David Legge left. By the beginning of August 2016,
some initial extracts for some of the templates had been provided and were in the process of
being validated. But there were still 9 templates for which absolutely no data extracts had
been provided. By September 2016, as the Programme Highlight Report for 6 September
2016 clearly demonstrates: JB 203, CGI had still only managed to provide usable extracts in
relation to a very small number of templates. Indeed, the final daily report on data
migration dated 29 September 2016, only showed 2 of the 30 templates as having been
completed. Extraction was not even completed by the time the Subcontract was terminated
in March 2017.
[222]       As set out earlier RN007 was based on a failure to provide extracts by either the
original deadline agreed between the parties in version 9 of the Project Plan of 22 December
2015, or by the later deadline in version 12 of the Project Plan of 3 February 2016.
David Legge gave clear evidence that he had not even been able to access the data in the
legacy systems by 3 February 2016 and in fact had only gained access to one of them from
Page 69 ⇓
69
BT by 4 March 2016: see paras 2.22 2.25 of David Legge’s witness statement. In the
circumstances, I agree with the submission made by Mr Cormack, it appears that RN007 was
factually well founded and CGI had materially breached its responsibilities on data
extraction.
[223]       Thereafter RN016 was based on the continuing failure to provide extracts by the
revised deadlines that had been agreed between the parties of 17 August 2016 in version 13,
and 29 September 2016 in version 16.6 of the Project Plans. These failures appear to be
vouched by the evidence. These are the deadlines in the relevant Project Plans. They were
the deadlines agreed between the parties in the context of the 22 April Conditions and the
re-plan and relief that was granted at that time. Again, it is clear that these deadlines were
not met. Data Migration was described as being in “special measures” throughout August
and September 2016: see for example JB 104 and JB 203. As summarised above, it is clear
that complete extracts had not been provided to Agilisys by the end of September 2016.
[224]       Mr Sandison made four distinct points regarding breaches by Agilisys in the
performance of the Subcontract in respect to data extraction:
No adequate data migration strategy had been provided by Agilisys;
The templates provided by Agilisys were inadequate;
Failure in leadership/management by Agilisys first in respect of a resource
provided by CGI, principally Mr Legge and secondly a failure to manage
CEC/BT.
Failure to produce a change management strategy.
I will consider each of these in turn.
[225]       It was not a contentious issue that one of the responsibilities on the part of Agilisys
was that a data migration strategy would be generated (see: section C of schedule 4.1).
Page 70 ⇓
70
[226]       The first question which therefore arises is this: was such a data migration strategy
produced? Secondly, was it produced timeously? Thirdly, was it adequate?
[227]       The position of Agilisys in respect to the above questions was that they should be
answered yes, on the other hand CGI’s position was that the answer to these questions was
no.
[228]       Agilisys’ starting point was that the solution design document JB 234-241 contained
the data migration strategy and that this was adequate.
[229]       I observe that at least initially there appeared on behalf of CGI to be a recognition
that this was a data migration strategy and that it was adequate in that it signed off
Milestone 2.1 JB 241. The purpose of it signing off this document was to indicate that
Agilisys had met the milestone for design of the system including the production of a data
migration strategy.
[230]       Mr Sheriffs was asked a number of questions in respect to the signing off between
pages 99 and 108 of the transcript of his evidence. He was the person within CGI who
actually signed off Milestone 2.1.
[231]       Mr Sheriffs explained in cross-examination why he had signed off this milestone,
when it was his position in evidence, that what was produced was inadequate and said this:
“These were produced following workshops between Agilisys and CEC. They were
presented for sign off by CGI along with client. This was done to a very compressed
timescale. I took Agilisys at their word that they were appropriate, and as CEC
approved them I did so at the same time.” (see: p101)
[232]       The test success criteria for certification was this: CGI and Agilisys confirmed that
agreed deliverables had been completed and approved by the Council as required.”
Mr Sandison’s position regarding the certification was this, that the certification by CGI
having regard to the whole circumstances did not amount to much and little or nothing as
regards the adequacy of the data migration strategy could be taken from it.
Page 71 ⇓
71
[233]       I do not accept Mr Sandison’s submission as to what can be taken from the signing
off by CGI. In terms of the structure of the contract Milestones and certification were an
important part of the mechanism for the operation of the Subcontract and the reaching of
Milestones had material consequences. It appeared to me to be a matter which is unlikely to
have been given as little consideration as Mr Sheriffs in his evidence sought to put forward.
I do not accept Mr Sheriffs’ evidence regarding this matter.
[234]       This evidence of Mr Sheriffs appeared to me to be an example of the rewriting of
history by CGI, namely: taking up one position at the time of the performance of the
Subcontract and taking up a different position later for the purposes of this action. I am
persuaded that the certification is an early indicator of the adequacy of the data migration
strategy produced by Agilisys.
[235]       That the data migration strategy produced was adequate I believe is supported by
other contemporaneous evidence, see: Mr Legge’s email JB 716 and the lack of any adverse
comment about the data migration strategy in the first report prepared by Mr Fulda. If there
was a failure to provide at the outset an adequate data migration strategy I believe it would
have been highlighted in this report given the nature and purpose in preparing it. I would
also note the email of Dr Speed of 23 February 2016 expressly accepting that the data
migration strategy was covered in the Solution Design Document (“SDD”).
[236]       I am unable to identify any other contemporaneous documents until late
September/October on behalf of CGI raising the issue of the lack of an adequate data
migration strategy with Agilisys.
[237]       In respect to the data migration strategy which was produced at the outset by
Agilisys it is I believe important to note the following part of the written submissions on the
part of Agilisys:
Page 72 ⇓
72
Further, to borrow Mr. Coyne’s language, these were documents produced
following ‘a series of meetings with the Authority, in order to gain an understanding of the
current business processes (on the legacy systems) and to agree how the processes would
operate on the new system’: see para. 4.13 in Mr. Coyne’s report. In particular, they
followed workshops with all of the key members of the Council on 6 October and 20
October 2015. They were then distilled into the draft SDD, which were the subject of
checks by various people at the Council. They set-out in copious detail: how certain
data was to be displayed in the new system; to the extent that it was known, areas
where that might require changes and cleansing of data in the legacy system; and a
record of decisions that had been taken about the overall design of the system. These
were all correlated to the specific requirements set out in the OBS. The data
migration strategy then sets out: (a) the process to be following in extracting,
transforming, and loading the data; (b) the level of data to be migrated to the live
and historic clients; and (c) a record of the various design decisions made about the
level of live and historic data that is to be migrated. Although hearsay evidence was
led by CGI that almost a year later, the CEC representatives did not understand and
were still querying the process. No actual evidence was led from CEC. No evidence
was led from anyone at those meetings. No cross-examination of Agilisys’ witnesses
was undertaken to suggest that those meetings had not fully covered the matters and
decisions recorded in the SDD.” (see: 5.6.5)
[238]       I believe that this is a very powerful submission clearly supporting the adequacy of
the data migration strategy produced in the SDDs and undermining the criticisms made on
behalf of CGI regarding this documentation.
[239]       CGI in the course of its evidence tried to develop a position through the evidence of
Mr Waters, Mr Brown and Mr Sheriffs that it was really only where Kevin Sander produced
a document on 23 November 2016 that anything approaching an adequate data migration
strategy was produced by Agilisys.
[240]       In respect to this issue I prefer the evidence of Mr Sander. I accepted the evidence
contained in his statement at paragraph 6.2.3 that all he had done in preparing this
document was review and consolidate what was already there in respect of a data migration
strategy. I also believe it is noteworthy in this context what is said by the witness at 6.4
and 6.5 of his statement which was as follows:
“Given the passage of time and change in stakeholders and general level of
frustration within the Council, I wrote a detailed Data Migration Strategy document
that ran to 100 plus pages. To do so, I copied information in from both the SDDs and
Page 73 ⇓
73
the data migration templates, reviewed each section with relevant stakeholders from
both the Council and CGI to ensure that we jointly captured any questions or
concerns they wanted to raise. I also reviewed the options paper which had been
previously produced by Julia Jackson to ensure that my document was consistent
with the recommendation made there. I then issued several iterations of the draft
document which progressively added in any changes or additional clarifications
required in order to respond to and close out the questions and concerns which we
had jointly captured.
6.5 However, we have worked perfectly well on other similar projects without this
consolidation and reformulation of information. In my view it became desirable in
this case only because by October 2016 the relationship between Agilisys, CGI and
the Council was perceived by the Council to be working so badly, and because so
much of the earlier discussion and knowledge transfer had been eroded by the
passage of time coupled with high turnover of CGI extraction personnel. My
motivation for writing a new document was only to restore confidence with Council
stakeholders. I perceived that consolidating all the already available information
into a single document and using that as a basis for review was the quickest and
most efficient way to achieve that.”
[241]       These passages reinforce his position as to what the document he produced
amounted to and that in particular it was not an adequate data migration strategy being
produced for the first time. The reasons he gives for the production of this document appear
to me cogent and explain why this document was produced, not because there was not
already an adequate data migration strategy but because of other problems which had
occurred in the implementation of the project and which had in particular been caused by
the changes in personnel within CGI. It was clear from the evidence that there had been
significant changes in CGI personnel in respect to data migration (see: Dr Hunt’s initial
report at 8.7.1).
[242]       Four detailed points were made by Mr Sandison in relation to the position of
Mr Sander which were as follows:
(a) In comparison to the existing sources of information which Mr Sander used,
Mr Sander noted that I probably documented a little more explanation and
Page 74 ⇓
74
context around some of the information. This included Mr Sander’s own
experience and knowledge of data migration.
(b) The strategy document brought together previously disparate sources of
information, including the templates and SDDs. These documents were, in
Mr Sander’s own description, not sufficient without the additional
(non-documented) “knowledge transfer” said to have been given to
David Legge at the outset of the project.
(c) Production of the document, including the investigative process that preceded
it, took around one month.
(d) The process of producing the document was a process that I went through
with the council workstream leads, and was intended to reflect the feedback
from that process in addition to any additional questionsthat had been
raised. Further versions of the document were produced which incorporated
further changes and revisions.
[243]       Point (a) does not suggest that there was anything significant added to what had
already been provided.
[244]       Point (b) emphasises that Mr Sander was bringing together information from a
number of sources (already available) and adds nothing further. I do not understand that
there is evidence which supports a lack of knowledge transfer to Mr Legge. When I look at
his evidence as a whole he appears generally happy with the information that was provided
by Agilisys.
[245]       Point (c) stresses the length of time it took Mr Sander to prepare this document,
however, Mr Sander at 6.4 and 6.5 of his statement gives an explanation as to why in the
context of the problems with the Subcontract it would have taken that long to produce this
Page 75 ⇓
75
further document. It does not suggest that anything significant had been added in this
document to what was already available or what had been provided to that point was
inadequate.
[246]       So far as the point made at paragraph (d) this is dealing with no more than feedback,
I think that would be expected in an iterative process. The mere fact that feedback had to be
taken account of does not mean that the original data migration strategy was in any sense
inadequate. All that can be taken from his taking account of feedback is this: as with any
document prepared issues may arise which cause it to be altered, that however does not
necessarily suggest the document was inadequate, it is merely a process of refining.
[247]       Further in respect to the issue of the adequacy of the data migration strategy there is
the evidence of Dr Hunt which she sets out as part of her consideration of data migration in
section 8 of her initial report. She considers this issue in some detail including consideration
of the general approach to data migration by Agilisys in section 8.4 and in particular
considers the adequacy of the data migration strategy under a number of headings and in
respect to certain specific criticism made on behalf of CGI regarding the strategy. Her views
are I think well reasoned and based on a full analysis of the relevant circumstances and
clearly support the view that the data migration strategy produced was adequate. I believe
having regard to her approach and reasoning that considerable weight can be attached to it.
I prefer her views on this matter to those given by Mr Coyne. I have already set out my
reasons for rejecting the evidence of Mr Coyne. I preferred the evidence of Dr Hunt to that
of Mr Coyne on this issue. I rejected his evidence for the reasons set out.
[248]       Overall I am persuaded that Agilisys produced at the outset (therefore timeously) an
adequate data migration strategy document.
Page 76 ⇓
76
[249]       Even, if contrary to my view the strategy document was not adequate and timeous I
am not persuaded that any such failure had causative substance. The evidence led by CGI
in respect to this was in my view sparse and not persuasive.
[250]       I believe an appropriate place to start consideration of this issue is the evidence of
Mr Legge. His position regarding a data migration strategy is given at paragraph 2.8 of his
statement:
“That (a data migration strategy) was not hugely important to me since my job was
to extract the data and the strategy would only come into play after that. The data
migration process can be described as extract, transform and load, and I was only
responsible for the extract part. I understood that Agilisys would subsequently
provide details of which data they wanted and the format of that, since that wasn’t
set out in the SDDs.”
[251]       The above answer I consider to be informative on this issue. Mr Legge was the man
on the ground initially doing the data extraction and the issue of the existence or otherwise
of a data migration strategy is of no importance to him. It is implicit in that answer that the
absence of such was not causative of any delay in the data extraction process. He was the
person on the ground for much of the project carrying out the work and it appears that he
was not held up by the lack of such a strategy.
[252]       Turning to Stephen Waters, his position regarding a lack of a data migration strategy
is given at 2.5 of his statement where he says:
“We didn’t have a data migration strategy from Agilisys when I started in April
2016. That was something we knew we needed, but as we didn’t have one no one
was clear on what the scope of work was. …there was no data migration strategy
produced by Agilisys until Kevin Sander produced one on 23rd November 2016.”
[253]       However, as to how a lack of such a strategy document impacted on the delivery of
the project and caused delay I am unable to identify anything further in his evidence
regarding this.
Page 77 ⇓
77
[254]       It is also clear that he implicitly accepts that Kevin Sander produced an adequate
data migration strategy and if all he was doing was consolidating (as I have earlier held)
then from the outset there had been an adequate strategy.
[255]       As regards the evidence of Colin Brown, he refers at paragraph 3.91 of his statement
to an email from Bridget Fox and at 3.91.4 he goes on to say this:
“Agilisys had not provided a data migration strategy (only a high level version
included in design documents) or a plan to engage the business in the data migration
stream. As such CGI was unaware of what Agilisys’ process was regarding ensuring
data integrity.”
[256]       His position was the same as Mr Waters on Mr Sander producing a data migration
strategy in November: “Agilisys started noticeably to improve” (see: 3.98). Again there is
an implicit acceptance that if what Mr Sander produced was a consolidation there was from
the outset an adequate data migration strategy.
[257]       This witness sets out what he regards as the critical importance of a strategy
document at page 79 of the transcript of his evidence:
“So strategy, get a good strategy, then you build your plan out, so this really should
have maybe fed in to that whole process of getting a good strategy and then getting
the plan, and I think that’s why we find data migration failing, because there’s not a
solid foundation to start building these plans off…”
[258]       What the witness makes clear in this answer is that from the outset a strategy needed
to be in place.
[259]       In light of the position which he took up about the necessity for this strategy
document to be there from the outset, at page 146 of the transcript he is asked this about a
lessons learned exercise put in place by him in August 2016:
“And if there had been problems with SDDs, if there had been problems with SDDs,
like outstanding information, again, that would have been called out by your team.
It would be a shocking state of affairs having spent a month intensively reviewing
the data migration workstream for CGI not to have realised if it was the case that
there was no data migration strategy?”
Page 78 ⇓
78
[260]       He answered the above question in this way:
“I think it is a shocking state of affairs there is no data migration strategy. I think
you’re absolutely right. As I say, we tried to work within the approach that was
suggested and we couldn’t impact at that time, but when the wheels do fall off
something like this, you do have to take it back and say ‘Okay, that hasn’t worked.
What do we need to ensure this works? What should we have done?’ Lessons
learned, it’s back to that, and I believe what should have been done was when
Kevin Sander came in and managed that stream correctly.”
[261]       What in the answer he does not give any proper explanation for is this: why, if this
matter was of such importance, was it not identified at the end of August following the
month of intense activity reviewing the data migration workstream, that there had been a
lack of a strategy document from the outset.
[262]       This appears to be an area where the contemporaneous documentation does not
support the position being advanced at proof by CGI. If a strategy document should have
been supplied at the outset, if it was not supplied, and if it was of critical importance as this
and other CGI witnesses seemed to suggest, I believe there would have been from the outset
documentation emanating from CGI and sent to Agilisys calling this out. In addition there
would have been a consistent pattern of contemporaneous documentation calling this out
with ever greater urgency. I do not believe there is such documentation. This lack of
contemporaneous documentation emanating from CGI calling out matters which it now
says were of importance is I believe a theme, across many of the issues and strongly
supports the view that there was no breach of obligations on the part of Agilisys which
caused or affected a breach by CGI of its obligations.
[263]       The next witness who gives evidence in respect of this issue is Mr Tom Fulda.
Mr Fulda prepared two audit reports which in some detail considered the implementation
of in particular the ERP project. Mr Fulda in his second report under project methodology at
Page 79 ⇓
79
5.2 identified as a criticism of Agilisys a lack of data migration strategy, although there is no
particular development of this point. This report was prepared in November 2016.
[264]       It is noteworthy that in his first report prepared in February 2016 he does not
recommend any form of rewriting of the design documents and does not note that there is a
requirement for a data migration strategy to be in place.
[265]       Given that it was, according to CGI witnesses, necessary from the outset, that an
adequate data migration strategy was in place and this was a significant issue, I would have
expected it to be identified by Mr Fulda in his first report. At pages 92 and 93 of the
transcript of his evidence he accepts no such issue was identified.
[266]       For reasons I have already set out, I was not prepared to accept the evidence of
Mr Fulda arising from his second report.
[267]       Ms Fox, Mr Graham and Mr Sheriffs it appears to me say nothing significant relative
to a lack of a strategy document.
[268]       I believe given the whole evidence that this issue of a lack of a data migration
strategy was sought by CGI at a later stage to be turned into an issue of significance, which it
was not during the course of implementation of the project. This I believe supports a
conclusion that Agilisys fulfilled its obligations in respect of the production timeously of an
adequate migration strategy and that in any event any lack of such a strategy document was
not causative of delay in the performance of CGI’s obligations. Beyond that any evidence of
actual causative effect of such a lack is vague and inspecific.
[269]       Thus I conclude for the foregoing reasons that a data migration strategy was
produced at the outset and that it was an adequate document. Even if I am wrong in that
view, it was not causative of any significant delay in the performance of CGI’s obligations.
Page 80 ⇓
80
[270]       The next issue is whether Agilisys failed to provide adequate templates for the data
extraction process in the sense that they were not fit for their purpose.
[271]       The data extraction process involved the use of templates. Specification of the
required extract is provided by the template. The template is typically a spreadsheet which
contains a sample worksheet with the required columns for the data extract identified. It
was a matter of agreement that these templates were to be prepared by Agilisys.
[272]       It was accepted at all hands that the data extraction process was an iterative one.
Thus templates would be used and what was recovered from the legacy system would be
examined and the templates amended and refined in order to obtain a better extract of data.
[273]       From October 2015 to June 2016 the person who was carrying out the extraction of
data from the legacy system was David Legge.
[274]       Looking at his witness statement at paragraph 2.34 he says that on 22 April 2016 he
obtained a set of updated templates from Agilisys. The only comment he makes about the
adequacy of these was that certain further templates were received in May (see: 2.35). He
does not raise any significant issues relative to the adequacy of the templates.
[275]       With respect to the new templates, provided in May, he made, what was clear from
his cross-examination a minor criticism about changing file names and the need for change
control (see: 2.42 of his statement).
[276]       In cross-examination he dealt with the position regarding the first set of templates
sent to him by Agilisys on 4 December 2015. He said this in an email:
“Just had a look through these, I don’t see anything untoward at this stage.” (see:
page 33 in the transcript of his evidence)
[277]       As at 21 January 2016 he describes a request that he is receiving from Agilisys as
being “the kind of routine thing which takes place in a data extraction process”.
Page 81 ⇓
81
[278]       At page 63 of the transcript he makes it clear that the one time he sought assistance
from Agilisys, it sought to provide assistance.
[279]       So far as the issue of any failures to manage him by Agilisys he makes a very general
criticism of this at paragraph 2.2 of his statement. This related to him being left to
communicate with BT and CEC without any assistance from Agilisys. However, when the
evidence which he gives in cross-examination at pages 28 to 31 of the transcript is
considered, he cannot identify anything Agilisys could have done by way of management
regarding the problems with CEC and BT over access by him to the legacy system which
arose at least in part from security concerns of CEC. In addition it is quite clear he knew
what he was doing relative to these matters and in relation in particular to BT he was
following a set process. The difficulties he was having in obtaining material from BT and
CEC did not arise through any failure in management by Agilisys of Mr Legge. The issue
was not one of management. It was an issue requiring CGI to procure that CEC complied
with its responsibilities, rather than simply not dealing with requests in an adequate and
timeous manner. Moreover, so far as being under the management and control of CGI
Mr Legge said this at page 55 of his cross-examination:
“Someone at CGI, and I can’t remember who it was, had said ‘Just deal directly with
Peter. Let him know how you’re getting on’. So reporting in this context does just
mean ‘Here’s what I’m doing. Here’s how I’m getting on.’ Doesn’t mean I do what
Peter specifically asks.”
[280]       Overall I do not think his evidence could be characterised as critical of the adequacy
of the templates. His evidence on this issue is important as he was the man on the ground
carrying out the extraction of data. Nor do I think any failure in management of Mr Legge
by Agilisys can be identified in his evidence. Beyond that I cannot see any causative
substance so far as delay is concerned which arises due to lack of management by Agilisys of
Mr Legge.
Page 82 ⇓
82
[281]       The next witness to work, hands on, in relation to the templates was Stephen Waters
who worked on this aspect of the project from September 2016 onwards (WS 1.4).
[282]       Mr Waters makes certain criticism of the adequacy of the templates at paragraph 2.7
of his statement. However, his first email after the receipt of these templates on 22 April
does not raise these issues (see 2.8 of WS). Nor does his email to Peter Hall on 26 April
referred to at paragraph 2.15 of his statement. If there was a significant problem with the
templates I would expect it to be raised in these emails. What he does say in the email of
26 April is that he is still awaiting “confirmation from the client that CGI can use live data”.
[283]       He refers at 2.31 of his statement to a document prepared by a Mr King (about
18 July 2016) in respect to the templates. Mr King was not led in evidence. I observe that in
his email Mr King says this: “Concern that Agilisys is being paid to do the data migration
and yet we are doing all the investigation work…” This takes no account of the fact that the
responsibility for data extraction was CGI’s and this misunderstanding may colour his
views. Mr Waters was unable to give any detailed evidence relative to the position
regarding templates in July/August 2016 as he was not directly involved with the ERP
project, only returning to it in September 2016 (see: paragraph 2.32).
[284]       Mr Waters between pages 81 and 85 of the transcript of his evidence is asked about
the adequacy of the templates, in the context of their being used in an iterative process. He
eventually is asked this question at page 85:
Q. Yes. If that’s right, that some extracts were being added to the SharePoint site
at this point, then would that not indicate that the Agilisys templates were
functioning satisfactorily at least in terms of delivering a first extract or first cut
of data?
A: Yes.
This was in relation to matters as at August 2016.
Page 83 ⇓
83
[285]       Data extraction, as was accepted at all hands, is iterative, so that after a first cut the
templates could be refined. This was the whole basis upon which the template solution
worked. It is clear from the above answer that the use of the templates was producing an
extract which could later be refined.
[286]       At 2.35 of his initial statement he says this:
“Around that time [September 2016] a serious issue arose regarding ‘relations’ not
being defined. The templates from Agilisys had a field for employee relations.
There were no further elements for relations within that, such as if an employee was
in a car leasing system, or had a season ticket for travel etc. and all of those elements
had to go into Business World. CGI had incorrectly assumed that Agilisys knew
what they were doing.”
[287]       He is then asked about this issue in cross-examination at page 93 of the transcript.
He is then referred to an email from Ms Johal in JB 1528 which says:
“Here are the valid relations for employee/pursuer employment.”
This email was written on 5 August. He accepts at page 95 that these “look like relations”.
[288]       Thereafter at pages 96/97 in the transcript it becomes apparent that the particular
examples of employee relations which he raised as a difficulty had in fact been covered in
the document earlier produced by Ms Johal.
[289]       The next person from CGI who had involvement in this issue was Mr Tom Fulda.
[290]       His principal involvement in the ERP project was the preparation of what were
described as two audit reports. The first report was dated 13 February 2016 (JB 808).
[291]       Mr Fulda in his statement at paragraph 2.5 sets out in full his recommendations
arising from this audit report. I do not read any of these recommendations as making any
implicit or explicit criticisms of the templates which had been produced by Agilisys.
[292]       He quotes from his report at paragraph 2.4 of his statement and says that the three
largest areas of concern were:
Page 84 ⇓
84
“Interfaces and data migration (large delays, multiple parties involved, too many
hand-offs, now an issue rather than a risk) and customer resources (multiple impacts,
see customer resources below).”
In this section I do not see anything about problems with the quality of templates being
identified. Nor do I see anything about a lack of a data migration strategy or failures to
manage by Agilisys.
[293]       The entire findings section of Mr Fulda’s report is in the following terms:
Findings
There are many risks, many of which are significant, but broadly affect just two
outcomes:
o Those that jeopardise go-live date
o Those that jeopardise solution adoption/benefits realisation and impact post
go-live hypercare (Well documented examples of local authorities in this
situation exist, e.g. Birmingham CC backlog in supplier payments. Extensive
bad publicity.)
The second outcome is more important than the first. A technical go-live (on
time delivery of tested solution that meets the specification) into a business that is
not prepared for it has both immediate and longer term costs: hypercare, repeat
training, delays to processing which can also be felt externally (delayed
payments/bad publicity), user engagement and morale/system credibility,
proliferation of poor practices/off-system working, delayed benefits-realisation,
time consuming and costly to turn around, etc.
Current version of the plan (v12) is coherent but there is now minimal
contingency. I cannot see a way of reshaping the plan to introduce more
contingency whilst retaining the same go-live date or introducing new risks.
A little contingency remains in the short period after the second payroll parallel
run (by increasing backfilling CEC training resource with Agilisys/CGI) but
hitting dates for parallel run is critical or a month is lost. Activities on critical
path leading up to that are tight and many are subject to significant risks (see
Appendix)
3 largest areas of concern are around interfaces and data migration (large delays,
multiple parties involved, too many hand-offs, now an issue rather than a risk)
and customer resources (multiple impacts, see Customer Resources below)
Taken individually, most risks can be mitigated to a greater or less extent (though
surely time is running out for data and interfaces) but taken collectively the
challenge is great.
On its own, extending the project timeline is likely to have limited impact on the
outcomes in the first point; it must be accompanied by other changes (see
Customer Resources below and Recommendations)
Although signed off, I am concerned about the degree of understanding of the
new solution and the resultant impacts. Documentation is extensive but has it
been understood? Insufficient/no ongoing involvement from senior business
users/process owners. Insufficient involvement of CEC project team members
Page 85 ⇓
85
with the project activities. Agilisys activities progress on best endeavours basis
(consultant I briefly met appeared diligent) but they are rather isolated from
customer scrutiny.
Fully expect Agilisys to deliver to spec (they are delivering their template
solution to a large degree) but a gap between this and customer
expectations/understanding is to be expected. It normally materialises
throughout the build phase but given relative hands off approach from customer
it is likely to materialise at UAT at earliest (adversely impacting UAT timeline)
and quite possibly after go live.
Test strategy document is rather loose, for example, SAT acceptance criteria
‘There should be no critical defects, and the number of major/minor defects
should be reasonable’
Risks and Issues: not comprehensively captured. Risks not updated in timely
fashion. Agilisys don’t have access to iTools and PMO input process appears to
have lapsed? Didn’t find an issue log (?)
Project documentation: appeared difficult to access Agilisys documentation. All
of the following documentation must be jointly and readily accessible: process
design, functional specs, technical specs, configuration notes, access
authorisations for roles, test scripts, end-user procedures, cutover planning.”
[294]       When the whole findings are looked at there does not appear to be any criticisms of
Agilisys templates or any other significant aspect of its implementation of this part of the
project.
[295]       In particular it appears to me to be noteworthy that the only specific comment on
templates is as follows:
“Fully expect Agilisys to deliver to spec (they are delivering their template solution
to a large degree)…”
[296]       On a fair reading this suggests that the templates part of the process is proceeding
properly and that there is no difficulty regarding the adequacy of the templates being
produced.
[297]       At page 93 of the transcript of his evidence he accepts that there is no
recommendation in his report regarding templates and their being altered in any way. Had
there been any material difficulties with the templates at that time I would have expected
this to have been raised within his report.
Page 86 ⇓
86
[298]       Against that whole background I find this comment at paragraph 3.2.1 in Mr Fulda’s
statement regarding data migration: “The quality of Agilisys material in this area was poor”
when speaking about the time of his first report to be entirely at odds with his findings
which he made at the time. This is perhaps another example of CGI rewriting. It
undermines the general acceptability of this witness’s evidence.
[299]       Mr Fulda did a further review in November 2016. In this review he made fairly
trenchant observations regarding the template solution (see 5.3.4 of his statement).
[300]       There is thus a seismic change in his view between February and November. In
section 5.3 of his supplementary audit report there is a root and branch criticism of the
approach and methodology of Agilisys in respect to data migration. It appears to me that if
these criticisms are justified they could equally have been made at the time of the February
review. The approach and methodology of Agilisys had not, as I understand it from the
evidence, changed during the period February to November. I therefore cannot understand
why if all of these criticisms are justified they are not set out in the report of his February
review. Moreover they are entirely at odds with Dr Hunt’s views regarding Agilisys’
approach at 2.5.2, 2.5.3; 8.4.1 to 8.4.5; 8.5.2 to 8.5.4 and 8.11.1 to 8.11.2. I find her analysis in
relation to this aspect of the case convincing and I believe it strongly supports the adequacy
of the templates.
[301]       I think the difference between the February and November report, at least in part,
relates to the context in which the two reports were prepared. At the time of the February
report there was no commercial dispute between the parties. By November there was quite
clearly such a dispute and Mr Fulda was aware of this. As Mr Fulda makes clear, at page 95
of the transcript of his evidence, by November he is concentrating on those elements which
have a commercial impact. As I understand it the question that he is considering is the
Page 87 ⇓
87
commercial impact on CGI in the context of its dispute with Agilisys. I believe to an extent
this context has affected the nature and degree of criticism of Agilisys.
[302]       At pages 113 and 114 of his transcript of evidence, Mr Fulda’s attention is drawn to
the difference between the terms of his report and the evidence of Mr Legge regarding the
adequacy of the templates and he was asked who was in the better position to judge. I
believe that the man on the ground doing the extraction process, namely Mr Legge, is the
best person to speak on this issue and this again casts a shadow over the opinion of
Mr Fulda which he expresses in his report.
[303]       Finally between pages 100 and 109 of the transcript of his evidence Mr Fulda is asked
about his conclusions relative to “Frontier”. It is clear from this passage of evidence that he
has failed to take account of evidence and that his conclusion is erroneous.
[304]       I would restate my position regarding Mr Fulda, I was not prepared to place any
reliance on his evidence except when independently supported.
[305]       Other witnesses led on behalf of CGI gave evidence regarding the adequacy of the
templates. First Andy Cleaver who says that the templates were inadequate (see:
paragraph 3 of his statement). He also says in the same paragraph that he had little
involvement with the ERP project. I take absolutely nothing from this wholly general and
inspecific evidence given on the basis of little involvement in the project.
[306]       Niall Sheriffs also gives evidence on this issue. At 3.54 of his statement he says the
following:
“Agilisys delivered some templates in December 2015 for data migration … I was
informed that they contained a list of field names but no field sizes, permitted values
or validation rules. It seems from what I was told that Agilisys reused templates
created for an earlier ERP client of theirs, Bristol City Council, as there were still
references to Avon Fire and Rescue settings within the templates.”
Page 88 ⇓
88
[307]       He does not say who informed him of these particular difficulties with the templates.
However, I observe that these difficulties have not at that stage been identified by Mr Legge
or Mr Fulda.
[308]       At paragraph 3.90 Mr Sheriffs sets out the findings section of Mr Fulda’s report of
13 February 2016, which I have already set out in full in this opinion. Mr Sheriffs then says
this at paragraph 3.91:
“This illustrates that there were many issues with Agilisys’ approach at that time.”
[309]       It does not appear to me that this comment is in any way justified by what is said
within the findings section. The findings section simply does not support that conclusion.
For reasons I have already set out I do not believe that the section can be said to be
supportive of criticisms of Agilisys.
[310]       The two sections of Mr Sheriffs evidence to which I have above referred are I believe
difficult to explain if what we have is a witness who is seeking to give evidence in a fair and
balanced way.
[311]       At 3.145 of his statement he sets out a workstream highlight report dated 4 April
2016 relative to data migration. There is nothing said about lack of a data migration strategy
nor is there anything about the inadequacy of templates contained within this document.
Nor is there any reference to some lack of management on the part of Agilisys as a problem
that is being encountered and which is causing delays. If these were real issues a document
such as this would have been an obvious place to expect to see such issues highlighted.
[312]       At paragraph 3.149, when dealing with the situation as at April 2016, Mr Sheriffs
makes further criticism of templates, repeating in large part his position stated at
paragraph 3.54 about which I have already commented.
Page 89 ⇓
89
[313]       This position which he gives as at April 2016 does not seem to fit with Mr Legge’s
position (see: 2.34 and 2.35 of Mr Legge’s statement).
[314]       His position in addition does not appear to fit in with the initial response to RN005 to
RN009 which he deals with at paragraph 3.150 of his statement. In respect to data extracts
he says this:
“This has been delayed due to BT, and now that the servers are in our control is
being scheduled.”
He further says:
“Without attributing fault CGI recognises that the replan is required due to
circumstances not under the direct responsibility of Agilisys and therefore relief from
any liquidated damages should be provided as a sole remedy to the Relief Notices
RN005 –RN009”
[315]       If there had been a delay caused by a failure by Agilisys to produce adequate
templates as is suggested at paragraph 3.149 of this witness’s statement or a failure by
Agilisys to manage BT or CEC or Mr Legge, or a failure to produce an adequate data
migration strategy, one would have expected something to be said about it in this document,
however, it is not. Moreover, one would have expected this initial response to RN005 to
RN009 to have been in wholly different terms. This initial response is wholly inconsistent
with the position this witness and CGI generally sought to advance in evidence.
[316]       In the ERP workstream report of 11 April in the last section no criticism is made of
the templates. All the templates Mr Sheriffs was criticising had been supplied on 22 April.
However, again I observe in the Workstream Highlight Report of 25 April 206 no criticism is
made of the templates. The problem which is causing concern is the data anonymisation
issue (see: 3.167 of his statement).
Page 90 ⇓
90
[317]       At 3.168 Mr Sheriffs criticises further templates provided by Agilisys but gives no
details as to the nature of those criticisms and what, if anything, flowed therefrom in terms
of causative effect, simply saying “CGI raised many issues”.
[318]       At paragraph 3.208 of his statement Mr Sheriffs sets out the terms of an internal
report of Graham King dated 18 July 2016, dealing in particular with templates, and data
migration, however, as I have said I heard no evidence from Mr King.
[319]       It is I find of assistance in considering Mr Sheriffs evidence to consider his position
relative to the response to Ms Ah-Wong’s email of 22 April. It would be appropriate at this
stage to set out the terms of this email (to which I have earlier made reference) in full:
[320]       The April conditions to which reference has been made were contained in an email
sent by Louise Ah-Wong of Agilisys to Niall Sheriffs (CGI):
“This note is regarding this week’s special ERP programme board and the decision to
extend the ERP project by two months.
Subject to the points below, I can confirm that Agilisys will not seek to make extra
charges to CGI for the costs incurred by Agilisys due to the delay to the project dates,
and the change to a December 1 Go Live date.
1. That CGI provides Agilisys with a detailed plan for the following areas (within
3 working days) for inclusion into the detailed replan :
o Development, test and Gold Environment provisions including the transfer
from the Agilisys temporary environment to the CGI environments in Wales
and Scotland
Pre-production and production environment delivery and their
proposed move (if any) to Wales
Delivery of practical management procedures for all environments
and their move into service including resolution of access issues
o Biz talk environment support and maintenance activities
o Detailed test plan activities
o The provision of test data and data migration activities under CGI
management
2. That CGI works with Agilisys to produce a joint detailed plan for the Council’s
review (based on the high level plan/critical path attached, this is the same
version that CGI/Agilisys discussed and has been presented to Jimmy Lumsden
the Council’s representative prior to this weeks’ project board meeting).
3. That CGI formally approves the revised ERP plan by the 29th April 2016 for joint
presentation to the Council for approval.
Page 91 ⇓
91
4. That CGI agrees to the CGI Cause in the EI Relief Notice RN003.
I hope you understand that Agilisys is not confident in CGI’s ability to meet the
deadlines on the provision of infrastructure services based on your performance to
date on both the EI and ERP projects. On the EI project this has been the cause of
delay. We need to be confident that CGI recognises this and the importance of
implementing infrastructure to agreed timescales. The new solution/approach that
CGI is proposing for the ERP project is more complicated and higher risk to manage
a responsibility that we expect CGI to manage, communicate clearly to us and
resource accordingly.
Please let me know your response to these conditions.”
[321]       It was not a matter of dispute that these conditions were accepted on behalf of CGI.
In its response to RN 005-9 (the composite response) (JB 30) CGI said this in the part headed
“Response”:
“As a consequence of RN 011 (payroll resource delay), a Go Live Re Plan to the
1st December has been agreed between CGI and the Council. The Council has agreed
to amend the associated ERP milestone dates to align with the replanned Go Live
dates. Therefore there will be no impact on associated LDs in regard to the ERP
project milestone relating to RN 005, RN 006, RN 007, RN 008, RN 009. The Council
has stated that the approval of the delayed implementation is contingent on there
being no additional financial cost to the Council.
CGI has confirmed to the Council that the time delay will be absorbed and the cost of
the CGI implementation will not increase.
Agilisys has confirmed to CGI that the time delay will be absorbed and the cost of
the Agilisys implementation will not increase. This confirmation was subject to four
criteria as defined in an email sent from Louise Ah-Wong to Niall Sheriffs on the
22nd of April 2016. CGI has agreed to meet these four criteria.”
[322]       Mr Sheriffs was a senior member of the management team in CGI relative to this
project by December 2015 (see: 1.3 and 1.4 of his statement).
[323]       If he had formed the view that the templates were inadequate at the stage he is
speaking about at 3.54 of his statement namely about December 2015 that there were
material problems with the templates (see: 3.94 of his statement) and as at February 2016,
that “many issues with Agilisys’ approach at that point” (see: 3.91 of his witness statement),
it does not fit and is not consistent with his acceptance of the points made in Ms Ah-Wong’s
Page 92 ⇓
92
email and in his granting relief at various stages to Agilisys. Those actings are not consistent
with the litany of criticisms of Agilisys this witness now puts forward including criticism of
the templates.
[324]       Mr Sheriffs’ explanation for what had happened at the time in respect to the granting
of relief in terms of Relief Notices and in respect to the response to the Ah-Wong email was
given at paragraph 3.7 of his statement:
“Agilisys were very quick to issue Relief Notices. Earlier in the programme CGI’s
stance was more conciliatory, with a willingness to accept without too much scrutiny
the suggestion that CGI might be the cause of delay, and the acceptance of CGI cause
in some relief notices. However it soon became clear that there was a tendency from
Agilisys to be disingenuously simple in the relief notices, claiming that all delays
were due to CGI when the reality was much more complex. It was a very large
programme with interdependencies between streams.”
[325]       This paragraph encapsulates a theme in the evidence of in particular more senior
members of CGI management who gave evidence. I have already commented to some
extent at an earlier point in my opinion about this supposed attitude of CGI. However, I
think this particular paragraph is illuminating. The point which the court is being asked to
accept is that a large commercial organisation such as CGI did not properly investigate
important matters such as Relief Notices and simply accepted what was said in these notices
“without too much scrutiny”. Such a statement in my view does not stand up to any
examination, far less close examination. The idea that Relief Notices were being accepted
and that an important document such as the Ah-Wong email was agreed without much
scrutiny simply seems to me to be wholly implausible.
[326]       Another matter referred to in this section of evidence is CGI being “conciliatory” and
this is another theme in the evidence of CGI witnesses. It just wanted to move matters
ahead and not to apportion blame. This is said to explain why it did not call out Agilisys on
matters which it now says that in retrospect it should.
Page 93 ⇓
93
[327]       Again, and I have to some extent commented on this earlier, such evidence stands up
to no real scrutiny. The relationship between CGI and Agilisys was a purely commercial
one. Because you want to be conciliatory, take a partnership attitude, make things work and
move things along does in no way stop the raising of, what are now said to be major issues,
with the other party. It may affect the language used when the issue is raised. It does not,
however, stop the issue itself being raised.
[328]       There is also a suggestion in the above paragraph of Mr Sheriffs’ statement that
Agilisys in some sense were not acting properly in their presentation of the Relief Notices,
they were quick to issue these and were disingenuously simple in the notices that they put
forth. Agilisys were doing no more when issuing Relief Notices than operating the agreed
contractual procedure. This evidence about Agilisys in some way acting improperly again
was a recurring issue in the evidence of members of CGI’s staff. I found no evidence in the
case to back up the suggestions made on behalf of CGI that in some way Agilisys had acted
in any way improperly in respect to the issue of Relief Notices.
[329]       I have dealt at some length with the evidence of Mr Sheriffs. I have done so in that
he was a witness, as I have already stated, who I found very unsatisfactory.
[330]       I turn now from Mr Sheriffs to Mr Colin Brown who also gave evidence on this issue
and was critical of the templates provided by Agilisys.
[331]       However, at page 123 of the transcript of his evidence he says this in regard to the
situation in August 2016 in respect to the templates:
Q. Aside from the point he [Mr King] makes about queries, it’s fair to say that he’s
not envisaging any particular difficulty in continuing on with the data
migration templates as is. That would be fair, isn’t it?
A. I think at this point he’s obviously read through the document and he’s given
his initial assessment, so at this point, no, he’s not raised that.”
Page 94 ⇓
94
[332]       One of the key criticisms that this witness makes about templates as is made by other
witnesses on behalf of CGI is the lack of specification of relations, however, the witness
appears to accept at page 129 of the transcript that this had been dealt with by Agilisys at
5 August 2016.
[333]       At page 137 he accepts that the first revised extracts have been produced by
30 August 2016. He is then referred to an email at page 139 which says this:
“There are no issues at present that will take us beyond the end of September for
delivering all the templates and as per an iterative development there are some
changes within the individual sprints, all noted within the report.”
[334]       He then accepts that an iterative process was being followed and that this involved
the checking for errors and then you keep coming back and go round in that circle”. He
then accepted at page 140 that during August 2016 a full and thorough review of the data
migration workstream had been carried out.
[335]       There is then a long section of evidence in cross-examination in which the witness is
asked: why after such a review, if there were any problems with the adequacy of the
templates, they were not called out at that stage, (see: pages 140 to 144). The answers the
witness gives I believe are rambling and do not answer the fundamental question, ie: why,
if they existed, were these inadequacies not called out by CGI at this point?
[336]       Lastly there was the expert evidence, as I understand it Dr Hunt did not have any
criticisms of the approach of Agilisys to data migration nor in respect to the adequacy of the
templates and I have set out the critical parts of her report earlier. I rejected the evidence of
Mr Coyne on this aspect of the case for the reasons I have set out.
[337]       For the reasons which I have referred to I do not think that the evidence supports
that there was any real issue with the adequacy of the templates. When looked at as a whole
Page 95 ⇓
95
the evidence supports the adequacy of the templates. It does not support the CGI position
that the templates were inadequate and in particular not fit for purpose.
[338]       The next issue is this: was there some causative breach in leadership or project
management of the ERP project in relation to the issue of data migration.
[339]       I do not believe that the issue of management properly arises in the context of data
migration. The two principal obligations on Agilisys were to produce a data migration
strategy and to produce adequate templates. I have analysed the situation in terms of those
two specific obligations and for the reasons I have given above I am persuaded that Agilisys
fulfilled its obligations. So far as management of Mr Legge is concerned for the reasons I
have earlier set out, there is no breach of any obligation. Turning to the issue of
management of CEC/BT I am not persuaded there was a breach of any such duty. It was
explicitly stated that it was for CGI to manage BT (see: OBS 1, 2, 3 and 30). I cannot see how
Agilisys have failed in some leadership role regarding CEC. It was the responsibility of CGI
to produce the extracts; it was CGI who had the contractual relationship with CEC and
critically it was for CGI to procure CEC would comply with its duties. Given the nature of
the difficulties with CEC which I have outlined this required CGI to procure CEC’s
compliance. I cannot see how any general leadership role in relation to the project or any
general management role on the part of Agilisys could have impacted on these difficulties
with CEC.
[340]       Moreover, even if I am wrong in my above conclusion regarding any issue of failure
to manage by Agilisys in respect to the data migration part of the project I believe that the
points put forward by CGI were vague and of no substance. An illustration of this is at
paragraph 207 of its written submissions it says:
“Mr Legge’s evidence contained multiple instances of Agilisys’ failure to manage
him properly.”
Page 96 ⇓
96
There is then a single reference to paragraph 2.2 of his statement about which I have already
commented. It is, for reasons I have given, entirely unclear what Agilisys by way of
management of Mr Legge or management of the project could do regarding the issues which
arose in respect of data migration.
[341]       Some issues are raised at paragraph 209 and thereafter of CGI’s written submissions
about the use of anonymised data, this however, is not a failure to manage rather it is if
anything a simple misunderstanding by each party of the other’s position.
[342]       As regards the management of CEC and Agilisys’ failure to do so there is no
specification of what at a practical level this should have amounted to. A good example of
this is in the evidence of Mr Sheriffs quoted at paragraph 227 of the submission for CGI. An
issue is raised by Mr Sheriffs who then says this: “Agilisys did not assist in managing this
process (obtaining agreement with CEC regarding access), as we had expected they would
but instead left CGI to deal with it” (emphasis added). He does not say what CGI expected
or what Agilisys could in fact do by way of management. CEC would not provide access. It
was for CGI to procure that they provided it. In this respect I believe that the points made
by Agilisys at paragraph 5.8.4 of its written submissions where it in summary refers to the
management structure within which Agilisys had to work and argues that that procedure
was used and Agilisys could no nothing more in respect of management has some
considerable force. These submissions fit in well with the detailed position of Dr Hunt as
regards project management set out in section 5 of her initial report. Her views are I believe
a correct analysis of the position. They also fit in with her analysis of the data migration
issues in part 8 of her initial report. As I have said I preferred her evidence to Mr Coyne on
this management issue.
Page 97 ⇓
97
[343]       There are also criticisms of the management by Agilisys made on the basis of the
evidence of Mr Waters (see: paragraphs 213 to 216 of the submission of CGI).
[344]       I have considered these various criticisms and in my view they amount in reality to
almost nothing.
[345]       Reference is made at paragraph 217 about the bringing in of Graham King of CGI as
an additional layer of management and that this was because of Agilisys’ failure to manage
data extraction. However, this is an example of the inversion of responsibilities as detailed
in the Subcontract. CGI was to carry out the data extraction. That was its responsibility. It
was for CGI to implement that part of the project. That this was the position is made
expressly clear by Louise Ah-Wong, if it was ever necessary to be made clear, in her 22 April
2016 email in which she set out that CGI required to produce a detailed plan for: the
provision of test data and data migration activities under CGI management” (emphasis
added). The terms of this email were, as I have said, accepted by CGI. This is a clear
acceptance by CGI that it was managing the process. Thus if CGI had to bring on board a
further person to manage the data extraction process it was because it had failed to properly
carry out its responsibilities relative to data extraction.
[346]       A criticism is made at paragraph 221 about when extracts could be regarded as
“complete”. I do not understand how this is a management issue. This point appears to
arise from a discussion between parties as to at what stage a particular extract was going to
be regarded as having been fully validated. It has nothing to do with management. It seems
to have been no more than a discussion between the parties as to what the term “complete”
would mean in the context of this particular part of the project and no more than that.
[347]       It is submitted at paragraph 224 that it was only when Mr Sander was appointed that
Agilisys properly engaged with its obligations of leadership and management.
Page 98 ⇓
98
[348]       I do not accept Mr Waters’ and Mr Brown’s evidence relative to this. I prefer the
evidence of Mr Sander for the reasons which I have already stated.
[349]       From paragraph 227 there is criticism of Agilisys’ management of CEC and flowing
from that CEC’s management of BT.
[350]       I would make the following points in regard to this criticism:
(1) In terms of the OBS set out earlier it was for CGI to manage BT.
(2) It was for CGI to procure that the Authority would comply with its duties,
including:
“Request the incumbent supplier to provide data.” (Schedule Part 6.1
Annex 2)
(3) The above two express obligations make it clear that the responsibility was on
CGI to manage both CEC and BT in respect to the issue of data migration.
(4) CEC was the only party with leverage in respect to BT. CGI was the only party
with any leverage with CEC to procure that it obtained the extracts.
On a plain reading it was for CGI to carry out its obligations in respect to
dealing with issues with BT and CEC.
(5) I believe on a proper construction the ultimate responsibility relative to BT and
CEC lay with CGI.
(6) For the reasons I have already stated CGI cannot turn a project leadership role
on the part of Agilisys into a duty which empties CGI’s responsibilities of any
content.
The next issue related to a failure to produce a change management strategy which was an
obligation incumbent on Agilisys. First the evidence relied on that this had not been
provided was this: “I would very much doubt that Agilisys provided a change management
strategy. I certainly never read one provided by Agilisys” (Colin Brown – see para 234 of
Page 99 ⇓
99
CGI’s written submissions). I am not prepared to rely on this single piece of evidence,
which does not amount to a positive statement that no business change strategy was
produced as a basis for holding such a strategy was not produced. In any event I have
earlier set out why I would not accept this witness’ evidence unless independently
supported.
[351]       As to the causative effect of any such failure at paragraph 235 again reference is
made to Mr Brown’s evidence and to a passage where he says two persons who were CEC
business users raised concerns about the templates. It gives no detail as to what their
concerns were, when they expressed these, the context in which they were expressed,
whether their concerns were justified and if so whether and how they were dealt with.
Absolutely nothing can be taken from this in respect to a failure to produce a change
management strategy or any causative effect that may have had.
[352]       At paragraph 236 of the submissions of CGI reference is made to a further section of
Mr Brown’s evidence at page 14 of the transcript of his evidence. He is asked a question
about data migration approach. He gives a lengthy answer, the essence of which seems to
be that Agilisys should speak to CEC about its data migration approach and get CEC to
agree it. As to what in causative terms is said to flow from this is entirely unclear. It may be
that this is said in relation to the issue of the archive solution which I deal with in terms of
the next question.
[353]       Within the evidence led by CGI there was further criticism of Agilisys in respect to
data migration including failures to: define data mappings; and produce data translation
rules. In respect to these matters I note that Dr Hunt rejected these criticisms at
paragraphs 8.10.4 and 8.10.6 of her initial report. I find the reasoning and conclusion in this
part of her report convincing.
Page 100 ⇓
100
[354]       Thus in summary upon the basis of the findings I have made to this point: in the
period September 2015 to June 2016, the dominant cause of delay was the inability of CGI to
procure CEC’s consent for Agilisys to have access to data from the legacy systems and
thereafter access to live data. In the period from June 2016 the root causes were: (a) the
continuing impact of the substantial initial delay in the data extraction work; and (b) a lack
of consistent or proper resources from CGI together with wider issues of environments
which I will shortly turn to discuss; and (c) CEC’s change of position in relation to the
volume and availability of historic data. I believe therefore the dominant cause of delay was
a CGI cause.
[355]       For the foregoing reasons I do not believe that CGI’s breach was caused or affected
by a breach by Agilisys of its obligations. It seems to me for the reasons outlined that CGI
breached its obligations to provide the necessary data extracts due to its failures as above set
out.
[356]       The next question is this: did that breach cause, or was it reasonably likely to cause,
Agilisys to fail to achieve the Milestone MS-2.5 in the ERP Implementation Plan by the
effective date? MS 2.5 per the Implementation Plan had a start date of 3 May 2016 and end
date of 26 July 2016.
[357]       Mr Sandison’s position was this: that any failure by CGI to provide data extracts in
breach of its obligations was not, in any event, the operative cause of Agilisys’ failure to
achieve the User Acceptance Testing (UAT) milestone. It was made clear by CEC that UAT
could not proceed without an adequate solution for historic data. The creation of an archive
solution was the responsibility of Agilisys, as was management of CEC at a project level and
change management. Agilisys failed both to generate a satisfactory solution but also to
manage CEC in order that a solution acceptable to all parties could be deployed.
Page 101 ⇓
101
[358]       This issue of the archive solution appears to have first been raised at an ERP project
meeting on 31 August 2016. An analysis of this issue was produced by Mr Sander
suggesting a data warehousing solution for historic data. This was rejected at a Project
Board meeting on 22 February 2017. On the basis of these points Mr Sandison submitted:
Agilisys accordingly failed in both its obligation to produce the archive solution in order to
allow UAT to proceed, but also in its obligation to lead the change management process
with CEC in order to produce a solution which was in an acceptable form for CEC.
[359]       Mr Cormack in reply submitted: in the period from June 2016 the root causes were:
(a) the continuing impact of the substantial initial delay in the data extraction work; and
(b) a lack of consistent or proper resources from CGI; and (c) in specific response to the
argument of Mr Sandison regarding the data archive solution he submitted that there had
been a change of position by CEC in relation to the volume and availability of historic data.
In reality, what happened in this period was that from the ERP Project Board meetings on 13
and 29 October 2016 onwards, CEC took the view that they needed to have absolutely all
historic data available to them. This represented a fundamental departure from the design
decisions recorded in the SDD. It did not represent a failure of management by Agilisys. If
this was anybody’s fault, it was CGI’s fault for failing in its responsibility to use its
commercial and contractual powers with CEC, and in particular to emphasise that this
amounted to an unauthorised change to the scope of the already signed-off design
documentation. Beyond that he argued that archiving was not in any event due to be done
until after the planned go live date (see: Appendix 1).
[360]       I am persuaded by the argument advanced by Mr Cormack. First there appears at a
fairly late stage to have been a material change in the requirements of CEC in respect to
archiving. The SDD recorded, in terms, what was to happen in respect to historic data. It
Page 102 ⇓
102
was to be dealt with as follows: there would be a read only instance of Business World that
amounted to an archive of historic data. As regards HR it was only to be historical payroll
information for the years 2008 to 2015 and in Accounts Receivable it would be all customer
invoices, customer notes, credit notes from 1 April 2009 onwards. In addition categories of
data were identified which were not to be archived. Thereafter CEC moved to a position
where it required all data to be archived and for it to be available seamlessly (emphasis
added). It is not a failure in management on the part of Agilisys that CEC have at a late
stage made such a unilateral decision. I do not see how any form of management could deal
with such an issue. In particular management could not in any sense seek to deal with this
issue until it arose. Thereafter Agilisys sought to deal with the matter through the
document prepared by Mr Sander. Secondly, in circumstances such as this it appears to me
that it was ultimately for CGI to exercise its contractual and commercial powers to
emphasise that this amounted to an unauthorised change. It is notable, as argued by
Mr Cormack, that while the archiving strategy created a delay ultimately CEC eventually
returned to the same basic approach outlined in the SDD. In any event I accept his
argument as to when archiving was to be available.
[361]       It appears to me that Dr Hunt’s careful analysis reaches the same conclusion (see:
section 8.9 of her original report). There was no breach of any obligation by Agilisys.
[362]       For the above reasons I do not accept Mr Sandison’s argument. CGI’s breach was
reasonably likely to and in fact eventually did cause Agilisys to fail to achieve Milestone 2.5
(completion of UAT).
[363]       In the original Implementation Plan, User Acceptance Testing was due to start on
3 May 2016 and conclude on 26 July 2016. It is obvious that, given CGI did not even have
Page 103 ⇓
103
access to the legacy systems until after the date for commencement of UAT, then the failure
to provide extracts was likely to cause Agilisys to miss this deadline.
[364]       When relief was granted in May 2016 and by agreement to the 22 April 2016
Conditions and milestones moved to comply with version 13 of the Project Plan, User
Acceptance Testing was due to start on 10 August 2016 and conclude on 8 November 2016.
Again, it is plain from the above that progress on extracts did not even commence in earnest
until around July/August 2016, and was not even complete by the revised deadline in later
Project Plans of 29 September 2016.
[365]       For the foregoing reasons I am persuaded that the delay in provision of data extract
was a dominant cause of delay as regards the ERP project and that as a direct consequence
Agilisys was likely to miss Milestone 2.5 for UAT (as set out in the Implementation Plan
which was due to start on 3 May 2016 and conclude on 26 July 2016 [see Schedule Part 6.1]).
It could not even have met the dates in the Project Plan. I will discuss the issue of
concurrent dominant causes later in this opinion.
ERP environments: RN001, RN008, RN 009, RN013, RN014 and RN015
[366]       The disputed Relief Notices in this group are RN009, RN014 and RN015.
[367]       RN009 concerned asserted delay of access to hosted ERP production and
pre-production environments and asserted that a CGI cause was reasonably likely to cause
Agilisys to fail to achieve Milestone 2.5 completion of UAT (issued 18 March 2016).
[368]       RN014 concerned asserted delay of access to ERP Environments and it asserted that a
CGI cause is continuing to cause Agilisys to fail to achieve all remaining Milestone Dates.
CGI have not provided reasonable access to the following hosted environments Gold, dev,
Page 104 ⇓
104
test, PPE-A and PPE-B including environment management services to support the delivery
of the next phase of the projects (issued 26 July 2016).
[369]       RN015 concerned asserted failure of provision of accessible ERP Environments and it
asserted that a CGI cause is continuing to cause delay to the ERP project and this may result
in Agilisys failing to achieve Milestone Dates. CGI have not provided all necessary services
and functionality to the following hosted environments Gold, dev, test, PPE-A and PPE-B
including environment management services to support the delivery of the next phase of the
project.
As a result of ongoing issues with CGI providing fully functioning ERP environments, there
have been delays to the Project:
On 13/6/2016, CGI accepted that it had not provided hosted environments and
agreed that delay to the project was a CGI cause. At this stage the delay was
expected to be limited to 4 weeks.
On 26/7/2016, CGI had still not provided usable pre-production environments.
This was a further 7 weeks of delay since CGI accepted the CGI cause on
13/6/2016. Agilisys raised Relief Notice RN014 to document the causes of delay.
As at 11/10/2016, CGI has not responded to RN014.” (issued 13 October 2016)
[370]       An environment in this context means: a physical or virtual system in which a
computer program can be run. It is normal practice for an organisation to have multiple
environments when delivering a new system e.g. a test environment and a production
environment.
[371]       The first question in relation to this group of Relief Notices is this: what are the
respective obligations of the parties (including as to the Milestone Dates in the
Implementation Plans and timing generally) in relation to the provision of fully functional
Page 105 ⇓
105
environments and related infrastructure, including suitable environment management
services, in relation to the ERP Project?
[372]       Agilisys contended that the Subcontract allocated the following responsibilities to
CGI either directly or by virtue of its obligation to procure that CEC perform its
responsibilities.
Section
Extract
OBS 1, OBS 2, OBS 3, Procure the Unit 4 Business World ERP solution and manage
OBS 30, OBS 47, OBS 51 the contractual relationship with Unit 4.
Provide infrastructure hosting and support of the Unit 4
Business World ERP solution
Provision of required system access to Agilisys to deliver
programme work.
Roles &
Responsibilities
document, Section C of
Schedule Part 4.1
Project Roles:
CGI providing hosting, technical, DBA and infrastructure
services (including the copy and creation of environments as
required and no fewer than 5 environments).
CGI Responsibilities and Deliverables:
Infrastructure build and hosting of the Agresso solution x 5
environments and the provision of access to Agilisys to maintain
the Agresso installation across the environments.
Management of Unit 4 contract and on-going support.
Agilisys contended that the following are Agilisys’ responsibilities, in terms of the
Subcontract:
Section
Roles &
Responsibilities
document, Section C of
Schedule Part 4.1
Extract
Project Roles:
Supplier leading delivery of this project including
commissioning and management of the supplier UNIT4
(through Agent relationship) implementation services. This will
cover specialist product support and bespoke development for
interest across instalment plans. The Supplier will rely on CGI
PMO for support. CGI to support with the Authority
Page 106 ⇓
106
governance and resources.
Supplier Responsibilities & Deliverables:
Client Management and Governance (Project Level)
Programme management of the implementation programme
and generation of all programme control documents.
Management of Unit 4 implementation services (Agent
Relationship) across the programme, including management of
bugs and resolution.
Facilitate design workshops to agree the implementation design
of the Agresso system…
Solutions architect resource which will generate the technical
solution design, interface strategy, data migration strategy, the
archiving solution and design of the supplier portal. Each of
these will produce a document deliverable for sign-off with the
Authority. The solution architect will manage the technical
resource provided by CGI.
[373]       I did not understand CGI to dispute the foregoing as being Agilisys’ obligations
other than to add one further responsibility: “The Supplier responsible for the system
design and build using their local government templates where applicable”. Mr Sandison
stressed that CGI’s obligations must be considered in terms of the general leadership
obligation of Agilisys in terms of the ERP project.
[374]       I consider that the obligations referred to by each side in respect to these Relief
Notices were incumbent on the parties and were relevant to the discussion of the various
issues relative to these Relief Notices.
[375]       CGI contended that the effect of the respective obligations was that whilst CGI were
to provide the environments, responsibility for the leadership of the project and the
provision of environment design were the responsibility of Agilisys. In respect to the issue
of design it was argued that the critical obligation was that incumbent on the solutions
architect to generate a “technical solution design”. In development of this argument it was
Page 107 ⇓
107
submitted that the design to be generated required to encompass a design for the complete
Business World solution (rather than the position put forward by Dr Hunt to the effect that
“solution” related to “the software aspect of the solution primarily”). Mr Sandison argued
that the interpretation of “technical solution design” which he was advancing was consistent
both with the words of the Responsibilities Document (which neither seek to restrict the
aspects of the design for which the Solutions Architect is responsible nor allocate
responsibilities for design elsewhere) and also with the evidence of Mr Coyne. Mr Coyne’s
evidence, which he invited the court to accept, was that the Solutions Architect was
responsible for designing the “technical aspects of a system”, including:
a. How much storage space would be required?
b. What performance profile would the storage need to have?
c. How much load would the system be under?
d. How much peak computing power would be required?
e. The size of the memory?
f. The requirement for swap space?
g. Who needs to connect to the system?
h. Network/Access requirements?
i. Tolerance for latency?
j. Testing requirements?
k. Where will users be connecting from?
l. What devices will they be using when they connect: PCs, laptop or mobile
phone?
[376]       Mr Coyne’s evidence was that these matters required to be designed prior to CGI
commencing the build of environments: because designing environments, certainly in this
Page 108 ⇓
108
security-sensitive space, only wants to be done once. Essentially, any rework in a data
centre where there’s secure data is generally costly, typically in terms of time, because
there’s a process that you need to go through to change anything within the environment.
You can’t just tweak something. All of the considerations about security impact and things
like that must be considered.
[377]       Mr Cormack’s position in summary was that I should prefer the analysis of Dr Hunt.
[378]       When considering the opinion evidence of an expert, evidence which is neither
relevant nor admissible is what a contractual provision means, that being a matter entirely
for the court. However, in terms of the Subcontract the term “technical solution design” is
not a defined term. It is first a generic term and secondly a technical term. Given the
foregoing in reaching a sound conclusion as to the meaning of the above term the court is I
consider entitled to have regard to expert technical evidence and guidance as to what such a
term means and what in practice is required to fulfil such an obligation. Accordingly
assistance can be sought from the views of Dr Hunt and Mr Coyne as regards this matter.
Each party directed my attention to the views of their respective expert in respect to this
question. Although placing a strict demarcation line in respect of such evidence may in
practice be difficult I believe that Dr Hunt’s and Mr Coyne’s evidence regarding this issue is
admissible. Dr Hunt at paragraph 6.2.2 to 6.2.6 of her initial report sets out her views:
6.2.2 As discussed in section 4.2, terms such as ‘infrastructure services’, ‘technical
solution design’ and ‘solution architect’ are generic and need to be read in the
context of a project that sits within a wider Programme.
6.2.3
A solution architect is typically someone with knowledge of the design issues
specific to a particular solution or application. In the context of an ERP
implementation within a much larger transition and transformation
programme I would expect the ‘technical solution design’ produced by such
an architect to be confined to the technical design for the ERP solution. I
would not expect an informed reader of this contract to expect an ERP
solutions architect to be able to design every detail of the infrastructure
required for the various environments. For example, large, complex
Page 109 ⇓
109
installations typically have separate architects responsible for networking and
security aspects of the infrastructure that apply across all applications and
which an ERP solution would need to work within.
6.2.4
Decisions and information about the data centre that would house the
environments and the network, security and other infrastructure that
comprised CEC’s IT systems and the network connections required to
connect the two were all the responsibility of CGI. In my view the primary
responsibility for designing the infrastructure to support the required
environments lay with CGI.
6.2.5
However, CGI did need some assistance from both Agilisys and Unit 4, who
needed to provide information about the infrastructure required by BW, such
as the number of servers required to support the expected user population
and database size.
6.2.6
Given that there was a requirement for a number of environments across the
two projects I would have expected CGI to appoint a systems architect or
technical architect to work with Agilisys and Unit 4 to produce the
infrastructure design. As CGI were also responsible for hosting, technical,
DBA and infrastructure services as well as supporting the system once it
went live I would also have expected them to appoint an environment
manager straightaway who would be involved in the design process and take
on support for the environments to ensure that they were properly setup and
maintained. As discussed in section 3 I would also expect those resources to
have been supported by a CGI project manager with responsibility for liaising
with BT to ensure timely access to information about CEC’s legacy
infrastructure.”
I find the analysis contained in this section of her report persuasive. It is clear, thorough and
well reasoned.
[379]       I would also refer to her evidence at paragraphs 2.2.1, 2.3.2 and 2.3.4 where she sets
forth a further analysis of this issue which is consistent with the views sets out in the
paragraphs I have set out in full above.
[380]       She explained in a little more detail in her re-examination her reasoning in respect to
this matter:
“So CGI were designing infrastructure for the wider programme. They were
designing and networking to connect Edinburgh’s offices with what was going to be
the new IT estate, some of which was going to be moved from BT and some of it was
going to be new. They had to take account of whatever CEC already had, what
CEC’s policies were, what their security arrangements were and what they wanted
Page 110 ⇓
110
them to be in the future, even down to the telephony. So whatever was designed for
the ERP and EI systems had to co-operate and work inside that. It was going to be in
CGI’s data centre along with everything else. So they had to design. The ERP and EI
environments were just the environments within that overall infrastructure design,
so it had to be a piece of work that they did.” (see: p 127 of the transcript of her
evidence)
[381]       Her analysis is consistent with the structure within which the Subcontract existed,
namely: it being a small part of CGI’s prime contract with CEC and that CGI was
responsible for not only the infrastructure build but the hosting of this within its data centre.
That is part of the “documentary, factual and commercial context” in which the relevant
words have to be construed. Such a construction does not damage the ordinary meaning of
the clause. It does not limit “technical solution design”. It gives the term its meaning within
its proper context in the way asserted by Mr Cormack. Mr Coyne in his opinion takes no
account of this context and if for no other reason I would have rejected his opinion in
relation to this issue.
[382]       I accordingly am persuaded that the technical solution design obligation incumbent
on Agilisys did not mean that the solutions architect was to produce the design for the
infrastructure. The “primary responsibility for designing the infrastructure to support the
required environments lay with CGI” (see: 6.2.4 Dr Hunt’s first report). I reject Mr Coyne’s
evidence, where it differs from Dr Hunt’s evidence on this issue. There are, I recognise,
areas where there is agreement.
[383]       In respect to the proper construction of Agilisys’ obligation: “Management of Unit 4
implementation services (Agent Relationship) across the programme, including
management of bugs and resolution.” Mr Cormack made a very detailed written
submission:
4.3.3. In order to understand the reference to management of ‘Unit 4
implementation services’, it is necessary to read the whole of the Subcontract
Page 111 ⇓
111
in its proper factual context. A fact that seems to have escaped later witnesses
and experts for CGI.
4.3.4.
First, in the Project Roles section of the Roles & Responsibilities document for
ERP, Unit 4 is expressly described as providing: ‘Business World software,
procurement solution and limited implementation services including
technical training services for client only (Note there is currently no budget
for external training of CGI support staff).’ There is, accordingly, a
recognition that the term 'implementation services' has a more limited and
specific meaning in this context. Further, that it would not include the
training provided by Unit 4 as part of installation. Secondly, this same
obligation is given slightly more detail in the OBS. In the OBS, there is an
obligation on CGI to ‘Procure training and implementation services from Unit
4 which will be managed by Agilisys as CGI agent. 175 days required 100 to
support the solution implementation, 25 for bespoke development, 50 for
training.’ That very same breakdown was provided by Claire Conaghan to
Niall Sherriffs the day after the Subcontract was signed in her email of 16
September 2015: JB [ ] at p. 8037. As she explained in that email, the 100 days
were to support specific delivery elements of the ERP project. Nothing was
mentioned about installation or installation services. This is a
contemporaneous account at exactly the time of the Subcontract, from the
person involved in drafting the OBS, and at a very early time when Claire
Conaghan would have had no reason or desire to minimise Agilisys’ role.
Fourthly, the reason for this was very clearly explained by Robert Price at
para. 6.17 of his Witness Statement (on which he was not challenged). It is,
also, consistent with the email exchange between Bob price and Terry Wiest
on 27 August 2015: JB 440. It is, also, consistent with the same distinction Unit
4 itself made between installation and implementation services in its email to
Niall Sherriffs on 21 September 2015: JB 465. It is also consistent with Raj
Sudra’s email of 4 November 2015 in which he stated ‘..but because of the
commercial relationships we are managing Unit 4 for their ‘specific’
deliverables.’: JB 542. At no point, from Claire Conaghan pointing CGI to Unit
4 in September 2015 onwards, was it ever raised with Agilisys that they
should be managing Unit 4.
4.3.5.
In short, CGI’s witnesses and expert have tried to elevate this obligation out
of all proportion and context. It is quite clear from the wider context referred
to above that the references to management of Unit 4 resource had nothing to
do with the base installation. It was, always, intended to relate to the
procurement of specialist resource from Unit 4 up to a specified amount of
days to provide support for configuration of the software, bespoke
development of the product, and training during the project.
I believe that this is a sound analysis of the contractual position. Implementation, when
looked at in the context of the Subcontract has a limited meaning and does not relate to the
issue of base installation. Agilisys’ responsibilities in respect to Unit 4 were limited to
Page 112 ⇓
112
supporting CGI where necessary and requested. Thus Agilisys’ management role in respect
to Unit 4 was limited in its scope and did not relate to installation. This I think is a sound
construction of the Subcontract and also fits in with the evidence and the way that matters
operated on the ground.
[384]       The next question is: did CGI breach its obligations by failing to provide fully
functional environments and related infrastructure, including suitable environment
management services, on the ERP Project? CGI in its written submissions said this in
relation to the above question:
“271. The five environments specified in the Responsibilities Document, and the
dates upon which they were recorded as being delivered were as follows (as set out
in the in the [sic] ERP environments status update of 9 September 2016):
a. Development environment: provided in December 2015.
b. Test environment: provided in December 2015.
c. Gold environment: provided on 25 March 2016.
d. PPE-A (part of the pre-production environment requirement): provided
on 7 June 2016.
e. PPE-A (part of the pre-production environment requirement): provided
on 30 June 2016.
f. Production environment: provided on 1 August 2016.
272. In addition, a ‘sandbox’ environment was identified by Agilisys as a
requirement in August 2016 to allow CEC staff familiarisation. The environment was
provided on 26 August 2016.
273. Environments were accordingly not provided within the contractual
timeframe.
274. Following provision of the environments on the dates above, further
specifications or requirements were raised by Agilisys. Peter Hall accepted that
Agilisys ‘habitually’ raised new defects with the environments. In addition,
difficulties were experienced with access to environments (both discussed in relation
to question 23 below). These issues were raised in RN016 (should be RN015) (which
concerned issues with environments which had been provided, rather than the
provision of environments themselves). For the reasons set out in relation to
Page 113 ⇓
113
question 23, it is submitted that there was no breach by CGI of its obligations in
relation to the issues raised by RN016 (should be RN015).”
[385]       Mr Cormack’s position was that CGI had breached its obligations in respect to
environments. It had failed first to provide the development environment within the time
required by the Implementation Plan. In development of this he argued CGI’s obligation to
build and provide access to the five environments was not given an express time limit.
However, he submitted that regard has to be had in this context to the various dates that
parties had agreed in the Implementation Plan in the Subcontract. He again submitted that
regard had to be had in this respect to dates agreed in Project Plans. I accepted this
approach for reasons I have set out.
[386]       He then drew my attention to the Implementation Plan in Schedule Part 6.1 where
parties had contracted for the “Agresso infrastructure buildactivity to take 55 days from 14
September 2015 to 27 November 2015. This was to be immediately followed by the
commencement of Build activity. He then submitted in actuality, the Development
Environment was not first made available until 9 December 2015. When it was made
available, it was unusable. It was not made fully available until 16 February 2016.
Accordingly, on any view, CGI failed to meet its responsibilities to provide the Development
Environment within the time necessarily required by the Implementation Plan.
[387]       It was also his position that CGI had failed to provide the PPE and Production
Environments in the time required by the Implementation Plan. In elaboration of this
submission the original Implementation Plan in Schedule Part 6.1 (consistently with
version 9 of the Project Plan) made no distinction between the different dates on which the
five environments would be made available. The intention at this stage was that all five
would be provided together. On any view, CGI utterly failed to build and provide access to
the PPE and Production Environments in those timescales.
Page 114 ⇓
114
[388]       Even looking at the dates agreed in later Project Plans and other more informal
agreements, CGI consistently failed to meet any of the deadlines agreed. In particular, it
missed the deadline of 1 March 2016 in version 9 of the Project Plan, the deadline of
31 March 2016 in version 12 of the Project Plan, the deadlines proposed by Kevin Bryce in
March 2016 and 29 April 2016, and the deadlines of 13 May and 4 July 2016 in version 13 of
the Project Plan. In fact, PPE-A was only made available at the beginning of July 2016, PPE-
B was not made available until mid-July 2016, and Production was not made available until
into August 2016.
[389]       It is submitted that, on any conceivable view, CGI plainly failed to build and provide
access to the Environments within the time required by the Implementation Plan.
[390]       In support of its position Agilisys produced a detailed chronology of the events
surrounding environments in the form of an appendix. Having considered the terms of that
Appendix I believe first it accurately set out the timeline; second it accurately sets out by
reference to oral evidence and documentation, a factual basis upon which I conclude that
CGI had breached its obligations as set out in the above question; it supports the factual
basis set forth in each of the Relief Notices. I have produced it as Appendix 2 to this
opinion. It clearly supports the submission that CGI had breached its obligations in respect
of provision of hosted environments, that being the issue raised in Relief Notice RN009 and
Relief Notice RN014. So far as breach of obligations asserted in RN015 given the position
taken up by CGI regarding this issue I will consider whether it breached its obligations in
the context of my consideration of the next question which is: was CGI’s breach or breaches
caused or affected by a breach by Agilisys of its obligations?
Page 115 ⇓
115
[391]       Before turning to that question I think it is helpful to set out the position taken up by
CGI in relation to the Relief Notices served regarding environments during the course of the
project.
[392]       Between the commencement of the Subcontract and June 2016, Agilisys issued six
Relief Notices relating, generally, to the environments as follows: (1) RN001 on 27 January
2016 for delay of access to ERP Development: JB 10; (2) RN003 on 28 January 2016 for delay
of hosted Biztalk infrastructure: JB 12; (3) RN004 on 10 February 2016, in part, for
continuing delay to Biztalk infrastructure: JB 13; (4) RN006 on 18 March 2016 for delay of
access to Biztalk development environment; (5) RN009 on 18 March 2016 for delay of access
to ERP PPE and Production Environments: JB 18; and (6) RN013 on 3 June 2016 for delay of
access to ERP PPE and Production Environments: JB 22.
[393]       CGI, either immediately or eventually, accepted all of these and purported to grant
relief. In particular: (1) they expressly accepted RN001 in their response of 10 February 2016
and granted relief by accepting that the dates in the Implementation Plan had been amended
to bring them into line with version 12 of the Project Plan, and they indicated that hosting of
the temporary environments would amount to ‘full and final compensation’ for that Relief
Notice; (2) they eventually accepted RN003, RN004, and RN006 as part of the 22 April
Conditions and their composite response on 1 May 2016 (see JB 30), and essentially granted
relief in the form of postponing Milestones in accordance with the re-planned dates in
version 13 of the Project Plan, and by confirming there would be no liability for Delay
Payments; and (3) RN013 was agreed as CGI Cause and relief was agreed to the extent of
recognising that Agilisys would be unlikely to be able to meet the deadline of System
Acceptance Testing.
Page 116 ⇓
116
[394]       CGI’s position in summary regarding this question was: CGI could not properly
fulfil its obligation to provide environments without the prior fulfilment by Agilisys of its
obligation to specify the design for these environments. It was Mr Sandison’s position that
Agilisys failed to discharge this obligation. This caused, or at the least materially affected,
the issues which were experienced with regard to environments.
[395]       The line taken by all of CGI’s witnesses was that the dominant cause of the delay in
respect to the provision of the environments was the failure by Agilisys at the outset to
provide a specification for the environments or put another way the failure by Agilisys to
advise them at the outset what its requirements were in respect to the various environments.
[396]       The witnesses for CGI who gave evidence on this issue were Paul Hamilton,
Kevin Bryce, Niall Sheriffs, Cliff Graham, Colin Brown and lastly Bridget Fox.
[397]       It was evident on looking at the evidence of these witnesses, when taken as a whole,
that their position at the stage of proof was this: this failure had existed from the start of the
ERP project and continued throughout the life of the project. On looking at their evidence as
a whole I understood their position at the proof to be that this had been a critical failure.
[398]       The criticality of that failure, according to the CGI witnesses, is perhaps best
captured in the evidence of Paul Hamilton who in his witness statement at paragraph 2.1
says this:
“We couldn’t do anything constructive until Agilisys told us what should be in the
environments. Essentially we were working blind.”
[399]       If that is a proper characterisation of the failure I would expect from the outset there
would be documentation emanating from CGI calling on Agilisys to produce this
specification. I would expect there to be considerable documentation thereafter to the same
effect (but with the demands made with increasing urgency) which produced a consistent
pattern of demands that such a specification should be produced by Agilisys. Looking at the
Page 117 ⇓
117
evidence of CGI’s witnesses and the contemporary documentation produced I am unable to
identify such a pattern in the documentation. This casts substantial doubt on whether there
was any failure to produce such a design. Nor does the position which CGI now advances
fit in with its responses to Relief Notices which I have earlier set out. This equally
undermines the position taken up by CGI at proof.
[400]       The first witness on behalf of CGI who spoke to this issue was: Paul Hamilton who
was CGI’s man on the ground relative to environments at the outset of the Subcontract.
[401]       At page 90 in the transcript of his evidence he is asked this question:
“Q. And are you able to identify to us from your witness statement or otherwise
any written communication, including an email, from you to Agilisys where
you are complaining about Agilisys not having provided an environment
design in relation to the development environment?
A: I am… off the top of my head I’m not aware of any, no.”
[402]       To put his answer fully in context he also says at page 91 that the issue had been
raised with Agilisys on a number of occasions verbally.
[403]       Further, at pages 57 and 58 in cross-examination he is asked about a document
prepared by a CGI colleague Sebastian Wood and he is directed to the fact that in this
document no reference is made to Agilisys providing a requirements document or a design
document. He is asked if this is a significant issue why is it not referred to there. He replies
by saying he does not know what was in Mr Wood’s mind.
[404]       I of course accept that he cannot know what was in Mr Wood’s mind. However,
given the nature of this document I believe it is odd if the issue of a design document was a
significant matter that it is not referred to therein and given the criticality of the issue
according to Mr Hamilton’s evidence it is odd that he could not point to a single document
in which this design had been sought by CGI. This inability on the part of this witness to
point to anything in the documentation raising this issue with Agilisys I believe strongly
Page 118 ⇓
118
supports the view that this lack of a design was not an issue of any significance and that
Agilisys did not breach any obligation in respect thereto.
[405]       The next witness from CGI who gave evidence on this issue was Kevin Bryce. From
about February 2016 he took over from Mr Hamilton as environment manager in respect to
the ERP project.
[406]       Once more it was a theme of his evidence that there was a necessity for a single
document setting out the requirements for the environments to be provided by Agilisys in
order for CGI to build the environments.
[407]       On 4 March 2016 he emailed Mr Hall of Agilisys (see: JB 881, page 11476). In that
email he asked for a copy of the development environment requirement (see: page 31/32 of
the transcript of his evidence). However, I note in the same email he says: “I am aware that
the ERP dev and test environment have already been provided”. Mr Hall replies the same
day setting out Agilisys’ position (see: 32/33 of this witness’s cross-examination).
[408]       Mr Bryce then accepts in his cross-examination that requirements had been set out by
Agilisys but in the form of “a series of conversations and in a couple of different emails”
(see: page 34 of the transcript). However, he again complains: “we lacked that single
document which we expected to see that said. This is what we need in the environments to
make it fit for purpose”.
[409]       It appears from the above passages of evidence that environments were able to be
built given the requirements stated by Agilisys. Secondly it appears that the concern on
Mr Bryce’s part is not that Agilisys failed to state its requirements but they were not set out
in the form of a single document.
[410]       When the evidence is analysed this need for a single document appears to be no
more than this: it was required to satisfy an internal requirement of CGI’s and not in reality
Page 119 ⇓
119
because CGI could not build the environments without such a single document. This is clear
from an answer given by Mr Bryce at page 110 of the transcript of his evidence where he
says this:
“Yes, that’s correct. So that was us trying to get the documentation in to a state that
GIS [an internal body within CGI] would accept it without having a formal
documented set of requirements from the application owners.”
[411]       This evidence is echoed in the evidence of Mr Hamilton at page 90 in the transcript
where he says this:
In CGI we are not allowed to simply stand up and make live servers and
environments without a design, so we were forced into writing our own HLD to
allow work to proceed within CGI. CGI has some very strict governance processes
around deploying servers. As you’d imagine, we support quite a lot of organisations
and our servers are tens of thousands in the number, so we had to draft our own
HLD, which was a very, very high-level diagram based on the environments matrix
which Agilisys had issued us.”
[412]       So once again the necessity for some single document appears to be in order to
satisfy internal requirements of CGI.
[413]       Mr Bryce makes a further comment on this issue at paragraph 2.7 of his statement
where he says this:
“Agilisys did not provide us with detailed requirements for the environment until
5 July 2016… and even then they were incomplete. CGI approaches things in a much
more structured and documented fashion which lends itself to an orderly design
solution, I get the impression that Agilisys have less structure. That led to an
incompatible work approach.”
[414]       Once more the requirement for a single document appeared to come from internal
CGI work practices.
[415]       I gained the impression that the issue of the single design document issued at the
outset in the end of the day was really no more than a difference in working practices
between CGI and Agilisys which had no causative effect so far as the issue of delay is
Page 120 ⇓
120
concerned. The issue did not appear to be: no requirements were provided by Agilisys but
they were not provided in a single document.
[416]       Returning to the evidence of Mr Bryce, at a number of points in cross-examination he
was asked to consider documents which had been produced at various stages by CGI in the
course of the project and asked why in this documentation the provision of a single
document setting out Agilisys’ requirements is not referred to (see: page 40-42; page 55,
lines 12-16, pages 67-72; page 83, lines 1-9; page 87 and 88; and pages 111 and 112).
[417]       In these documents the lack of reference to the absence of a design document and of
Agilisys being called out in relation to this issue by CGI strongly points to there being no
issue in respect to this matter and that Agilisys had fulfilled its obligations regarding this.
[418]       Mr Bryce was the person on the ground who was in charge of the production of the
environments. If there was a real problem with this lack of provision of a single document
setting out the requirements which was materially hampering CGI in the building of the
environments it would have been called out somewhere by this witness and in particular in
the various documents to which he was directed. Given his position in relation to
environments, given his character which seemed careful and given the bureaucratic,
structured nature of the way in which CGI according to him approached its work in respect
to matters such as environments then if this lack of a single document setting out the
specifications was a real and significant issue it would have been documented by him. It
was not.
[419]       The next witness who gave evidence in respect to this issue from CGI was
Mr Sheriffs. As I have already said he was referred to the April email from Louise
Ah-Wong, I set out the text of that email earlier on.
Page 121 ⇓
121
[420]       He conceded that the conditions contained in the email had been accepted by CGI.
Between pages 60 and 62 he was asked why CGI would accept such conditions if Agilisys
had significantly contributed to the problems as regards provision of environments. His
response for reasons I have already detailed I did not find plausible. I have commented on
this to some extent earlier. I would add that in particular in relation to environments, if the
position as now put forward by CGI is correct then I believe these conditions would not
have been accepted and the reply would have been something like this: Agilisys is to blame
for the delay because it did not provide a specification for environments. This email is
dealing explicitly with the issue of environments.
[421]       Mr Sheriffs was not a low level functionary in CGI. He worked at a high level within
that organisation and was a senior person in respect to both the programme and the projects.
Given his position within CGI I simply did not find his position convincing as to why this
document was agreed to if the position was as he now set out in his evidence.
[422]       It appeared to me that the response to the Ah-Wong email was inconsistent with the
position which CGI was seeking to advance at the proof relative to environment
specification.
[423]       Cliff Graham also gave evidence on this issue.
[424]       He was also questioned about Ms Ah-Wong’s email in the course of his
cross-examination.
[425]       His initial response to the Ah-Wong email is contained at page 12725 in the JB. It
was this:
“My initial impression is that the requests from Agilisys are not unreasonable.”
(see: page 59 of the transcript)
[426]       Mr Graham was then asked at page 59 was the above response not inconsistent with
his earlier having said in his evidence that he had been told that no requirements had been
Page 122 ⇓
122
set out by Agilisys beyond “rudimentary instructions” (see: page 55, lines 15-23). The
questioner is careful to ensure that the witness understands the question by asking him to
confirm his answer.
[427]       In response to the above question Mr Graham gives a series of answers between
pages 60 and 64 of the transcript culminating, when it is plain to him that the response
which he had given to the Ah-Wong email at the time is inconsistent with the position CGI
is now taking he says: he did not have the information at the time when the response to the
email was drafted (see: page 63, lines 7-13 of the transcript). However earlier in his
evidence at page 55, lines 15 to 23 of the transcript he had said that he did have the
information. I believe that this witness’s change of position on what was a matter of some
importance undermined his evidence. I do not think on this matter that his evidence was
acceptable.
[428]       There are thus two senior members of the CGI team who at the time took up
positions with respect to the Ah-Wong email which are wholly inconsistent with the
position which is now being put forward by CGI. Again I think that the contemporaneous
documentation shows the true position, namely: there was no issue on CGI’s part of any
significance regarding a failure to provide a single requirements document relative to
environments.
[429]       The next witness for CGI was Colin Brown who took over the ERP project from
Mr Sheriffs in July 2016. When he came on board it was his position that there were
problems regarding the ERP project. CGI accordingly carried out what was described as a
lessons learned exercise. As I understand it this was a review of what had occurred to date
in order to identify problems which had occurred and means of solving these problems.
Page 123 ⇓
123
[430]       At pages 40 and 41 of the transcript of evidence he was asked about the content of
the document which was prepared as a result of the lessons learned exercise and in
particular he was asked if this document raised issues of lack of information necessary for
environments to be built being provided by Agilisys. He answered that no it did not raise
that matter, however, he said that this was because the document was going to be made
available to the client, CEC.
[431]       This document appeared to me to be a clear opportunity to set forth, what as I
understand it, CGI is saying was a critical issue for it at the time of the implementation of
the project. If this was a real issue at the time I would have expected this to be clearly set out
in this document. The reason given by Mr Brown as to why it was not clearly articulated
therein makes no sense if it was in fact the critical issue. What is the point in carrying out
the exercise and preparing this document if critical issues are not to be raised in the
document.
[432]       In respect to Ms Ah-Wong’s email he gave no plausible explanation as to why CGI
had accepted the conditions contained therein. He tried to say this was a matter for others
in CGI, however, given his relationship to the response (see: page 65) this did not make any
sense. He repeated this answer seeking to say that it was other persons who were
responsible (see: page 65).
[433]       This matter is returned to at page 75 and 76 of the transcript. His answer there
saying that Agilisys was at fault does not explain why therefore CGI had agreed to the
conditions.
[434]       Looking at the evidence led on behalf of CGI on this issue as a whole, I observe that:
over a significant period of time; from the stand point of persons who were working on the
ground in order to build the environments; from the perspective of persons higher up in
Page 124 ⇓
124
CGI’s management; and in various types of contemporaneous documentation there is a
significant lack of evidence that there was a real issue regarding a failure to provide a single
document specifying the requirements of the environment by Agilisys. This evidence I
believe strongly points to this not having been an issue of any significance at the time and
that Agilisys had not breached any obligations in relation thereto.
[435]       Between paragraphs 322 and 340 of his submissions Mr Sandison sought to highlight
contemporaneous documents in which this was raised. At paragraphs 325 and 326 he
contended that the documents prepared after the lessons learned exercise raise the issue.
Nothing referred to there says that there was a failure by Agilisys to provide a specification
document at the outset. At its highest this is said “As subject matter experts, Agilisys/Unit 4
should have provided greater detail…”. That is not the same thing. It does not say Agilisys
should have produced at the outset a document setting out its requirements. It does not say
what precisely that document should have contained. This is no more than a passing
reference in this document. If there was an obligation on Agilisys to produce this document
at the outset and its failure to produce this had been the dominant cause of delay, this
document would have contained more than such a general passing reference.
[436]       At paragraphs 327-330 reference is made to a Partnership Board meeting on 7 March
2016 and actions taken following that. The issue of CGI providing environments is raised.
However, I note that it is dealt with by it being recorded that “…this was an area of work
that would be ongoing throughout the life of the project” which does not suggest that this
was an area in which there would be supply by Agilisys of a single document at the outset.
It reflects the reality of the situation that there was bound to be fine tuning in requirements
during the lifetime of the project.
Page 125 ⇓
125
[437]       There was certain internal CGI documentation which was founded upon at
paragraph 332 and 333 of CGI’s written submission, and in particular two emails were
founded upon: one from a Chris Bennet on 4 November 2015, I did not hear any evidence
from this witness and one from Sebastian Wood in December 2015, who was also not led as
a witness. They both date to a period early in the project and do not appear to have been
followed up with Agilisys. They amount to very little, if this was a significant issue.
[438]       There is further support for the conclusion that there was no difficulty in respect to
provision of requirements by Agilisys and this is found in the responses at the time to
various Relief Notices.
[439]       RN001, reading short said:
“The responsibilities have not been met due to CGI not providing a fully functioning
ERP development environment which can be accessed by Agilisys users.” (which is
dated 27/1/16).
[440]       CGI’s response was to grant relief. If the issue of a single document setting out
specifications which was required from the outset had been a significant issue I cannot see
that relief would have been granted in terms of this Relief Notice. In the evidence of CGI’s
witnesses there is nothing which persuades me that there is any explanation as to why relief
was granted if there was a real issue from the outset regarding a lack of specification of
environments by Agilisys.
[441]       There is an associated issue and that is the payment by CGI to Agilisys to set up its
own temporary ERP development environment on Microsoft Azure to enable progress to be
made. CGI paid an extra £40,000 to Agilisys to carry this out (see: the evidence of
Mr Sheriffs at page 36 of the transcript).
[442]       If the environment problems were caused by the failure to provide a single
document at the outset setting out specifications it is difficult to understand CGI’s position
Page 126 ⇓
126
in doing this. Rather this supports the position that CGI had not put in sufficient resources
to produce the environments timeously.
[443]       RN013 relates to “ERP - delay of access to hosted environments”. It is dated 3 June
2016. The response to that was this:
“CGI have not provided hosted environments and have not confirmed a date by
which these will be accessible.”
“CGI agree that this delay has the likelihood of Agilisys not meeting the system
acceptance testing 11/7/16 date within ERP plan version 13 and is due to CGI cause
as described above.”
[444]       Again this response is wholly inconsistent with the position now being taken by CGI
and no plausible explanation was advanced for this inconsistency.
[445]       It was put to Ms Fox in cross-examination at pages 138 and 139 of the transcript of
her evidence that this would not have been CGI’s response to RN013 if Agilisys had not
provided the specification for environments. She says “Possibly yes … May have been”.
[446]       I believe the only accurate answer to that question is a straightforward yes. Her
explanation as to why the response to this was not: Agilisys did not provide a specification
of its requirements at the outset and therefore no relief will be granted” was this: “It was not
flagged to me to put it into a response”. This answer does not bear any close scrutiny. If
this was an issue then at June 2016, she must have known about it.
[447]       RN014 also related to delay of access to ERP environments and was dated 26 July
2016.
[448]       At page 114 in her cross-examination Ms Fox is asked about the initial response to
this Relief Notice. Ms Fox sent an email to other CGI colleagues, it was in response to
RN014, once more in this email there was no reference to there being a lack of an
environment specification. The explanation given for the lack of reference to this issue in
Page 127 ⇓
127
this document by this witness I believe does not hold water. If this was an important issue
surely it would be at the top of the list of matters mentioned.
[449]       Overall I am persuaded that looking to all the contemporaneous documentation the
position now being put forward by CGI is inconsistent with the position that it put forward
at the time and no plausible explanation is put forward explaining that. Its position is
inconsistent to its responses to Relief Notices; its response to the Ah-Wong email and in
general terms with the contemporaneous documentation which does not raise this issue of a
specification. I believe that this shows that the question of a specification of environments
being provided by Agilisys at the outset was not a significant issue at the time and that CGI
are attempting to rewrite history. Overall I am not persuaded that any failure to provide
environments on the part of CGI was caused or contributed to by any failure by Agilisys to
provide a design specification.
[450]       I have set out earlier my view as to the scope of the obligation to provide a technical
solution design. In terms of that limited obligation Dr Hunt does level certain criticisms at
Agilisys. She opines that at the outset the delivery of requirements was somewhat
piecemeal (see: her initial report at 6.5.4 and 6.8.6). However, I accept her conclusion that
this failure was not causative of anything other than minimal delay.
[451]       Moreover, insofar as evidence was led on behalf of CGI that any delay flowed from
this alleged failure I believe that the evidence either did not show that any delay had taken
place or alternatively if there had been a delay that it was minimal.
[452]       The two principal witnesses regarding this issue were Mr Hamilton and Mr Bryce.
[453]       First looking at Mr Hamilton, an issue was raised in his evidence by him about not
being aware that internet connections was a requirement in relation to the environments.
However, his attention was directed to documentation emanating from Mr Wood (another
Page 128 ⇓
128
CGI employee) and he appeared to know about the necessity for internet connections on
13 November 2015 (see: page 62). He further later accepted at page 88 that this had been
known about by CGI from 5 November.
[454]       Beyond that at page 84 he confirmed that the points which had been raised in the
snagging list by Agilisys in December 2015 were dealt with within a matter of a few days.
[455]       A further point which he had raised in relation to lack of specification of the
environments was that it was unknown as at December 2015 that Agilisys required Excel.
However he eventually accepted at page 87, lines 16 to 18 of the transcript that this had in
fact been known about on 11 November 2015.
[456]       His evidence when looked at as a whole did not support any significant delays
arising from a failure to specify environments.
[457]       Turning to Mr Bryce, at about pages 128 to 133 of his cross-examination Mr Bryce
made a number of complaints relative to the issue of lack of requirements and these were
related to a period as at 5 July 2016. However, these issues do not appear to have caused
any real delay.
[458]       He complained about not being informed prior to that about the need for BIF and
Excelerator, however, at page 140 of the transcript of his evidence he accepted that these
were known about as at 5 November 2015.
[459]       His evidence does not support any significant delay being caused by this issue.
[460]       The next issue raised by Mr Sandison which he submitted was a cause of delay was
lack of management on the part of Agilisys in respect to environments. I am unable to
identify with any precision what any such failure amounted to. Mr Sandison made a
criticism of Agilisys at paragraphs 289 to 291 of his submissions to the effect that Agilisys
had failed in its management responsibilities by directing CGI to Unit 4 for the details of the
Page 129 ⇓
129
environments. I observe that Unit 4 were installing in the environments the ERP solution,
namely: Business World. It seems to me in no sense surprising in those circumstances that
in respect to the installation of Business World Agilisys would direct CGI to Unit 4.
Business World was Unit 4’s proprietary product. In these circumstances it would be
entirely appropriate to direct CGI to Unit 4 for advice.
[461]       Lastly Mr Sandison referred to difficulties which were also caused by issues
experienced by Agilisys in accessing environments. CGI engaged Advance7, an external set
of consultants, to attempt to resolve these problems. Following Advance7’s
recommendations, performance issues were improved. Kevin Bryce’s evidence was that this
did not prevent Agilisys from proceeding with testing. Difficulties were also experienced at
an earlier stage with access by Agilisys to environments using the laptops used by Agilisys
contractors. Kevin Bryce’s evidence was that:
“the challenge that we had, which I hadn’t encountered before, was that the
individuals that Agilisys were using as contractors used their own device, so that
made the ability to allow people to connect remotely with a personal laptop to a
client network or, indeed, our own corporate network a bit of a challenge.”
Use of personal laptops posed a particular difficulty because CGI “had no ability to know if
they were patched, if they had antivirus software on, if they were running up-to-date
antivirus, and as soon as we allowed those to connect to our network it then introduced a
level of risk.” The contractors did not wish to use separate CEC laptops, as suggested by
CGI in December 2015, as this would have involved the Agilisys contractors having to carry
two laptops. The difficulty was not resolved until mid-2016, when the original CGI
suggestion of use of CEC laptops was ultimately accepted by Agilisys. Mr Sandison
accordingly argued that delay in access to the environments was caused by this issue and
was due to fault on the part of Agilisys.
Page 130 ⇓
130
[462]       In reply Mr Cormack said this: First, there was no contractual obligation (express or
implied) for Agilisys to provide some form of standardised laptop, whatever is meant by
that term. Further, there was no clear explanation or consensus on what a standardised
laptop would be or would involve.
[463]       I am unable to identify any evidence relied upon by CGI which sets out the
contractual basis for such an obligation nor in the evidence is it possible to identify what is
meant by “a standardised laptop”. Dr Hunt looks at this issue at paragraph 6.6.12 of her
initial report and I find her analysis of this issue compelling. If there was any party at fault
in respect to this issue it was CGI.
[464]       Mr Cormack secondly argued: the contemporaneous email evidence clearly shows
that Peter Hall was always willing to consider the use of either CGI provided or CEC
laptops if that was the only way to perform the Subcontract. Every time the issue was
actually raised by CGI, he responded positively: see in particular emails at JB 633, JB 878,
and JB 1281. The actual evidence led by CGI was often merely rumour and hearsay. For
example, Kevin Bryce suggested in his witness statement that there had been some form of
refusal to use CEC laptops in December 2015. But he had absolutely no direct knowledge,
had no details of it, and was unable to explain Peter Hall’s contemporaneous email to the
opposite effect. The highest the evidence got in relation to this issue for CGI was that Peter
Hall had expressed to Paul Hamilton a concern about having his contractors carry two
laptops and about software installed on their laptops being necessary for their work. An
expression of concern is not a refusal. And it is not anything remotely approaching a breach
of any contractual responsibility or obligation.
[465]       I find this line of argument persuasive. There simply was no body of cogent
evidence that Agilisys had refused either a CGI or CEC laptop. At its highest, Mr Hall had
Page 131 ⇓
131
raised a reasonable point about the necessity to use two laptops. That is not a refusal.
Again, to return to a point I have made on a number of occasions, if this was a real issue as a
matter causative of delay the contemporaneous documentation does not support this.
[466]       In conclusion, for the foregoing reasons, I do not accept that Agilisys failed to fulfil
its obligations in terms of the environments part of the project. CGI was unable to provide
environments in the form and at the time required as a result of their own failures.
[467]       From the evidence as a whole, Agilisys submitted the dominant and effective cause
of CGI’s failure to set-up the Environments was, plainly, an underestimation of the time,
resources, and effort that would be involved in setting them up.
[468]       Agilisys has prepared a detailed chronology of the events relating to the
Environments to which I earlier referred and I would reiterate that I believe having
considered this document it accurately sets out the timeline in respect of environments; and
accurately sets out the causes of the environments not being produced timeously by CGI
and supports the factual position put forward by Agilisys in the relevant Relief Notices.
[469]       Agilisys submitted in summary form, what can be taken from the chronology is as
follows. In the initial period from September 2015 December 2015, CGI appears to have
simply not understood the urgency of its obligations in relation to Environments within the
context of the Implementation Plan at the start of the Subcontract. This was, possibly, due to
a misunderstanding on their part about what would be involved when they entered into the
Subcontract. There is the typical high turnover of different people being involved in
September and October 2015 and then being replaced by others in November 2015. By early
November 2015, CGI had managed to arrange an installation from Unit 4. Sebastian Wood
of CGI, in particular, seems to have been the person responsible for meeting with Agilisys
and others for the purposes of producing High Level Designs.
Page 132 ⇓
132
[470]       In December 2015, access was provided but the development environment was not
fit for purpose. The major issue seeming to be that CGI struggled for an inordinate amount
of time to overcome an internal security protocol in relation to providing internet access. By
early January 2016, CGI had asked Agilisys to set-up their own temporary environments at
their cost, which was done. That allowed build work to commence. By late February 2016,
CGI had recruited Kevin Bryce as an environments manager to help with setting up the
remaining PPE and Production Environments. Between March 2016 and July 2016, CGI
experienced significant difficulties in deciding where to set-up servers, then setting-up the
servers, and then arranging Unit 4 to install the base software. It was, also, interrupted by
the data centre move at the end of June 2016. By mid-July 2016, it was recorded that the
environments were essentially unusable due to performance issues. This was investigated
by CGI and was not ultimately resolved until use of a workaround in September 2016. But
the workaround was not ever going to be fully acceptable for use in order to complete UAT.
[471]       The clearest evidence on the root cause came from CGI’s own contemporaneous
internal email from Niall Sheriffs of 24 April 2016 (JB 1069 at p 12724). In that email, he
expressly stated The real issue with the ERP project from our point of view is that we
under-estimated the effort to get the environment and data migration team resourced and
effective.Further, it is supported by the fact that from January 2016 June 2016, whether at
Partnership Board meetings or in responses to Relief Notices, CGI repeatedly acknowledged
that it was failing to meet its responsibilities.
[472]       This general underestimation seems to have been exacerbated by the very
bureaucratic and technocratic approach taken within CGI to provision of environments. The
contrast between the time taken for Agilisys to set-up their temporary environments in
January (a matter of a couple of weeks), and the time taken for CGI to set-up environments
Page 133 ⇓
133
is striking. This was explained by a number of witnesses to be due to a “matrix
management chain of command” with the GTO internal team that meant there was no direct
control between the CGI project team at CEC and the GTO unit: see Niall Sheriffs on Day 9
at p 13 of the Transcript. It was, also, accepted there were substantial lead times for
procuring work. And the evidence, amply, demonstrates the lack of consistency of people
being involved.
[473]       Whilst it is not intended to necessarily be a criticism that CGI are so formalistic and
bureaucratic, it should have been a significant factor in CGI’s thinking when it took on
responsibility for environments. By way of example, as Paul Hamilton explained in
cross-examination, the issue of internet connectivity in the Development Environment was
technically very simple. Indeed, he managed to provide it within a day of it being
identified. As he explained, the issue was it then took almost 2 months (from mid-December
2015 to mid-February 2016) for CGI’s internal security protocols to be resolved to allow a
permanent internet connection. He accepted this was a CGI issue: Day 11, page 69, of the
Transcript.
[474]       Given the original Implementation Plan allowed just over 2 months for the entire
completion of all of the infrastructure building, it is not clear how CGI intended to fulfil its
obligations in the Subcontract within the framework of such processes. But if, for example,
CGI needed to have highly documented requirements or specifications for the purposes of
operating its own internal procedures and the arms length arrangement with GTO, that is
not a failing on the part of Agilisys - CGI needed to tell Agilisys clearly that such
documentation needed to be provided for their purposes.
[475]       I accept the above submission for all of the reasons I have hereinbefore set out.
Page 134 ⇓
134
[476]       In conclusion CGI did not meet its obligations in relation to provision of
environments in respect of the dates in the Implementation Plan or any other agreed dates in
Project Plans. Its failure was not caused or affected by any breach of obligations by Agilisys.
[477]       In the questions posed to the court the issue of “suitable environment management
services” is raised. This issue does not appear explicitly to form a material part of Agilisys’
submissions. It is referred to briefly in CGI’s submissions at paragraphs 313 and 314. There
was almost no evidence on this issue. It appears from such evidence as there was (see:
paras 313 and 314) that CGI fulfilled its obligations regarding this.
[478]       The final question in respect of environments: did any breaches cause, or were they
reasonably likely to cause, Agilisys to fail to achieve milestones from MS2.5 onwards in the
ERP Implementation Plan by the effective dates? The effective dates in terms of the
Implementation Plan for MS2.5 were: start date 3 May 2016 and end date 26 July 2016.
[479]       Mr Sandison’s position was this in respect to the above question. Milestone MS2.5
founded upon by Agilisys was the commencement of UAT. He submitted that in any event,
delay to the provision of ERP environments (irrespective of the cause of that delay) was not
the operative cause of the failure to achieve UAT (and subsequent milestones) by the
required date. By the autumn of 2016, it was clear that serious issues in relation to data
migration to which he had earlier referred would prevent the achievement of UAT. Colin
Brown’s evidence was that, “The ERP programme would have been delayed due to data
migration issues, regardless of any environments issues.”
[480]       Mr Brown’s evidence was supported by the “Business World ERP Re-Plan Proposal”
dated 25 November 2016. Mr Brown’s evidence was that this was a ‘joint’ document,
produced by CGI and Agilisys. The document stated, Business World (ERP) will not
Page 135 ⇓
135
deliver to the planned go live of 3rd of April 2017. The reason for this is primarily due to the
Data Migration stream not running to plan.”
[481]       At 5.12.1 to 5.12.4 of his submission Mr Cormack said this:
5.12.1. The interdependence between data migration and UAT was repeatedly
acknowledged at ERP Project Board Meetings, Programme Highlight Reports, and in
correspondence between May 2016 and September 2016. It is also vouched by the
sequencing of data migration prior to UAT in the Project Plans (or, at least, in the
later ones that the bulk of data migration would be completed) and the original
Implementation Plan. Indeed, the Council raised concerns in October 2016 at the ERP
Project Board in relation to the fact that not all data was going to have been migrated
before UAT. Finally, it was expressly identified as a dependency for UAT in the
Project Initiation Document: see JB 147 at p. 17 of 30.
5.12.2. In the original Implementation Plan, User Acceptance Testing was due to
start on 3 May 2016 and conclude on 26 July 2016. It is obvious that, given CGI did
not even have access to the legacy systems until after the date for commencement of
UAT, then the failure to provide extracts was likely to cause Agilisys to miss this
deadline.
5.12.3. When relief was granted in May 2016 and by agreement to the 22 April 2016
Conditions and milestones moved to comply with version 13 of the Project Plan,
User Acceptance Testing was due to start on 10 August 2016 and conclude on 8
November 2016. Again, it is plain from the above that progress on extracts did not
even commence in earnest until around July / August 2016. And was not even
complete by the revised deadline in later Project Plans of 29 September 2016.
5.12.4 In all of these circumstances, it is plain that, along with environments, this
was a persistent and dominant cause of delay on the ERP project. And that, as a
direct consequence, Agilisys was reasonably likely to miss Milestone 2.5 for UAT
both in the original Implementation Plan and in relation to the postponed Milestone
dates accepted and agreed under other Relief Notices.”
[482]       It appears to me that there are two relevant causes of delay operating concurrently in
respect to the ERP project, namely: CGI’s failure in respect of environments and in respect
of data migration. It does appear that irrespective of delays in environments, data migration
would have stopped the achievement of Milestone MS2.5 and future dates. Equally delay in
provision of environments would, irrespective of delays in data migration, have stopped the
achievement of Milestone MS2.5.
Page 136 ⇓
136
ERP Biztalk: RN004 and RN006
[483]       Relief Notices 004 and 006 concerned the Biztalk development environment for the
ERP project. Biztalk environments were also required for the EI project. The provision of
the Biztalk environments for EI is not the subject of dispute in the present proof.
[484]       Relief Notices 004 and 006 asserted that the claimed CGI cause would cause Agilisys
to fail to achieve Milestone MS2.5: User Acceptance Testing (“UAT).
[485]       The issue in respect to these environments concerned interfaces. An interface is a
means of exchanging data between two systems or software applications. The BW software
required batch interfaces with various back office systems run by CEC (in particular the
Housing, Benefits, Bank, and Parking systems). In order to do so, it required the existing
interfaces (that were in place for the existing PPSL, Oracle, and iTrent) to be replaced with
interfaces to Business World via BizTalk: see, generally, Dr Hunt’s Report at paras 9.1.1
9.1.4.
[486]       The first question in relation to this group of Relief Notices was this: what were the
respective obligations of the parties (including as to the Milestone Dates in the
Implementation Plans and timing generally) in relation to the provision of experienced
Biztalk development resources and/or access to the Biztalk development environment?
[487]       Agilisys contended that the following responsibilities were allocated to CGI in terms
of the Subcontract.
Section
Extract
OBS 1, OBS 2, OBS 3,
OBS 30, and OBS 47
Provide experienced Biztalk development redevelop all required
interfaces into and out of the Unit 4 Business World solution
through the Biztalk middleware, using the Agilisys Biztalk
integration framework.
Manage the contractual relationship with BT enabling CGI to
work as its agent.” (via CEC)
Page 137 ⇓
137
Roles &
Responsibilities
document, Section C of
Schedule Part 4.1
Project Roles:
CGI providing data migration and integration services.
CGI Responsibilities & Deliverables:
Technical resource to define the technical specifications of the
Agresso interfaces, replacing all interfaces into and out of the
legacy systems and redesign of the integration with Capita Axis
income management system. The technical resource will be
appropriately skilled in Biztalk middleware and will be
responsible for all coding, testing and documentation to complete
the interfaces, working to the Agresso solutions architect.
Supplier management of existing applications that Agresso will
interface with, co-ordinating any support required to complete
integration testing. Key applications where Agresso will need to
feed data into are Northgate Swift and Capita Axis.
Agilisys contended that the Subcontract allocated the following responsibilities to Agilisys
in relation to the work on interfaces and integration on the ERP project.
Section
Extract
OBS 1, OBS 2, OBS 3,
OBS 30, and OBS 47
Working with CEC, define the technical specification of each
interface required into and out of the Unit 4 Business World
solution.”
Roles &
Responsibilities
document, Section C of
Schedule Part 4.1
Supplier Responsibilities & Deliverables:
Facilitate design workshops to agree the implementation design
of the Agresso system this will cover all business areas and
include interfaces, data migration and reporting. Deliverables
will include the workshop materials and the overall solution
design document.”
“Solutions architect resource which will generate the technical
solution design, interface strategy, data migration strategy, the
archiving solution and design of the supplier portal. Each of
these will produce a document deliverable for sign-off with the
Authority. The solution architect will manage the technical
resource provided by CGI.”
Page 138 ⇓
138
I consider that the above obligations were incumbent on parties. CGI in its written
submissions developed the following position in relation to obligations incumbent on
Agilisys:
98. Agilisys was under obligations as the leader of the ERP Programme. These
obligations were extensive, and included the provision of the solutions
architect resource to which the CGI responsibility makes reference. Agilisys’s
obligations with regard to the ERP programme are discussed more fully in the
‘Subcontract: roles and responsibilities’ section of these Written Submissions,
above. It is submitted that these obligations gave rise to two specific obligations
in the context of the Biztalk development environment:
a. As part of Agilisys’ overall responsibility for design and leadership, the
provision of adequate and detailed specifications for the Biztalk
development environment was an Agilisys responsibility.
b. Agilisys was responsible for managing the Biztalk developers provided
by CGI. This was in any event made clear in the drafting of the
Responsibilities Document: CGI was responsible for the provision of the
‘resource’, which was ‘working to the Agresso solutions architect.’
99. Accordingly it is submitted that CGI’s obligations in relation to the Biztalk
environments for ERP were subject to Agilisys’ obligation to lead the ERP
programme and, in particular, Agilisys’ responsibility to specify environment
design. There was no obligation on CGI to deliver the Biztalk environments by
a particular date.
[488]       The response of Agilisys was this:
The basic division of responsibilities was that Agilisys was responsible for
designing the integration strategy and identifying and defining the necessary
technical specifications of the interfaces. CGI were responsible for all matters relating
to the procurement and redevelopment of interfaces through Biztalk. This included,
by necessary implication, the creation of the Biztalk environments being the place
where that work would be undertaken.
[489]       I am satisfied that the above analysis by Agilisys is correct. It is pointed out by
Dr Hunt at paragraph 9.2.3 of her report: there is no specific responsibility within the
Subcontract for providing Biztalk environments. However, given CGI’s responsibility to
provide all infrastructure, hosting and support for the BW solution and the obligations as
identified in the OBS incumbent on CGI that is, as submitted by Agilisys, not surprising.
Page 139 ⇓
139
[490]       The only substantial difference between parties as regards obligations incumbent
upon them appeared to be in respect to the general management role of Agilisys and in
particular regarding the managing of the Biztalk developers provided by CGI. There is in
express obligation on the solution architect “to manage the resource provided by CGI.” and
in addition it is explicitly stated that the technical resource: “is working to the Agresso
solutions architect”. Reading these two provisions together there is a clear management
aspect to the role of the solutions architect in respect of the Biztalk resource. I have already
commented on the scope of the general management role of Agilisys. The scope of the
management obligations of the solutions architect has to be considered in terms of the
general management responsibilities and the roles and responsibilities as a whole imposed
on parties in respect of this area of the project. In particular it has to be construed in the
context of CGI’s obligation to “redevelop all required interfaces” and “CGI providing data
migration and integration services”. Thus the management role cannot empty these
responsibilities of any content. It seems to me that a sound construction of the obligations of
the solutions architect is this: she was to be the principal person within CGI to initiate the
management functions identified by Dr Hunt, which I set out earlier and in addition was
obliged to provide such assistance as required to the Biztalk resource and such assistance as
requested by the Biztalk resource. CGI’s position appeared to be this. It was a failure in this
responsibility of management if, first, environments were not provided in conformity to a
timeline and, second, if the environments were not of the appropriate quality. That is a clear
example of emptying the responsibilities of CGI and rendering them devoid of content. It is
creating, by use of a management obligation, an obligation on Agilisys which amounts to a
guarantee that CGI performed its responsibilities in terms of the Subcontract. I reject this for
all of the reasons I have outlined in the course of my opinion.
Page 140 ⇓
140
[491]       As regards timing, no express date is given in respect to the delivery of Biztalk
Environments. However in terms of the Implementation Plan at Schedule Part 6.1 the build
period for interfaces and integration was 1 December 2015 to 30 June 2016. Thus in order for
Agilisys to meet these dates the interfaces necessarily had to be provided well before that
latter date. By issuing AG003 on 13 May 2016 CGI accepted that it had not produced the
Biztalk environments within the time necessarily required by the Subcontract as set out in
the Implementation Plan.
[492]       Having considered the responsibilities of parties in respect to this part of the project,
the next question is: did CGI breach its obligations by failing to provide access to Biztalk
development resources and/or access to the Biztalk development environment?
[493]       Mr Cormack submitted in respect to this issue:
6.2.2. Insofar as Biztalk environments are concerned, the initial background was
neatly summarised in the email of 17 February 2016 from Bob Price to Bridget Fox,
and the attached slide: JB 820. From an early point in October 2015, Agilisys
requested that CGI provide the necessary Biztalk infrastructure and were working on
providing it. It suffered identical problems as the ERP Development Environment
and EI Environments; namely, despite Agilisys having identified the need for an
internet connection as early as October 2015, it took an inordinate amount of time for
CGI to overcome its own security protocols on this issue. This is neatly captured in
an internal CGI email in December 2015 discussing the two Biztalk developers that
had been employed by CGI. They were complaining about the lack of work. As
Richard Salmond, candidly, says: ‘They have been waiting for the development
environment to be available before they could do any development. This has been “it
will be ready tomorrow” for weeks and should now be available.” Despite earlier
escalations within the project, this was an on-going delay when RN004 was issued on
10 February 2016. The issuing of RN004 led directly to AG003, which is discussed
below.
6.2.3. Insofar as Biztalk resource is concerned, Agilisys’ principal complaint was
that, as a result of the general re-plan of ERP that took place in late 2015, the time
period for Biztalk development had been compressed and, accordingly, more
resource was required. This was the fundamental concern when RN004 was issued.
[494]       RN004 was issued in direct response to the fact that: (a) the Biztalk development
environment was not set-up; and (b) in his email of 10 February 2016, Barry Speed of CGI
Page 141 ⇓
141
had expressly said to Peter Hall of Agilisys that CGI had insufficient resource to cover both
the EI and ERP requirements: JB 824.
[495]       The sequence, after RN004, was that on 18 February 2016 CGI asked Agilisys to
provide a quote to provide the Biztalk resource: see JB 826. Kevin Bryce of CGI was
involved in setting-up the Biztalk Environments and was initially targeting a delivery date
of 14 March 2016: see JB 908. This date was missed and led to the issuing of RN006 on
18 March 2016. And, thereafter, the Biztalk environment was eventually made available to
Agilisys around mid-April 2016: see Kevin Bryce WS at para 2.24- 2.25.
[496]       Eventually, CGI formalised the informal agreement reached in early March 2016 and
granted AG003 on 13 May 2016: see Bridget Fox WS at para 3.9 and see AG003 in JB 983.
The effect was to transfer the responsibility for interface and Biztalk development to
Agilisys.
[497]       The above sequence of events appears to confirm the failure to provide on the part of
CGI sufficient Biztalk resources and the relevant environment within the time necessary for
the Implementation Plan. Mr Cormack submitted that the effect of AG003 was to put CGI in
immediate breach of its obligations, which perhaps reflects an acceptance they had not
properly performed their responsibilities. This is because AG003 updated CGI’s obligations
to a straightforward obligation to provide the required Biztalk infrastructure, any necessary
technical support and environment management services, for a minimum of
3 environments, by the 1st of March.However, as can be seen from the sequence of events,
the Biztalk Environments were not provided.
[498]       I agree with this analysis of the effect of AG003. I further agree with the submission
of Mr Cormack at 6.3.1 of his written submissions where he argues the root cause or reason
for CGI’s failure was its own inability to correctly resource the Biztalk development work
Page 142 ⇓
142
within any sort of reasonable period. In my view the evidence of CGI’s witnesses essentially
confirms it: see para 3.85 and 3.101 of Niall SheriffsWS, and para 3.6 of Bridget Fox’s WS.
Both sections of Mr Sheriffsstatement explicitly accept that CGI had a resources issue as
does the referred to section of Ms Fox’s statement. This, it appears to me, has nothing to do
with a lack of management by CGI but is purely and simply a failure on CGI’s part to
provide the appropriate resources.
[499]       In addition to the above it took an inordinate length of time to overcome issues with
internet connectivity, set up of and providing access to the Biztalk environments.
[500]       In my view the above position is entirely supported by the analysis of Dr Hunt
between 9.3.8 and 9.3.15 of her initial report, where she clearly sets out the timeline in
respect to this aspect of the case and analyses the root cause of the difficulties and delay.
[501]       I have considered the written submissions made by CGI in relation to this question,
in particular at paragraph 108. There is nothing therein which causes me to depart from my
view in respect to this question. I cannot see how in the light of AG003 CGI is not in breach
of its obligations.
[502]       The next question is: was CGI’s breach of its obligations in respect to Biztalk caused
or contributed to by Agilisys’ breach of its obligations. CGI’s position is that it was caused
or contributed to by Agilisys’ breach. This position was founded on first a failure to provide
an adequate specification. My attention was directed to the evidence of Mr Sheriffs at 3.84.1
to 3.84.3 of his witness statement. However, in the course of his cross-examination, at least
to a material extent, he accepts that prior to 10 December 2015 (the date about which he is
talking in this part of his statement) there had been notification by Agilisys to CGI of the
requirements he mentions at 3.84.1 to 3.84.3 of his statement. In particular internet
connection was identified by Agilisys in October 2015.
Page 143 ⇓
143
[503]       The only other reference to a lack of specification relied on by CGI is an email of
18 April 2016 from Mr Olejnik of CGI to Agilisys. This does refer to a problem relative to
specification by Agilisys. In the first part of the email there is a general criticism of the
specification provided by Agilisys but no detail is provided as to what were the problems,
how significant they were and how long the issue took to be dealt with. I did not hear any
evidence from Mr Olejnik. The above is completely lacking in specification and does not for
that reason establish that there was a material problem with specifications being provided
by Agilisys. The second half of the email refers to one further requirement being specified,
again there is no detail regarding this issue. There is insufficient in this document as a
whole to establish that delay in providing requirements by Agilisys was causative of delay
in the provision of resources to Biztalk and or access to the Biztalk development
environment. Once more if there had been a significant problem in the provision of
requirements the contemporaneous documentation would have been much more
voluminous than this and I would have been referred to such by CGI. The
contemporaneous documentation, as I have said, tends to support the conclusion that the
effective cause of the delay was lack of resources being provided by CGI and difficulties in
access.
[504]       Once more, criticism is also levelled in respect to Agilisysfailure to manage the
developers (see para 118 written submission CGI). This raises the issue of failure to manage
timelines and quality. I have already to some extent dealt with this issue earlier. I would
add this: if the obligation to manage required management relative to timelines, I am not
aware from the evidence in what way Agilisys failed to manage so that timelines were not
adhered to. The problem was lack of resources provided by CGI. I cannot see how any
management by Agilisys could affect this issue. Nor how it could affect any issue of access.
Page 144 ⇓
144
Equally in respect to quality I do not see in what way any management in respect to
“quality”, whatever that might mean, had any causative impact given the root causes I have
identified for the failures. In conclusion, it is wholly unclear in what way Agilisys failed to
manage the developers. Secondly it is wholly unclear how this led to any delay. I do not
find this line of argument persuasive.
[505]       Thereafter, the final question is this: did any breaches cause, or were they reasonably
likely to cause, Agilisys to fail to achieve the Milestone MS-2.5 in the ERP Implementation
Plan by the effective date?
[506]       CGI’s position in respect to this was as follows:
The UAT milestone was not dependent upon the provision of Biztalk resource and
environments alone. In particular, the achievement of UAT in May 2016 was
dependent on other aspects of the programme, including payroll, data extracts and
data migration. For the reasons set out in relation to those relief notices, the failure of
Agilisys to meet its obligations in those areas would have prevented the achievement
of UAT in any event. It is accordingly submitted that any breach by CGI in respect of
question 6 was not in any event the operative cause of the failure to achieve UAT.
[507]       The effective date for UAT had a start date of 3 May 2016 and end date 26 July 2016
per Schedule Part 6.1. Biztalk environments became available in April 2016. In the period
up to April 2016 there were three concurrent causes operating: the wider issue of
environments as earlier discussed in this opinion; the issue of data migration as earlier
discussed; and, lastly, the issue of the failure to provide the Biztalk environments.
Thereafter for the reasons I have already set out the dominant causes were the issues of data
migration and the wider environment issues. Thus the three causes would have had a
cumulative effect on the achievement of this Milestone.
Page 145 ⇓
145
Relief Notice 008 ERP Interfaces
[508]       Relief Notice 008 concerned the provision of technical specifications for interfaces to
existing feeder systems. It was issued by email on 21 March 2016. Agilisys asserted that a
failure to provide those specifications caused Agilisys to fail to achieve Milestone MS2.5 and
relates generally to relief from delay under version 12 of the Project Plan.
[509]       The particular matter in relation to interfaces which was at the core of the dispute
concerning interfaces was: interface specifications. These are helpfully defined in the
following section of the written submissions of Agilisys:
7.2.2. This section relates to the specific batch interface specifications that were
required as part of the redevelopment of the required interfaces to the legacy systems
(in email correspondence they, and the legacy systems, are often referred to more
colloquially as ‘feeders’, ‘feeder files’, and ‘feeder output documents’). The basics of
this concept are, helpfully, described at: paras. 6.13 and 8.49 of Peter Hall’s WS,
para. 5.5 of Ravinder Johal’s WS. As there explained, interface specifications are
Microsoft Word documents that detail the format and content of the output of data
from the legacy systems. They are critical in order for Agilisys to develop a
specification of how that data output can be translated into an acceptable format for
BW. This specification was then required to allow the CGI Biztalk developers to
write programming code in Biztalk to complete their work.
[510]       The first question in respect to this section is: what were the respective obligations of
the parties (including as to the Milestone Dates in the Implementation Plans and timing
generally) in relation to the provision of a complete set of Agresso technical interface
specifications? Agilisys contended that the following responsibilities were allocated to CGI
in terms of the Subcontract in relation to the work on interfaces and integration on the ERP
project:
Section
OBS 1, OBS 2, OBS 3,
OBS 30, and OBS 47
Extract
Provide experienced Biztalk development redevelop all
required interfaces into and out of the Unit 4 Business
World solution through the Biztalk middleware, using the
Agilisys Biztalk integration framework.
Manage the contractual relationship with BT enabling
CGI to work as its agent.” (via CEC)
Page 146 ⇓
146
Roles & Responsibilities Project Roles:
document, Section C of
Schedule Part 4.1
CGI providing data migration and integration services.”
CGI Responsibilities & Deliverables:
Technical resource to define the technical specifications of
the Agresso interfaces, replacing all interfaces into and out
of the legacy systems and redesign of the integration with
Capita Axis income management system. The technical
resource will be appropriately skilled in Biztalk
middleware and will be responsible for all coding, testing
and documentation to complete the interfaces, working to
the Agresso solutions architect.
Supplier management of existing applications that
Agresso will interface with, co-ordinating any support
required to complete integration testing. Key applications
where Agresso will need to feed data into are Northgate
Swift and Capita Axis.
As prime contractor/partner with the Authority, leverage
commercial powers to ensure that the Authority fulfil
obligations such as timely decisions, participation in
workshops and testing in accordance with the
implementation plan) and ensure approval of designs for
EI are not unreasonably withheld.
Annex 2:
CGI shall procure that the Authority will comply with
the duties listed below, as required to deliver the Authority
functions listed in Schedule 2.1 (Service Description):
Provide business specifications for interfaces and reports
Request the Incumbent Supplier to provide data extracts
in accordance with specifications provided by the
Supplier.
Page 147 ⇓
147
The Subcontract allocates the following responsibilities to Agilisys in relation to the work on
interfaces and integration on the ERP project:
Section
Extract
OBS 1, OBS 2, OBS 3, Working with CEC, define the technical specification of
OBS 30, and OBS 47
each interface required into and out of the Unit 4 Business
World solution.
Roles & Responsibilities
document, Section C of
Schedule Part 4.1
Supplier Responsibilities & Deliverables:
Facilitate design workshops to agree the implementation
design of the Agresso system this will cover all business
areas and include interfaces, data migration and reporting.
Deliverables will include the workshop materials and the
overall solution design document.
Solutions architect resource which will generate the
technical solution design, interface strategy, data
migration strategy, the archiving solution and design of
the supplier portal. Each of these will produce a document
deliverable for sign-off with the Authority. The solution
architect will manage the technical resource provided by
CGI.
[511]       CGI relied in particular on the general obligation in the responsibilities document
that: “Agilisys was responsible for client management and governance (project level).” It
also relied on the solutions architect’s responsibility to prepare an interface strategy. It
submitted that these obligations had been breached.
[512]       I consider that the obligations above referred to by each side in respect to this Relief
Notice were incumbent on the parties and were relevant to the discussion of the various
issues relative to this Relief Notice.
[513]       The position of CGI in summary was this: the delays to the provision of interface
specifications were caused by a combination of the delay in BT responding to requests
(which was a matter outwith the control of CGI, and which could only be resolved by
Page 148 ⇓
148
escalation through CEC) and a delay in obtaining approval from CEC. Agilisys were
responsible for CEC management and governance at a project level. Responsibility for
resolving these delays were with Agilisys, rather than with CGI, as a result, there was no
breach of CGI’s contractual responsibilities in relation to the provision of interface
specifications.
[514]       There was in essence no dispute between parties as to what was the factual cause of
the delay (see: 7.3.1 to 7.3.3 of the submissions for Agilisys and paragraphs 250 to 257 of the
submissions of CGI).
[515]       The dispute was entirely a legal one as to whose was the legal responsibility to deal
with these issues, namely: the issue with BT.
[516]       I would observe regarding this legal issue as follows:
[517]       First, it was for CGI to “manage the contractual relationship with BT enabling CGI to
work as its agent”.
[518]       Secondly: “CGI shall procure that the Authority will comply with the duties …
Request the Incumbent Supplier (BT) to provide data extracts in accordance with
specifications provided by the Supplier.” “…provide business specifications for interfaces
and reports.”
[519]       Thirdly, it was CGI’s responsibility to “leverage commercial powers to ensure that
the Authority fulfil obligations…”
[520]       Fourthly: “Agilisys was responsible for client management and governance (project
level)”. Having regard to these four responsibilities I am persuaded that the principal and
the ultimate responsibility in obtaining the specifications from BT fell on CGI. It was its
responsibility to procure that CEC complied with its responsibilities. I have already
discussed the nature and scope of that responsibility. It was CGI who had the responsibility
Page 149 ⇓
149
to exercise its commercial powers. It was to manage the contractual relationship with BT.
These responsibilities sat within the context of Agilisys playing a subordinate role to CGI for
the reasons already set out. When these responsibilities are read together the responsibility
for client management on the part of Agilisys is clearly a subordinate one. It was given the
difficulties encountered with BT and CEC for CGI to exercise its power to procure CEC’s
compliance in order to resolve the delay issues and to exercise its financial leverage in
respect of CEC. The principal and ultimate responsibility to obtain the specifications was
incumbent on CGI. That at a practical level it was a failure by CGI is illustrated by this
passage in Mr Sheriffs’ evidence.
At an operational level, Agilisys were able to interact through the weekly meetings
we had with BT, and with the progress through the tracker of issues. There was no
direct means for management escalation and interaction from Agilisys, or from CGI,
as BT would only deal with CEC at that level.” (Supplementary statement of
Mr Sheriffs at 4.11)
[521]       This passage of evidence illustrates the importance of CEC in obtaining the
specifications and it was CGI who had the power to procure CEC complied with its
responsibilities.
[522]       So far as an interface strategy being prepared by the solutions architect there was I
believe no evidence before the court that the non-existence of such a strategy had any impact
in relation to the issues of obtaining specifications from BT.
[523]       CGI in its written submissions raised the issue of no date being specified in the
contract as to when the specifications were to be provided.
[524]       The specifications were held by BT. In respect to the issue of timing it is clear from
the nature of interface specifications that they were required at the outset in order to allow
work to progress on ERP interfaces. Agilisys submitted as follows as regards the issue of
timing:
Page 150 ⇓
150
“Agilisys requested the interface specifications at an earlier point. They had, in fact,
made it clear that there were going to be required pre-contract and as early as
27 February 2015: see the Expert Report of Dr Hunt at para 3.2.4 (footnote 14)
referring to JB 402. They are recorded as already being delayed in the Weekly
Highlights Reports for the project around October / November 2015, they by January
2016 they were recorded as severely delayed, CGI failed to obtain the majority of
them until well after April 2016, and indeed some minor work was on-going on them
still in February 2017. It is submitted the factual position is entirely accurately
recorded in RN008.”
[525]       Taking into account the explanation given by Agilisys as to when interface
specifications were needed by Agilisys they were not provided within the time necessarily
required by the Implementation Plan. By the time of the issue of the Relief Notice it was
accepted by CGI that only 34 from 50 specifications had been provided and the rest were not
provided until 1 April 2016.
[526]       Having held that the responsibility for resolving the delays lay with CGI, and my
above view as to when these specifications were necessarily required in terms of the
Implementation Plan, there was a breach of CGI’s contractual responsibilities in relation to
the provision of interface specifications. It did not provide the full specifications in the time
reasonably required by the Implementation Plan. The root cause of its failure was the
difficulty in obtaining the information from BT.
[527]       The next question is: was CGI’s breach caused or affected by a breach by Agilisys of
its obligations? For the reasons given above my answer to this question is no. I am unable
to identify any evidence which would support a conclusion that CGI’s breach was caused or
affected by a breach by Agilisys of its obligations.
[528]       The final question in respect to these Relief Notices is this: did the breach by CGI
cause, or was it reasonably likely to cause, Agilisys to fail to achieve the Milestone MS-2.5 in
the ERP Implementation Plan by the effective date?
Page 151 ⇓
151
[529]       CGI’s position in respect to this question was this: RN008 is founded upon an
anticipated failure to achieve UAT. For the reasons given in answer to other questions, it is
submitted that the operative cause of delay was not any failure by CGI to discharge its
obligations in respect of interface specifications. UAT was not achieved following the
provision of the missing specifications on 1 April 2016. The reason for this was not CGI’s
failure to provide specifications, but rather was Agilisys’ failures to discharge its obligations
in respect of data migration and environments.
[530]       I have already held that Agilisys did not fail to discharge its obligation in respect of
data migration and environments. Accordingly my answer to this question is yes. It was a
concurrent cause along with the wider issues of data migrations and environments more
generally in relation to which I have already identified CGI cause.
RN002 and RN012: EI environments
[531]       RN002 was issued to CGI by email on 1 February 2016. It was concerned with the
provision of development, test and production instances of Northgate iWorld to allow build
and configuration of the Agilisys Digital module and to provide test data to allow proving
and testing of the configured modules.
[532]       Relief Notice 012 was concerned with the provision of the infrastructure required to
allow Systems Integration Testing to commence. Systems Integration Testing was required
in order for Milestone MS-OSC1.1 to be achieved. Milestone MS-OSC1.1 was the Go-Live of
the AD platform, myGovScot integration and initial transactions.
[533]       Turning to look first at RN002, the first question in respect to this Relief Notice was
this: what were the respective obligations of the parties (including as to the Milestone Dates
in the Implementation Plans and timing generally) in relation to the procurement,
Page 152 ⇓
152
installation and provision of access to a development, test and production instance of the
Northgate API and the test data?
[534]       Agilisys contended that the Subcontract allocated the following responsibilities to
CGI in relation to third parties:
Section
OBS 21
Extract
Technical architecture working in partnership with
Agilisys”
Procure and manage the contractual relationship with all
3rd parties providing support across the Enterprise
Integration platform”
Procurement of required 3rd party APIs and supplier
portals as required to deliver the online transactions to
meet the requirement.”
“Installation of 3rd party portals and APIs for AD
integration and provision of development and test
environments of 3rd party portals and APIs”
NB. The term ‘Third Party Suppliers’ is defined in
OBS 21 as ‘Third parties provide solutions which
comprise functional components of this service, which may
be hosted in the Authority estate, or externally.’
Roles & Responsibilities
document, Section C of
Schedule Part 4.1
Project Roles:
CGI responsible for negotiation of subcontract, licence
and payments to line of business application portals or
APIs (CGI have put aside budget for APIs) liaise with
Authority & BT
CGI responsible for contract management and delivery of
the 3rd parties across the current Authority Applications in
scope of the Enterprise Integration suite e.g. the MDM
infosphere product, CRM, LPG, Payment engine etc…
CGI Responsibilities & Deliverables:
Procurement of required 3rd party APIs and supplier
portals as required to deliver the online transactions
outlines within the OBS.
Installation of 3rd party portals and APIs for AD
Page 153 ⇓
153
integration and provision of development and test
environments of 3rd party portals and APIs.
As prime contractor/partner with the Authority, leverage
commercial powers to ensure that the Authority fulfil
obligations such as timely decisions, participation in
workshops and testing in accordance with the
implementation plan) and ensure approval of designs for
EI are not unreasonably withheld.
[535]       Agilisys contended that the Subcontract allocated the following responsibilities to
Agilisys in respect to third parties:
Section
Extract
OBS 21
Replace the online transactions with the AD product
with modules as required to deliver the functionality as
outlined within the OBS
Roles & Responsibilities
document, Section C of
Schedule Part 4.1
Project Roles:
The Supplier responsible for defining 3rd party
requirements including APIs or online portals (GovTech,
Oracle, Xerox, Pitney-Bowes, APP, Northgate, Idox)
Supplier Responsibilities & Deliverables:
Replace the online transactions with the AD product
with modules as required to deliver the functionality as
outlined within the OBS
[536]       Lastly Agilisys contended that the Subcontract allocated the following
responsibilities to CGI in relation to environments (including Biztalk environments) on EI:
Page 154 ⇓
Section
OBS 21
154
Extract
Provision of dev/test environments that are on the same
version of the software/environments as the production
system of Line of business applications for AD integration
development and training.
Infrastructure hosting and support of the Enterprise
Integration platform apart from Oracle RightNow SAAS
provision.
Provide Biztalk infrastructure to replace the WebSphere
middleware product
Redevelop the Websphere middleware rules and
translation maps to Biztalk with experienced Biztalk
development capability.” (and a slight variation on
these words elsewhere)
Roles & Responsibilities
document, Section C of
Schedule Part 4.1
Project Roles:
CGI providing infrastructure and connectivity services
and security management covering IT security
management plan and data protection etc. (The Supplier to
feed into plans)
CGI providing biz talk integration licences and technical
training to CGI development and support team.
CGI providing BizTalk Resources to redesign all
integration through websphere to go through Biztalk
where required”
CGI Responsibilities & Deliverables:
Infrastructure hosting and support of the Enterprise
Integration platform apart from Agilisys Digital and
Agilisys Engage products hosted by Agilisys.
Provision of dev/test environments that are on the same
version of the software/environments as the production
system of Line of business applications for AD integration
development and training.
[537]       The particular responsibility CGI emphasised within the above was “The Supplier
responsible for defining third party requirements including APIs…”.
Page 155 ⇓
155
[538]       Moreover, CGI in its submissions stressed Agilisys’ responsibility as the manager of
the EI programme and the general responsibilities of management and leadership. It was
CGI’s submission that these management/leadership responsibilities had been breached. I
have earlier set out my views regarding the scope of the general management responsibility.
[539]       I consider that the obligations referred to by each side in respect to these Relief
Notices as set out above were incumbent on the parties and were relevant to the discussion
of the various issues relative to these Relief Notices.
[540]       In summary CGI’s position became this:
Accordingly it is submitted that whilst CGI had specific obligations under OBS 21 in
respect of API provision, these responsibilities were subject to Agilisys’s overall
responsibility for the EI project and for the AD product. In particular, they were
subject to Agilisys’ responsibilities firstly to specify the time at which the
dependencies were required and secondly to specify properly the design of the
environments.
[541]       In respect to design, the obligation of Agilisys was explicitly confined to it being
“responsible for defining 3rd party requirements including APIs or online portals”. This
obligation has to be understood in terms of the general nature of the AD product. This is set
out by Claire Conaghan in her witness statement between paragraphs 8.4 to 8.6. What
importantly can be taken from the foregoing description is this: the AD product has certain
services fully integrated as built in modules, which require connection through APIs to
CEC’s system. In other cases the AD has no built in module and here the product relies on
integration with either SSOs or APIs. Thus it is not the position that CGI were required in
terms of the contract to develop and build bespoke APIs. Rather in terms of the Subcontract
at OBS 21 CGI were to take a tiered approach using (1) published APIs; (2) failing which
SSOs; and (3) failing which electronic forms. Thus Agilisys’ responsibility on a proper
understanding is to advise CGI as to the APIs necessary and not to give a specification on
how to build the API.
Page 156 ⇓
156
[542]       I consider so far as timing is concerned if party one is depending on another to fulfil
a particular aspect of its obligation in order to start work in terms of its obligations then the
first party has a duty to advise the second party at what time the dependency is required
assuming that is not clear from the nature of the work itself.
[543]       Did CGI breach its obligations by failing to provide access to a development, test and
production instance of the Northgate API and test data? Agilisys helpfully produced a
timeline in relation to this issue which I have appended as Appendix 3 to this opinion. On
reviewing this document I believe it accurately sets out the relevant chronology and the
factual basis for the failure.
[544]       The above chronology in my view clearly establishes a breach by CGI. The
Northgate API and associated environment were requested in September 2015 to support
Agilisys’ build activities for Drop 1/MS-OSC1.1, but were not actually delivered until
December 2015. Once delivered, the data anonymisation in the cloned environment was so
severe as to make it impossible for Agilisys to test the API properly, and that issue was not
resolved until late January 2016. This was most clearly spoken to by David Buchanan in his
WS and in his evidence in court.
[545]       It was submitted by Agilisys that in the original Implementation Plan in Schedule
Part 6.1 of the Subcontract, the period allowed for build was 21 September 2015 to
26 February 2016. As David Buchanan explained in his WS, “This [access to Northgate
iWorld API] was fundamental to the work to be undertaken by the AD team and having the
access was necessary for around 80% of the build work.That is unsurprising as Northgate
was the line of business system for the Revenues, Benefits, and Housing element of AD that
was to make up the majority of the substance of Drop 1.
Page 157 ⇓
157
[546]       I am persuaded by the above submission and accordingly CGI breached its
obligations. It did not produce the Northgate instance and test data in the time necessarily
required by the Implementation Plan.
[547]       The next question is whether CGI’s breach was caused or affected by a breach by
Agilisys of its obligations. The first issue raised by CGI is a failure to specify the time at
which the dependencies were required.
[548]       It does not appear from the evidence that specific dates were given by Agilisys.
However, I accept the evidence of David Buchanan and Claire Conaghan referred to in
Appendix 3 that Northgate API was “fundamental” and “critically important” and it
therefore should have been obvious to CGI that Northgate API would be required at an
early stage. I do not think that Agilisys breached its obligations to advise as to when the
specifications would be required.
[549]       CGI in its written submission in respect to the issue of lack of specifications relied on
certain parts of the evidence from Richard Salmond. His evidence was to the effect that
there was a fundamental failure in this regard. It is noteworthy: (1) he refers to not a single
contemporaneous document in which this issue was raised with Agilisys; (2) there is no
reference to the issue being escalated; and (3) there is even a lack of internal
contemporaneous documentation which is referred to. One email is referred to (see:
supplementary statement 2.7). However, I do not find it surprising that one email was sent
regarding specification of APIs. It would be surprising if no questions had to be asked of
Agilisys. As is accepted by CGI certain of these questions were answered by reference to
architecture documentation, again suggesting there was no significant issue.
[550]       I find it is noteworthy that the evidence of Mr Salmond is relied on by CGI. As he
points out in his witness statement: “In general my involvement in both EI and ERP was
Page 158 ⇓
158
peripheral and was focused on changes to the business” (para 4.2 of his witness statement).
His lack of direct involvement means that his evidence such as it is on this issue is of little
value.
[551]       In the whole circumstances, given the lack of contemporaneous documentation and
his peripheral role I am not prepared to accept Mr Salmond’s evidence on this issue. It is
also noticeable that his evidence is wholly lacking in detail as to what questions were not
answered and as to when they were asked and what effect, if any, the failure to provide
specifications caused.
[552]       In addition I accept the analysis of Dr Hunt in relation to this issue. It is once again a
well-reasoned analysis of the position (see: 9.25 to 9.3.7 of her original report).
[553]       The final question in respect to RN002 is this: if the answer to the above question is
yes, did that breach cause, or was it reasonably likely to cause, Agilisys to fail to achieve the
Drop 1 Milestone MS-OSC 1.1 in the EI Implementation Plan by the effective date? CGI’s
position was that the root cause was the failure to give adequate notice in respect of the EI
environment.
[554]       Agilisys’ response to this argument was:
“8.8.1 The Court is invited to conclude that the two dominant causes of Agilisys’
inability to meet the originally planned Milestone of MS-OSC 1.1 in the
Subcontract were both the Northgate API issue and the Environments
(particularly in this context the Biztalk Environments). These were, it is
submitted, the dominant causes operating concurrently in the period
September 2015 January 2016.
8.8.2
It is, further, submitted that (whilst the Northgate API issue was resolved by
late January 2016) the dominant causes of delay in the period after January
2016 continued to be the unavailability of the Biztalk, Development, and Test
Environments (also joined by the Production and Pre-Production
Environments as outlined in RN010). As RN010 clearly stated: ‘Agilisys has
made it clear that it is in a position to deliver all required functionality for milestone
MS-OSC1.1 with the exception of the Biztalk integration which is dependent on
CGI delivering on its Biztalk responsibilities.’ As will be recalled, CGI accepted
RN010 and granted relief on 20 May 2016.”
Page 159 ⇓
159
[555]       I am persuaded that the analysis of Agilisys is correct. CGI had a further position in
relation to the final question regarding RN002 to the effect that in any event the relevant
milestone would not have been achieved owing to the inadequacy of Agilisys’ AD product.
I will discuss this question when considering RN012.
[556]       Turning to RN012 I first want to deal with a preliminary argument. As I said earlier
there was a specific argument advanced by Mr Sandison in terms of RN002 and RN012
based upon his position that changes to the contract which did not follow change control
were invalid. His position was this:
The Transformation Plan provided that the Milestone was to be achieved by 1 April
2016. Following the delays to the provision of the EI development environment
caused by the failure of Agilisys to specify requirements with sufficient time, the
Go-Live date was moved to 29 April (discussed at §74 above). On 18 February 2016,
Agilisys proposed a revised plan following the delay which split the Go-Live into
two separate ‘drops’ in June 2016. A version control note signed on 28 April 2016
(which was stated to have no contractual force) scheduled drop 1 for 16 June and
drop 2 for 30 June. This plan was subject to further (non-contractual) variations. By
September 2016, the EI project had been split into six separate drops (also referred to
as ‘phases’), with the first (drop 1) to take place on 24 November 2016) and the last
(drop 6) to take place on 23 March 2017. These drops, and their contents, were
summarised in an ‘EI phase timeline’ document issued on 7 September 2016. There
was, however, no contractual variation (utilising the Change Control procedure) to
the single milestone in the Transformation Plan.
[557]       Given the views I have earlier expressed, the division of the EI project into drops did
not produce a valid alteration to the Subcontract terms, there was only a single milestone.
[558]       There was a second preliminary issue regarding this group of Relief Notices, namely
this: an objection to certain evidence led by CGI was taken by Agilisys. This in particular
related to the evidence led by CGI about what were described as strategic issues with the EI
Project [these strategic issues related to alleged inadequacies with the AD product.] The
objection as developed came to this:
“8.9.2 The thrust of all of CGI’s witnesses that gave any detailed evidence directly
about the EI project was to focus on the strategic issues that came to a head
Page 160 ⇓
160
around November 2016. It is submitted that these issues are quite simply
utterly irrelevant.
8.9.3
This is, principally, because Agilisys did not issue or seek to rely on any
Relief Notices relating to the deferral of Drop 1. The Relief Notices relied
upon in relation to the EI project were all related to the delay that caused the
prior deferrals to the Go Live date of November 2016.
8.9.4
If the submission is that it is relevant because it shows that Agilisys would
never have been able to meet the Milestone, it is flawed because: (a) it is
misconceived as a matter of law because it involves judging Relief Notices
retrospectively, which is the same fallacy discussed in Walter Lilly & Co
Limited v. Mackay & Anor [2012] EWHC 1773 at para. 365 (Akenhead J); and
(b) it is misconceived as a matter of fact because all of the witnesses were
clear and consistent that Agilisys was in a position to meet the Milestone
the Council’s reticence related to the potential ability of AD to meet certain
requirements it wished to see achieved in later drops.
[559]       Earlier in his submissions at 3.4.5 Mr Cormack submitted:
Thirdly, when read as a whole, the Subcontract required Relief Notices to be issued
and determined in the currency of the Subcontract prior to completion. This,
accordingly, requires the Court to engage in an essentially prospective analysis of
what was delaying the projects at any given time. In general terms, the Court will not
approach issues of delay from an entirely retrospective perspective. That this is the
correct approach is clear from the following features of the Subcontract: (a) Clause 31
is intimately linked to the form of liquidated damages in the Subcontract called
Delay Payments and specified in Section C of Schedule Part 7.1 of the Subcontract;
(b) such Delay Payments were immediately payable once a Key Milestone had not
been achieved: Para. 1.1 of Section C of Schedule Part 7.1, and other than in very
specific circumstances on the EI project were non-refundable: Para. 1.5 of Section C of
Schedule Part 7.1. Accordingly, there is no post-completion mechanism for a review
of the parties’ respective entitlements; and (c) the notice and response provisions in
Clause 31 required notice to be given and a response to be made ‘as soon as reasonably
practicable’ (Clause 31.2 and 31.4 of Subcontract), and much of the language is
inherently prospective in nature (‘the Milestone Date shall be postponed’ and ‘shall have
no liability to pay any Delay Payments’). In practical terms, this means that: ‘In the
context of this contractual based approach to extension, one cannot therefore do a purely
retrospective exercise. What one can not do is to identify the last of a number of events which
delayed completion and then say it was that last event at the end which caused the overall
delay to the Works. One needs to consider what critically delayed the Works as they went
along’: cf. Walter Lilly & Co Limited v. Mackay & Anor [2012] EWHC 1773 at para. 365
(Akenhead J).
[560]       CGI’s response to the objection was this:
It should also be noted that this question is directed at milestone MS-OSC1.1. As set
out at §346 below, there was only one contractual milestone for EI Go-Live, with the
Page 161 ⇓
161
subsequent division into ‘drops’ having no contractual effect. Agilisys’ contentions in
relation to the relevance of subsequent drops in their Note of Objections are therefore
misconceived. An objection is also taken by Agilisys on the basis that there are no
averments to support these passages of CGI’s witness statements. This objection is
ill-founded. The relevant averments can be found at Articles 33-37 of
Condescendence in Action CA54/17.
[561]       As I have said, I regard the division of the milestone into drops as having no
contractual effect. Nevertheless, the objection by Agilisys appears to be well founded for
this reason: no relief is sought relating to the deferral of drop 1, but rather to the prior
deferrals. Mr Sandison’s approach requires retrospectively judging the Relief Notices and I
agree with Mr Cormack that this is not the correct approach. I am persuaded that his
construction of the Subcontract is correct that Relief Notices are to be determined in the
currency of the Subcontract and that this requires a prospective approach to be taken.
[562]       Before turning to look at the various responsibilities incumbent on each party
regarding RN012, it is perhaps helpful in relation to RN012 to understand the nature of the
AD platform and this is conveniently summarised by Ms Conaghan in her witness statement
between paragraphs 8.4 and 8.6 (to which I earlier referred in passing):
“8.4 The AD platform provides local authorities with an integrated single
platform to manage digital contact with its customers. Citizens can access the
platform by web, tablet and smartphone, and officers (e.g. in the contact
centre and face to face) can use the same platform through the AD Contact
module (enabling contact centre staff to undertake the same transactions on
behalf of the customer.)
8.5 AD provides certain services fully integrated ‘out of the box’ as built in
modules (such as Council Tax transactions, Benefits Account, Housing
Account) with others added through single sign-on (“SSO”) integration with
3rd party portals (such as Benefits applications, Registrars bookings, bulky
waste booking). Integration with 3rd party portals to provide SSO (to enable
one customer account to be authenticated and used across all the online
portals and the platform to manage the customers session information across
all the online portals which would remove the necessity for customers to
relog in to each online portal as they moved around the CEC website). For
this to work, the 3rd parties were required to undertake development to
integrate with the AD federated identify services so they could receive the
Page 162 ⇓
162
customer log on information from AD and match it to their own
authentication userid.
8.6 Integration between AD and line of business systems ‘out of the box’ is
implemented via “connectors” or APIs provided by the line of business
systems. The connectors/APIs allow data collected by residents online to be
automatically entered into the line of business systems without the need to
re-key data or contact a council officer to do it for them.t. [sic] All other
solutions within the AD platform are integrated through SSO to online
portals as described above.”
In addition to the responsibilities incumbent upon Agilisys referred to in its previous
submissions in relation to RN002, CGI referred to a number of responsibilities of Agilisys
which specifically related to the EI project.
(1) Client Management and Governance (Project level).
(2) Programme management of the implementation programme and generation of all
programme control documents.
(3) Solutions architect to deliver the overall architecture of the solution and the
delivery roadmap, working with the CGI technical architect across the infrastructure
architecture.
(4) Take the lead in the digital architecture design with the Authority enterprise
architect and design authority, supported by the CGI technical architect.
(5) Replace the online transactions with the AD product with modules as required to
deliver the functionality within the OBS.
(6) Provide transaction data extract to be shown within the Oracle RightNow CRM
system.
[…]
(8) Implement the Agilisys Engage product.
(9) Integrate the AD product with the Scottish government authentication service.
(10) Business Analysis resource to work with the Authority team to define the scope
of the transactions and align to the AD modules.
(11) Technical resource to undertake the electronic forms development to provide
functionality not within the AD product.
[…]
Page 163 ⇓
163
(13) Project management of the transactional solutions to the timelines outlined with
the OBS, generating associated project control documents.
I accept these are incumbent on Agilisys.
[563]       Agilisys referred to the same responsibilities as in respect to RN002. Once more CGI
emphasised the general leadership and management role of Agilisys. I have already set out
my views regarding the nature and scope of this obligation. On a proper construction I
believe that the principal obligation incumbent on Agilisys was this: “The Supplier
responsible for defining third party requirements including APIs…” On a sound
construction of the foregoing it was incumbent upon Agilisys to identify what interfaces
were required and the extent of that duty is as I have earlier defined it. I do not believe any
of the individual responsibilities relied on by CGI expands on the scope of that duty
incumbent upon Agilisys. This I consider fits in with the obligation on the solutions
architect to deliver the overall architecture.” (emphasis added). Thus as part of the overall
architecture of the solution there is a necessity for APIs and Agilisys have to inform CGI
what APIs are required. However in terms of OBS 21 “the technical architecture” is a matter
for CGI “working in partnership with Agilisys” and this fits in with the evidence of
Ms Conaghan that Agilisys were responsible for “logical design”, ie what is necessary by
way of APIs but CGI was responsible for “physical design” although she accepted that in
order for logical design to work certain features of the infrastructure will require to be
specified; thus Agilisys to work “in partnership” with CGI relative to technical architecture.
Thus I believe Agilisys’ responsibilities were restricted as set out by Ms Conaghan.
[564]       The next question was whether CGI breached its obligations by failing to provide
access to fully functional EI environments? CGI’s response to this question was this: “The
Page 164 ⇓
164
required environments were provided by 11th August. Accordingly CGI failed to provide
the required environments by the contractually required dates” (para 355).
[565]       The next question: was CGI’s breach caused or affected by a breach by Agilisys of its
obligations? The same argument was advanced here by CGI as in relation to the ERP
environments, namely delay being caused by failure by Agilisys to provide an adequate
specification. The first piece of evidence relied on by CGI in respect to this was a short
passage in the evidence of Kevin Bryce at 4.1 of his witness statement. First this witness had
little to do with the EI project as he said in paragraph 4.1 of his witness statement. The
evidence to which I was directed was entirely general in nature. It did not identify what
particular requirements were not specified and what effect if any this had in respect of CGI’s
ability to fulfil its obligations. Reliance is placed on a passage in Richard Salmond’s
evidence. I have already commented on this passage of evidence and for the reasons I have
given I attach no weight in this evidence.
[566]       There is next reference to an email of Steve O’Sullivan of CGI where he asks for a
“consolidated document for EI environment Agilisys requirements?” Agilisys’ response is
set out in full at paragraph 360 of CGI’s submissions. Agilisys’ position in short was “all
necessary documents” were already available and that the “single required artefact”
requested by CGI was the responsibility of Kenny McAllister of CGI. It is submitted by CGI
that this approach is a breach by Agilisys of its obligations. I do not agree, Agilisys explain
that all necessary documents were available. It is saying it provided the information
necessary and it was dependent on Mr McAllister’s knowledge of “how these things were
set up”. This seems an entirely reasonable approach given Mr McAllister’s job and the
responsibilities of CGI regarding “technical architecture”; “procurement of required 3rd
party APIs” and “installation of 3rd party portals and APIs”. Mr McAllister in his evidence
Page 165 ⇓
165
said he produced a draft solutions document, although it had gaps (see: para 361, CGI
submissions). However, there is no specification as to when this was done; what the gaps
were and what effect, if any, it had on CGI in carrying out its obligations. I again observe
that in its submissions CGI refer to a single email in relation to EI environments issued at the
time of the implementation of the project regarding the issue of specification and it again
relates to production of a single document. Had this been an issue of significance I would
have expected the contemporaneous documentation to be littered with requests for
requirements.
[567]       For the foregoing reasons I am not persuaded that any failure to provide a
specification of requirements in relation to this aspect of the project caused or affected CGI’s
compliance with its obligations. Nor do I think any failure in general management
responsibilities on the part of Agilisys caused or affected CGI’s compliance with its
obligations.
[568]       The final question in relation to these Relief Notices was this: did any breaches cause
or were reasonably likely to cause Agilisys to fail to achieve the Milestone MS OSC1.1 in the
Implementation Plan by the effective date? Milestone MS-OSC1.1 was the Go-Live
Milestone for the EI project which according to Schedule Part 6.1 was 1 April 2016. CGI
submitted that the operative cause of the failure to reach Milestone MS-OSC1.1 was the
unsuitability of the AD product provided by Agilisys. On the basis that I am wrong in
relation to the issue of the objection to evidence regarding the suitability of the AD product I
will now turn to this issue.
[569]       Mr Cormack argued:
8.10. In any event, the four strategic issues identified with the EI Project at the
time did not represent a breach by Agilisys of any of its obligations and
did not have any causative effect on delay to the project.
Page 166 ⇓
166
8.10.1. The evidence led by CGI on this issue fell woefully short of establishing any
breach of responsibilities or contractual obligations on the part of Agilisys.
8.10.2. The common thread was that all of the issues were areas where CEC had
decided it wanted to make changes to the scope of the Project and were then
unhappy that these scope changes could not be instantly produced. It is
submitted that Dr. Hunt's analysis of these issues was entirely correct in her
Supplementary Report and in the Joint Statement of Experts. Mr. Coyne's
criticisms were impressionistic, reliant on statements of opinion by other
people, and were not backed up by any actual technical analysis of the AD
product, or the issues, or proper evidential foundation.
8.10.3. The four strategic issues, and Agilisys position in relation to them, are
summarised, as follows.
8.10.4. The first issue is Business Authentication. This arose from a perverse stance
taken by CEC, coupled with the lack of any commercial support by CGI
when that stance was taken in late 2016.
8.10.5. As appears to be common ground, Agilisys' obligations were specified in
the column headed Agilisys in OBS 21. The specific obligation on Agilisys in
relation to authentication was to 'integrate the AD product with the Scottish
government authentication service'. This related to the actual functional
requirement from CEC in OBS 21 that specified 'Customers are identified and
authenticated by use of the Scottish Government Citizen Account Services and
Council's identity & access management platform.' This is entirely consistent
with the Implementation Plan that mentions in Milestone MS-OSC1.1
'myGovScot integration'. It is, also, entirely consistent with the Project
Overview, and Supplier Responsibilities & Deliverables that only mention
integration with the Scottish government authentication service.
Accordingly, Agilisys was contractually obliged to provide integration with
the Scottish Government's myGovScot account for customers. The Court
will search in vain for a requirement that a separate and bespoke solution
for business authentication required to be provided. It, also, did not feature
anywhere in, and is inconsistent with, the Conceptual Architecture for EI
that was signed-off by CGI in February 2016: see JB 651, particularly at
p. 9979 9980, and JB 357 for sign-off by CGI
8.10.6. The stance taken by CEC in late 2016 came to be a classic Catch 22 from
Agilisys point of view: (a) CEC remained insistent on the contractual
obligation that authentication be done through MyGovScot, as opposed to
any other method; but (b) at the same time, MyGovScot did not have an
established system for business authentication and CEC were unwilling to
contemplate that authentication of businesses by an individual
representative through that method for unspecified 'data protection' issues
(despite having been willing to contemplate that at an earlier stage). It is
submitted that Agilisys were entirely correct to identify that what CEC were
requesting was the development of an entirely separate and bespoke method
Page 167 ⇓
167
of business authentication that featured nowhere in OBS 21 or the Roles &
Responsibilities in Schedule Part 4.1: see JB 1. Agilisys endeavoured to assist
CEC in finding a solution. But they simply were not supported by CGI
leveraging its commercial powers with CEC to make it clear this amounted
to a change to the project scope.
8.10.7. The second issue is Gandlake. This, again, arose from changing requirements
from CEC. As Mr. Fulda himself accepted in evidence, the issue in relation to
Gandlake was that it contained functions that OBS 21 had not identified. This
is recorded in the Gandlake Summary issued by Agilisys in June 2016: see
JB 2220, and in particular p. 22792. As was said then: 'During the Enterprise
Integration project implementation, it has since been identified that the actual
functionality provided by the Gandlake product delivers much more functionality
than that outlined in OBS 13... Gandlake in fact delivers additional functionality to
citizens and... to a large number of staff. As a consequence of this mismatch, the
Council raised a Change Request (CR 154 please see Appendix A) for Agilisys to
assess and define the requirements for staff users that had been missed." Agilisys
very clearly identified in that report that the Agilisys Digital product did not
provide some of this additional functionality and, as such, that a technical
workaround would have to be found. Again, it is extremely difficult to
understand the basis on which this later became founded on as a breach of
Agilisys' obligations. Agilisys never undertook to provide a system with
functionality that was nowhere mentioned in OBS 13 or 21. Again, it is
submitted that Dr. Hunt's analysis of the issues here was entirely correct.
8.10.8. The third issue was CRM Replacement. This reflects, in effect, an attempt by
CEC to accelerate and impose upon Agilisys an obligation it had never
undertaken. This is, perhaps, the most striking of the four issues. It is plain
beyond peradventure that OBS 21 does not require replacement of CRM. The
term 'CRM' is defined in OBS 21 under reference to the existing CRM system,
Oracle RightNow. Agilisys' obligations in OBS 21 in relation to CRM were as
follows: 'Provide transaction data extracts to be imported into the Oracle
RightNow CRM system.' This corresponded to CGI's obligation to 'Provide
Right Now expertise to undertake changes delivered as a catalogue item' and
'Import customer transaction data extracted from AD into the Oracle RightNow
CRM'
8.10.9. The Supplier Solution in OBS 21 itself stated: 'The Supplier shall deliver existing
agent-facing capabilities through the Authority's current CRM platform at
Operational Service Commencement Date, and shall integrate the customer history
record from the Agilisys Digital platform with the current CRM. The Supplier shall
ensure that the agent-facing functionality of the CRM provides visibility of
transactions that have been completed through the online services.' It went on to
state that 'The Supplier shall deliver, within six months from Operational Service
Commencement Date, a roadmap and business case detailing its plans to replace or
maintain the current CRMylatform: The same references are clearly made in
Schedule Part 4.1 of the Subcontract.
Page 168 ⇓
168
8.10.10. CGI's own witnesses, Richard Salmond and Tom Fulda, both confirmed that
it was never understood to be within the scope of OBS 21 to replace CRM.
Mr. Coyne is really the only witness who seriously suggests another
construction, and he has no proper basis for doing so. It is submitted that he
materially damaged his credibility and reliability as an independent expert
witness by becoming an advocate for CGI on this point.
8.10.11. The fourth, and final, issue was the issue of an agent authenticating on behalf
of a citizen. This, in effect, arose from two conflicting requirements. Agilisys
was obliged to use MyGovScot to authenticate customers. But that method of
authentication meant that if an agent wanted to do an online transaction on
behalf of the customer over the phone then it would potentially need to go
through the authentication system. This solution was unacceptable to CEC
despite being the inexorable logic of the system they had contracted for. It did
not affect Drop 1. And by the time of rescission of the Subcontract, Agilisys
had a fully worked out solution to the issue.
8.10.12. A number of other issues were canvassed but it is submitted that they can be
quickly dealt with in submissions:
8.10.13. Parking: the issue was the refusal of the 3rd party supplier NSL Limited to
work with Agilisys and CGI in providing an SSO to the Xerox Easipermits
portal. As was explained by Claire Conaghan, NSL Limited had been
overlooked in the managed contracts being taken over by CGI: para. 12.10.4
of her WS. As all of the Agilisys witnesses explained, there was an
Easipermits online portal provided by Xerox capable of integrating with AD
over an SSO. The difficulty was that NSL refused to admit it existed or would
be suitable for months. If the responsibility for this lay anywhere, it plainly
fell under CGI's remit of procurement and installation of 3rd party portals.
They, ultimately, took the commercial risk that it might be difficult to procure
such portals as transpired to be the case in relation to Parking.
8.10.14. Housing: the issue with Housing was not canvassed with any of the Agilisys
witnesses and was not spoken to by any of CGI's factual witnesses. As
Dr. Hunt observes in the Joint Statement at Para. 72, no detailed complaint at
a technical or project level is actually made nor is any actual period of delay
identified. It is submitted that this issue is utterly irrelevant.
[570]       In reply CGI submitted:
364. The Go-Live Milestone was split into a number of ‘drops’, to take place
between November 2016 and March 2017 [para 346 above]. Completion of the
EI project to achieve the Milestone accordingly required the achievement of
each drop. None of the drops was achieved.
365. David Buchanan’s evidence concerning the position in relation to the Go-Live
Milestone was that CEC were ‘confident’ in relation to the status of the
activities required for Drop 1, but that there was concern about ‘the wider
Page 169 ⇓
169
strategic challenges which we [sic] relevant to the future drops.’ It is submitted
that this analysis ignores the close relationship between all of the drops. It was
necessary for CEC to be confident not only that Drop 1 could be achieved, but
also that the remaining drops were viable. Tom Fulda’s evidence in relation to
CEC’s position was that,
Once they had decided to Go Live with Phase 1 they would have burned their
bridges. It was deferred to allow Agilisys to satisfy CEC that the further business
requirements to be met by subsequent phases could be met.
366. The Court is invited to accept Mr Fulda’s analysis, which is supported by the
evidence of CEC’s attitude to Drop 1 in November 2016. On 16 November 2016,
Colin Ronald of CEC emailed Greg White and Cliff Graham of CGI to advice
that CEC,
were not prepared to accept part of a solution that ties the Council into a technical
platform that is unable to deliver what was stated as part of the contract. The dates
below do not align with the Council supporting such a decision prior to the go/no
go decision date for Drop 1.
367. Following this message, the decision was taken to defer Drop 1. Andrew
Couch’s evidence was that the decision was a decision made at a ‘senior’ level
by representatives of CEC, CGI and Agilisys. The reason expressed for this in
the email sent by Cliff Graham to CGI and CEC advising of the decision was
that,
we have collectively agreed to defer delivery of Drop 1 of the El project (previously
scheduled for the 24th November 2016) until we have completed activity on
establishing a full and detailed project delivery plan for all subsequent phases of the
project.
368. Mr Graham’s evidence was that this decision took place against a background
where,
Agilisys were becoming increasingly less confident of Phases 2 to 5 and their
timescales for delivery; and also because we were apparently some way off an
agreement on the four critical issues, which were Gandlake replacement, CRM, AD
contact and business authentication.
369. It is submitted that it was the unsuitability of the AD product which was the
operative cause of both the deferral of Drop 1 and also of the ultimate failure to
achieve all drops (and thus the Go-Live Milestone). This was manifested in the
‘critical issues’ identified by Mr Graham, in addition to further difficulties
experienced in relation to the fitness for purpose of AD.
370. Agilisys were aware that CEC had chosen the AD product because of the
opportunity which it offered to extend the services which they provided ‘in a
cost-effective way’. David Buchanan’s evidence was that Agilisys Digital was
Page 170 ⇓
170
‘absolutely a flagship product. A whole business strategy had been predicated
on its success’. Similarly, CGI were told by Agilisys that ‘AD could do a lot of
the work more cheaply and more flexibly.’ AD, however, had not previously
been used in a situation where Agilisys was acting in a prime-subcontractor
relationship.
371. The performance of AD was contrary to CGI and CEC’s expectation that it
would be a Commercial Off the Shelf (COTS) or ‘plug and play’ solution. David
Buchanan’s evidence was that on ‘[a]lmost every occasion there are going to be
levels of bespoking or customisation’ with AD. As Claire Conaghan
acknowledged in her evidence, a COTS solution was not possible:
AD could never be that because of the number of back-office systems that would
need to be integrated to enable transactions to go from the customer self-service
through to the back-office systems.
372. Furthermore, the amount of ‘bespoking’ required for the AD product varied
between clients. It was conceded by Ms Conaghan that, although the work had
not been completed, ‘significant development’ had been estimated by Agilisys
to deal with the scale of the exercise required for CEC, given the ‘wide range of
services much more than other customers’ which CEC required. Mr Buchanan
accepted that it was a ‘fair reflection’ that AD had never been delivered in a
project as complex as that required by CEC.
373. Kenny McAllister’s evidence was that:
I was of the view that AD had been sold as a complete solution, but the off-the-shelf,
self-service part of AD moved a lot, so it was not a complete solution. Not
everything was within the AD platform. Agilisys were having to use a number of
third parties in order to deliver the proposed solution.
374. This problem identified by Mr McAllister third-party integration for AD
caused particular difficulties.
375. As a result, the AD product required more interfaces in practice than had been
envisaged at the time of entering into the Subcontract. Anna Vale’s evidence
was that:
We had understood pre-contract that the EI product had pre-existing interfaces,
and you’ll see in the letter of 29 September that we do go on to describe the fact that
a lot of Agilisys appeared to change their policy and presented then to us that they
were going to move forward with their product in a specific way, which meant that
we would need greater interfaces than we had originally envisaged in the contract.
376. The unsuitability of the AD product was evident in the specific areas identified
below. In respect of each these areas, an issue arose as to the ability of the AD
product to meet the requirements of OBS21. In respect of each of these issues, it
is submitted that the AD product was unable to deliver the contractually-
Page 171 ⇓
171
required functionality. In any event, however and irrespective of any possible
modifications which might ultimately have been made to the AD Product no
solution to any of the issues had been found by the time at which Agilisys
chose to leave the project in March 2017. On any view, therefore, it is submitted
that the failures of AD in this regard were the operative cause of delay to the
Go-Live milestone.
Gandlake
377. Gandlake was the function which allowed certain records (relating to revenues
and benefits) to be viewed. The requirements for Gandlake replacement were
set out in OBS 13 and OBS 21. OBS 13 provided:
The Supplier [Agilisys] will initially will provide all the requirements listed in this
paragraph 6 utilising the current version of Citizen Access Portal (Gandlake), non-
version-specific, which will be based within the current Supplier’s cloud
infrastructure, retaining or enhancing all current functionality utilised by the
Authority at the Service Cut Over Date.
After undertaking transformation, the Supplier will then provide all of the
requirements listed in this paragraph 6 utilising Agilisys Digital, non-version-
specific, which will be based on cloud infrastructure retaining or enhancing all
current functionality utilised by the Authority at the Service Cut Over Date.
378. In respect of two of the requirements of OBS 13 (OBS13.1: ‘View their Council
Tax and Non-Domestic Rates and Recovery Notices’; and OBS 13.5:
‘Automatically receive files transferred from the Income and Benefits
Management System, where they can be viewed by citizens’), OBS13 provided
as an Agilisys requirement,
Through the Enterprise Integration (OBS 21) programme, replace the Gandlake
Account view with the AD product with modules to deliver the functionality.
379. OBS 21.1 provided that Agilisys was to
Provide customer access across MyAccount Revenues and Benefits replacing
Gandlake for customers.
380. Despite these requirements expressly based upon the functionality of the
legacy system, rather than the functionality that AD might be able to offer
Claire Conaghan acknowledged that, following conclusion of the Subcontract
and the specifications in OBS21,
we then reviewed the functionality of Gandlake, how it was used, what was
available for both customers and staff that wasn’t outlined in the OBS
requirements, there were big gaps from what we could do and what the Gandlake
product did.
Page 172 ⇓
172
381. David Buchanan’s evidence was to the same effect. It is submitted this position
advanced on behalf of Agilisys fails to engage with the terms of OBS21. OBS21
did not limit the functions which AD was to perform: it simply stated that
Gandlake was to be replaced by AD. That this involved functionality which
had not been anticipated by Agilisys was not relevant to the scope of the
contractual obligation. Moreover, Agilisys, as leaders of the EI project, ought to
have engaged with CEC in order to manage the change from Gandlake to AD,
and to identify any difficulties in relation to functionality. Agilisys failed in
respect of both of these obligations. Dr Hunt acknowledged that no agreed
solution had been found at the time Agilisys left the project.
382. The Notifiable Default issued on 13 December set out that Agilisys had ‘no
solution’ to the requirement of OBS 21.1 noted at para 379 above. It is
submitted that this analysis was correct. No solution had been identified which
would have given CEC sufficient confidence to proceed with Drop 1.
Business authentication
383. OBS21 required use of AD for transactions both by individual citizens and also
by businesses. The identifies [sic] of these users had to be authenticated.
OBS21.1 required that Agilisys ‘Integrate the AD product with the Scottish
Government authentication service.’
384. The AD product was not designed to separate citizens and businesses. Previous
local authority customers of Agilisys had not had such a requirement. CEC’s
concerns in this regard were raised in an email from Paul Neary to Claire
Conaghan on 27 September 2016. Claire Conaghan acknowledge that she could
‘see where [CEC] were coming from’. Solutions were discussed (including the
use of the existing authentication method for individual citizens for
businesses), but no agreement was reached with CEC as to the way forward.
David Buchanan’s evidence was that ‘a fundamental design change of how the
[AD] platform is built’ would have been required in order to meet CEC’s
requirements.
385. Business authentication was raised in the Notifiable Default issued by CGI on
13 December 2016. This stated that Agilisys had ‘failed to provide a viable
solution’ in respect of business authentication, and noted that,
Agilisys' proposed solution of using the existing "citizen" authentication to meet
this requirement is unacceptable and has been rejected by the Council. Citizen
authentication cannot be used to fulfil this requirement due to the associated
information governance issues and as an 'expert' in local government
transformation, that has implemented the AD product numerous times before,
Agilisys should have known this.
Page 173 ⇓
173
386. No solution to business authentication had been identified by the time Agilisys
left the project. The Court is accordingly invited to accept Mr Coyne’s view that
in respect of business authentication, ‘CGI’s position of default appears to be
correct’.
Customer Relationship Management (CRM)
387. CRM was the,
customer relationship management system. If a citizen were to phone CEC, a call
handler would complete the task over the phone and report the outcome on the
CRM. If the same transaction were to be done by the citizen online, then AD would
need to be able to record the dialogue onto the CRM.
388. OBS 21.7 required the provision of ‘customer facing and agent facing portals’.
In respect of these two different contact routes, OBS21.7 required,
Development of the CRM functionality for agents and other council staff should
match that available to citizens within the customer portal or as web transaction.
389. OBS 21.7 required that Agilisys was to ‘develop new online and agent-facing
transactions in the Agilisys Digital solution to meet the specific requirements of
this OBS’. This included the requirement, noted above, for matching
functionality for customer and agent transactions. OBS21.7 further provided
that the Agilisys ‘solution shall enable the staff and customer use the same web
forms.’
390. Paul Neary’s evidence was that,
The Agilisys design documents do not make clear how it would handle the multi-
channel requirements of being able to raise an incident as a “citizen” via the web
and as a “contact centre agent” on behalf of a citizen calling into the contact centre
via telephone.
391. Similarly, Tom Fulda’s evidence was that,
CEC did not consider Agilisys Digital (“AD”) … to be capable of meeting all the
CRM-type obligations specified in OBS21.7. Sch 4.1 states a ‘roadmap and
business case detailing its plans to replace or maintain the current CRM platform’
would be produced. The absence of this roadmap/business case became a strategic
issue.
392. This issue was to be addressed by the ‘AD Contact’ solution which was part of
Agilisys Digital, and was to be delivered as part of Drop 2 (as for Drop 1 the
legacy system, Oracle CRM, was to be retained). The AD Contact solution,
however, was unable to deliver the functionality required by OBS21.7. The
Notifiable Default of 13 December 2016 set out,
Page 174 ⇓
174
On Tuesday 20 September 2016 Agilisys provided a demonstration of the
functionality of their "AD Contact" solution. This highlighted that the solution
would only provide functionality for agents to act as a simple proxy for citizen on-
line transactions, and only when the citizen has an online account. There is no
facility to process transactions in this way for citizens who do not have, or do not
wish an online account, and contravenes the DBS requirement that the solution
should provide the same experience irrespective the [sic] channel used for the
interaction.
393. A ‘CRM Options Document’ was produced in February 2017, but no solution
had been agreed at the time that Agilisys left the project. Accordingly, it is
submitted that Agilisys failed to meet the requirements of OBS21.7.
Phone-only contact
394. Phone-only contact also related to the requirements of OBS 21.7. OBS 21.7
required Agilisys to ‘develop new online and agent-facing transactions in the
Agilisys Digital solution’. Accordingly the OBS required that AD could be used
for a transaction that was ‘online’ (that is, where the customer initiated the
contact directly online), or ‘agent-facing’ (that is, where the customer
telephoned a CEC agent, with the agent then carrying out the transaction rather
than the customer doing so directly).
395. This requirement raised similar issues to those found in relation to CRM. The
Notifiable Default of 13 December stated,
Due to the design of the Agilisys proposed "AD Contact" solution; any on-line
transactions which are required to be performed by a contact centre agent on behalf
of a citizen, require the citizen to have previously been registered with mygov.scot.
This is an issue for citizens contacting the Council by phone, as highlighted by the
Council at the demonstration of the AD Contact solution by Agilisys on Tuesday
20th September 2016.
396. The changes required to AD to enable this requirement to be met were still
‘being worked on’ at the time of termination.
Parking solution
397. OBS 21.25 provided,
The Supplier [Agilisys] shall provide the necessary systems integration to internal
and external line of business systems to support the programme described in Annex
C of the Authority’s OBS.
Page 175 ⇓
175
398. These requirements (which were also set out in OBS 21.25) included the
following for ‘Parking’:
Provide the identified Parking transactions using the EI Platform.
The Supplier intends to use an SSO to the SIDEM Portal to fulfil this requirement.
399. Claire Conaghan’s evidence was that ‘We knew the portal was available as our
AD team had been talking to another of our clients implementing the portal
there’. In cross examination, however, she acknowledged that the project with
Agilisys’ other client (the London Borough of Hammersmith and Fulham) had
not reached the stage where a successful integration had been achieved.
400. A Notifiable Default, identifying the inability of AD to interface with the SiDem
portal was issued on 18 February 2016. The issue had not been solved by the
time of the Notifiable Default issued on 13 December, which noted that the
proposals made by Agilisys to resolve the issue in the interim were
‘commercially complex and unacceptable’. Dr Hunt’s evidence in relation to
whether AD was able to interface with SiDem at the date Agilisys left the
project was that ‘[i]t couldn’t’.
Housing
401. The requirements for housing were, as for parking, found in OBS21.25. This
provided that Agilisys was to,
Provide the identified Housing transactions using the EI Platform.
The Supplier intends to deliver these capabilities using an SSO to the Northgate
Housing Portal, combined with the forms capabilities of Agilisys Digital.
402. The Notifiable Default of 13 December 2016 stated that there was ‘no
confidence that Agilisys will deliver the integration at the date identified in the
roadmap.’ The Court is invited to accept Mr Coyne’s analysis that,
Given the various re-planning activities that occurred in respect of EI, and the
failure to meet previous milestones, it is understandable that confidence was low.
Suitability: conclusion
403. It is submitted that the issues identified in the Notifiable Default of
13 December 2016 individually and collectively represented the major
obstacle to progress to Go-Live in respect of the AD Product. Irrespective of
whether a solution to these issues would ever technically have been possible,
no solution had been identified by the Drop 1 date (and, indeed, no solution
had been identified by the time at which Agilisys left the project). The failure to
Page 176 ⇓
176
resolve these issues, together with Agilisys’ earlier failures of management,
caused CEC to lose confidence in the AD product. As a result, CEC refused to
proceed to Drop 1, with the situation not remedied prior to Agilisys leaving the
project. This loss of confidence, caused by the failings of the AD product, was
the cause of the failure to meet the Go-Live Milestone.”
[571]       As regards the first issue, namely: business authentication, I am persuaded by the
analysis of Agilisys contained at paragraph 8.10.5 and 8.10.6 of the submissions. The
submissions advanced by CGI in relation to business authentication do not answer the
points made by Mr Cormack. Agilisys could integrate its AD product with the Scottish
Government authentication service, which was the specific obligation incumbent upon it.
The difficulty was CEC’s insistence on the use of the authentication service which did not
have an established procedure for business clients. That is not a failure in the AD product,
it is CEC wanting something more than integration with the authentication service.
[572]       I clearly prefer Agilisys’ analysis in relation to CRM. On a sound construction of
OBS 21 there was no obligation on Agilisys to replace the current CRM platform. Its
obligation regarding replacement extended only to the preparation of a roadmap “defining
its plans to replace or maintain the current CRM platform” within six months of the
operational commencement date.
[573]       In respect to agent authentication again I prefer the analysis of Mr Cormack.
[574]       As regards Gandlake, I agree with the submissions of Mr Cormack.
[575]       The next issue is parking. As argued by Mr Cormack this issue appeared to relate to
a 3rd party portal. An explicit obligation imposed on CGI was the “Procurement of required
3rd party APIs and supplier portals”. Accordingly the difficulty arose here from a failure by
CGI to procure the necessary portal.
[576]       As regards housing, I agree with the position taken by Mr Cormack.
Page 177 ⇓
177
[577]       Dr Hunt in her supplementary report carefully analyses each of these areas and gives
detailed reasons why she believes that no criticism can properly be made of Agilisys in
respect to these matters. I find her evidence in respect to this part of the case convincing.
Overall I prefer the position of Agilisys in relation to these issues.
[578]       A final point made by Mr Sandison was this: drop 1 was postponed as a matter of
agreement between Agilisys, CGI and CEC. He developed this argument as follows: all that
we need to establish on our analysis is that it had not been made to work at the point that
the milestone that is complained of having been missed was due to go ahead, and that lack
of demonstration that it could be made to work in the various regards led to a decision
among all three of the interested parties that the events that would constitute the occurrence
of the milestone would not in fact occur. Thus the EI milestone was not missed because of a
CGI failure but because of a collective decision not to proceed to the milestone. It does
appear that there was a joint decision. However, on a proper analysis what underlay and
necessitated this decision was CGI’s failures as earlier identified. These were the effective
underlying causes of the decision.
[579]       In the end I accepted the following submissions made by Agilisys:
8.8.1. The Court is invited to conclude that the two dominant causes of Agilisys’
inability to meet the originally planned Milestone of MS-OSC 1.1 in the
Subcontract were both the Northgate API issue and the Environments
(particularly in this context the Biztalk Environments). These were, it is
submitted, the dominant causes operating concurrently in the period
September 2015 January 2016.
8.8.2.
It is, further, submitted that (whilst the Northgate API issue was resolved by
late January 2016) the dominant causes of delay in the period after January
2016 continued to be the unavailability of the Biztalk, Development, and Test
Environments (also joined by the Production and Pre-Production
Environments as outlined in RN010). As RN010 clearly stated: “Agilisys has
made it clear that it is in a position to deliver all required functionality for milestone
MS-OSC1.1 with the exception of the Biztalk integration which is dependent on
CGI delivering on its Biztalk responsibilities.” As will be recalled, CGI accepted
RN010 and granted relief on 20 May 2016.
Page 178 ⇓
178
RN005 & RN011 Payroll
[580]       Agilisys did not make detailed submissions in respect to these matters on the basis
that on its understanding of the way that the case had developed there was no apparent
dispute regarding payroll and therefore it was not clear to Agilisys that there was any
relevant dispute between the parties. CGI had accepted RN011 in its response to that Relief
Notice. This was done on the basis that CEC had agreed that its inability to provide a
payroll resource to enable configuration and build work of the BW solution amounted to
Authority cause under the Prime Contract. Mr Cormack noted that Mr Coyne had revisited
these Relief Notices in his report. But the purpose of that does not seem to be clear and
seems to take no account of the fact that CEC accepted it was in breach of its responsibilities.
The above understanding that there was no dispute, however, proved to be wrong and short
written submissions were advanced by CGI.
[581]       Relief Notices 005 and 011 concerned the provision of information by CEC’s payroll
team to enable the completion of the build element of the ERP project. CGI in its response to
RN011 accept that CGI had the following responsibilities in terms of OBS 1:
“’Provide business information, knowledge and decisions to enable the solution to be
configured. The necessary information and support to enable Payroll configuration
to complete on the scheduled date on 31st March 2016, was not met.”
[582]       In addition, Agilisys had relied on the following responsibility incumbent on CGI in
RN011:
“CGI has a responsibility in the subcontract Schedule 3 to: ‘procure that the
Authority fulfils its roles and duties under this Agreement and as defined in the
Implementation Plans and in Schedule 2.1 (Services Description) and in Schedule 4.1
(Supplier Solution)’. The Authority has not fulfilled its role of providing business
information, knowledge and decisions to enable the completion of the build process
for the Payroll element of the ERP project.”
Page 179 ⇓
179
[583]       CGI relied on the following responsibilities of Agilisys’ “Project Roles” that Agilisys
was responsible for “Leading delivery of this project”, with CGI responsible for “support
with the Authority governance and resources.” Agilisys was further stated to be responsible
for, inter alia:
“1. Client management and Governance (Project level)
2. Programme management of the implementation programme and generation of
all programme control documents
[…]
8. Change management resource to lead the Authority through the change
programme.”
[584]       In addition it relied on
Change management was a responsibility of the Application Architect role which
should have been carried out by Claire Conaghan (discussed at §28 above). In
addition, leadership of the collaborative work with CEC was the responsibility of the
Transformation Director (Louise Ah-Wong), as discussed in relation to the roles and
responsibilities of parties under the Subcontract at §§16ff above.
I accept all of the above referred to obligations are incumbent on parties.
[585]       The critical responsibilities are those incumbent on CGI. It was incumbent on CGI in
respect of payroll to procure that CEC met its obligations, procure being defined as I have
earlier set out. This is consistent with CGI’s responsibility for support with the Authority
governance and resource. Agilisys was, for the reasons I have already set out, playing a
subordinate role in respect to CEC. In order to obtain business information, knowledge and
decisions from CEC, it is the power to procure which is critical. Where CEC is not providing
information requested and by the response to RN011 that was accepted I do not see how
some general management responsibility incumbent on Agilisys could in any way impact on
the failure by CEC to provide the necessary information. The obligation of Agilisys to
manage has to be seen within the context of CGI’s obligation to procure the business
Page 180 ⇓
180
information etc as earlier set out. It cannot simply avoid that obligation and empty it of
content by blaming a failure of management on the part of Agilisys which is both vague and
nebulous. Nor can I identify how the issue of change management impacts on this issue.
Finally the role of Transformation Director is again subordinate to the role of CGI procuring
that CEC fulfilled its duties.
[586]       It is clear, from the narrative set out in RN011 as follows:
“Identification of the data required from the Payroll team was initially requested by
Agilisys from the Council during the Design Workshops held in September and
October 2015 in the form of templates. A consolidated list of outstanding data
requirements and the dates by which this information was required was
incorporated in the Payroll Detailed Design Document that was approved by the
Council on 27th November 2011. Subsequent updates to this list of outstanding
actions and data requests has been provided on a monthly basis both in presentation
to the ERP Project Board on 15th December 2015, 19th January 2016 and 16th February
2016 and as a detailed analysis provided to the Payroll team following these Project
Board meetings. Agilisys received an email from the Council Payroll workstream
lead on 7th March 2016 advising ‘I am unable to support this work in the manner in
which Agilisys require.’ Subsequent discussion with the Council Payroll team
following this email clearly identified that there was insufficient resource within the
Council Payroll team to provide the necessary information and knowledge to enable
the configuration of the BW Payroll to continue. The Council did manage to provide
a resource with Payroll experience on 29th March.”
that Agilisys had done what it could in order to obtain the information. It is clear from that
narrative that it was thereafter for CGI to procure performance by CEC of its responsibilities.
[587]       The actual failures in management relied on by CGI are based on first an email of
Mr Sheriffs of 7 April 2016. He says, inter alia, “the escalation of the lack of payroll
information was handled in a very low key manner”. He then sets out a number of points.
These points all relate to the period after 7 March 2016 when CEC’s workstream leader
Grant Craig had accepted “unable to support this work in the manner in which Agilisys
required”. I cannot see how what he states amounts to a serious criticism in respect to
management. The next matter relied on is a response by Hugh Dunn to Ms Ah-Wong
complaining of a lack of availability of CEC resources and his comment “he has never been
Page 181 ⇓
181
made aware of the time critical resource requirement until it has caused a delay”. These
comments have to be seen within the context of the response to RN011 which accepted that:
“the necessary information and support to enable payroll configuration to complete on the
31st of March 2016 was not met.” This appears to me on a fair reading to be wholly
inconsistent with the position now being advanced by CGI. There is reference in the
response to Agilisys leading the project, however, it is clear that no particular management
failure is identified on the part of Agilisys. Those observations in addition have to be seen in
the context of the narrative to which I have above referred which sets out that this
information had been sought for a substantial period from CEC.
[588]       I am persuaded by the foregoing that: (1) there was a breach of its obligations by
CGI; (2) it was not caused or affected by any breach of Agilisys’ responsibilities. I am
further persuaded that in conjunction with environment difficulties it would have caused a
failure to achieve the relevant milestone.
Conclusion
[589]       For the foregoing reasons I would answer the questions as follows:
Relief Notice 002 3rd Party APIs and Supplier Portals: To answer question 2 yes,
question 3 noand question 4 yes.
Relief Notices 004 & 006 Biztalk: To answer question 6 yes, question 7 noand
question 8 yes.
Relief Notices 005 and 011 Payroll Information: To answer question 10 yes,
question 11 noand question 12 yes.
Relief Notices 007 and 016 - Data Migration: To answer question 14 yes, question
15 noand question 16 yes.
Relief Notice 008 Interface Specifications: To answer question 18 yes, question 19
noand question 20 yes.
Page 182 ⇓
182
Relief Notices 009, 014, and 015 ERP Environments: To answer question 22 yes,
question 23 noand question 24 yes.
Relief Notice 012 EI Environments: To answer question 26 yes, question 27 no
and question 28 yes.
[590]       I shall put the case out By Order to determine the orders necessary, and reserve all
issues of expenses meantime.
Page 183 ⇓
Date
October 2015
183
APPENDIX 1
DATA MIGRATION
Chronology
On 6 October 2015, and further in the course of October 2015, Agilisys holds a series meetings with CEC representatives
for the purposes of discussing, communicating, and understanding the data migration strategy in order to feed into the
Solution Design Documents: see JB 244, and reference in SDDs (for example at p. 5208 in JB 235), Ravinder Johal’s WS at
para. 4.2.1, and see reference to 5 further planned meetings in the Project Board slides in October 2015 at JB 72 73
(particularly at p. 2298).
November 2015
On 2 November 2015, Ravinder Johal and Peter Hall exchanged emails about the request for a PPSL database extract: JB
533. Ravinder records that she had agreed with PPSL leads they would need access to this data. She records her statement
that “we will need access to this data for migration so happy to sign any NDA as we will be doing the migration and to do so we will
have access to more council data than this including HR and payroll” On 18 November 2015, David Legge emailed Ravinder
Johal after having reviewed the SDD and asks various questions about extracting from PPSL in order to assist with
requests for data from BT [JB 1753]. At this point, it is clear that Agilisys has already requested access to the PPSL
database and it is acknowledge there are issues being raised by CEC, even prior to signing off the SDD, about access to
live data.
On 18 November 2015, Ravinder Johal and Charles Wyndham Jones exchange an email in which Agilisys clearly explain
Page 184 ⇓
184
that they do not intend to take the approach of producing a Data Migration Specification which details source data
location (database table and field), transformation rules and the new data location (database table and field) in the ERP
solution. Ravinder clearly explains the strategy of defining the data to be moved and then using the templates to move it:
JB 568 and Ravinder Johal’s WS at para. 4.2.6.
Between 24 November 2015 and 2 December 2015, the SDD for each workstream was signed off containing the data
migration strategy: see JB 234 241.
On 25 November 2015, Peter Hall emails a Data Migration Overview document to David Camilleri of CEC: JB 590. It is of
importance to note a number of aspects of this document that clearly set-out Agilisys’ strategy. These include: ‘Principle 2
Migrate early, Migrate often’ at p. 9115. The point being that if migrations are done early and often enough then it reduces
time pressure and allows it to be done in parallel with the build; and ‘Principle 4 Iterate Data Loads’ where it explains the
changing nature of the templates. It sets out in detail the various data to be migrated. This document is not discussed by
any of CGI’s witnesses, despite the fact that criticisms are repeatedly made of not having communicated the precise
strategy and approach to CEC.
On 27 November 2015, the Milestone Date for MS2.1 (Design Specification Approved) in the Implementation Plan is
reached (JB 1 @ page number 1119) and Milestone 2.1 is signed-off by CGI: JB 241. It is plain from looking at the PID, the
wording in the Subcontract, and subsequent references that everyone viewed these as containing the data migration
strategy discussed and agreed at the design meetings recorded in them.
Page 185 ⇓
185
December 2015
On 4 December 2015, Martin Spelman sends the first draft HRP templates to David Legge by email: JB 620 - 632. David
Legge repled to the email to say that he had ‘Just had a look through these – I don’t see anything untoward at this stage’: in JB
716. He, also, explained that he was not making as much headway with PPSL as he would like. He said that getting access
to HR/ Payroll and Finance is problematic due to personal data.
On 14 December 2015, Ravinder Johal and David Legge exchanged further emails. The particular topic was PPSL extracts.
The issue of personal data is raised again. Ravinder Johal asks ‘what do we need to do to get round this?: in JB 683. David
Legge replied to say that, if Ravinder could confirm that what he has suggested met her needs, he would raise a request
with BT. This is clear evidence that: (a) Ravinder was being helpful and co-operative; and (b) that David Legge was
indicating he was intending to take up the issue of live data.
January 2016
In January 2016, Agilisys and CGI representatives undertook lengthy steps to take matters forward, and to ensure that
David Legge fully understood the templates and data migration strategy as follows.
On 6 January 2016, Peter Hall and Barry Speed meet to discuss various issues including Data Migration. Barry Speed
indicates that he is arranging a face-to-face meting the following week to review the Data Migration strategy and outline
plan, and to review the status of extracts with David Legge in preparation for that: see JB 709.
Page 186 ⇓
186
On 7 January 2016, Martin Spellman emailed David Legge to check how he is progressing with data extracts. David Legge
emailed back to say that ‘I still don’t have access to ANY of the legacy systems. Its getting beyond a joke now.’: JB 716
On 12 January 2016, the meeting discussed the week before takes place between Peter Hall, David Legge, and Barry
Speed: see Ravinder Johal’s WS at para. 4.15.6 and JB 722. In that email chain, David Legge plainly recognises the Solution
Design Documents as being the current data migration strategy. This meeting was used as part of knowledge transfer and
involved. Ravinder Johal recalled “discussing and explaining the templates to him. I recall that he understood them and only
raised routine queries with them at that time which I was able to address.”: Para. 4.15.6 of Ravinder Johal’s WS. Following that
meeting, Barry Speed takes over escalating matters with BT from David Legge: see the WS of Niall Sheriffs at para 3.64.
On 14 January 2016, BT email CEC to ask them to confirm that access to data can be given to David Legge under his
request. Derek Masson of CEC asks Niall Sherriffs whether it has gone through the approved process.
Between 15 and 18 January 2016, in light of the ‘unfortunately cumbersome process’, David Legge starts to take steps to make
a request for live data: see JB 732
On 21 January 2016, Martin Spelman emailed, again, on having heard that David Legge had obtained access to the legacy
systems. David Legge explained that he had managed to use the Transoft tool to extract certain data, but goes on to say
Getting access to HR/Payroll and Finance still seems to be problematic. Looks like the issue is around the fact that they don’t want us
working with personal data and they do not have test instances that have been anonymised.’: JB 738. David Legge, also, emailed
Page 187 ⇓
187
Ravinder Johal looking for drop dead dates for data. Again, Ravinder Johal made it clear that she did not want
anonymised data and that there was ‘no reason why I cannot see that as migration will need to any way for the live tests and
parallel runs’: JB 739.
February 2016
At the beginning of February 2016, it was confirmed that David Legge had access to Oracle and iTrent. But the access
granted to him by BT was such that he was not actually allowed to do any extracts and was read only. His access was
conditional on not accessing live data other than to anonymise it: see emails at JB 790, and 796.
On 13 February 2016, Tom Fulda issues his report to Niall Sherriffs and Barry Speed: JB 808. The Court will recall that Mr.
Fulda was taken through the broadly positive statements that he made in that report in relation to the ‘extensive
documentation’, the lack of any criticism of the absence of a data migration strategy or issues with the templates in cross-
examination. This was, entirely, consistent with the fact that on 16 February 2016, following receipt of that report, Dr.
Speed emailed his updated actions for the ERP Workstream in which he noted that ‘Data migration strategy (in Tom Fulda
report). Clarified this is fully covered in the approved design documents.’ Accordingly, again, as they plainly state themselves, it
is clear that everyone at that stage recognised the Data Migration Strategy was contained in the SDD.
On 15 February 2016, the ongoing delay in getting access to data has clearly been escalated to Niall Sherriffs. David Legge
emailed him about the issue and emphasised that there was increasing concern about the impact the delay was having on
the migration timescales. He emphasises the need for live data to Niall Sherriffs: see for discussion in Peter Hall’s WS at
para. 8.42.4, David Legge’s WS at para. 2.22, emails at JB 816 and 818. This is important because it emphasises that the
Page 188 ⇓
188
issue was already escalated to the head of the entire programme within CGI. Agilisys could not have escalated it higher
themselves.
On 22 February 2016, Julia Jackson issues Migration Report to David Camilleri of CEC: JB 837. This is important. As is
recorded at p.11200, the purpose of the document was to address the one outstanding design decision on data migration
noted in the R2R SDD. This is the ‘further meetings’ that was later referred to by Colin Brown in his evidence, which he
claimed had never taken place but which were always intended to be addressed by this document.
March May 2016
By the end of this period, Mr. Legge had left CGI. He still did not have full access to the legacy systems. And he had not
passed any data extracts to Agilisys: see para. 2.38 of David Legge’s WS. As he said in evidence: “We didn't get to the point
of being able to transfer any data to Agilisys.”see Day 13, page 65, Line 6 of the Transcript. Accordingly, very little actual
progress was made with data extraction.
But it is helpful to highlight the following aspects of this period that show that CGI struggled to obtain CEC’s agreement
to get access to live data and that that was the principal cause of delay:
On 18 March 2016, Agilisys issued RN007 relating to Data Migration delays: Peter Hall’s WS 8.43, and see RN007
at JB 16. It was issued then because issued because CGI's delay in providing full data extracts to Agilisys was
going to affect their ability to meet the System Acceptance Testing Milestone. By this stage, CGI was almost 6
weeks late in delivering full data extracts in accordance with the Implementation Plan v 12 deadline of 3 February
Page 189 ⇓
189
2016. This was ultimately accepted insofar as Relief was granted as a result of the ERP re-plan agreed with the
Council at the end of April 2016: see response to RN007 at JB30.
By early April 2016, Kevin Bryce, Stephen Waters, and David Legge were meeting with CEC to try to persuade
them to let Agilisys have access to live data: see JB 993. Following this meeting it was identified by Derek Brown
of CGI (who bears to have been involved in data security), and having discussed it with CEC, that a statement of
why access was required needed to be put together to obtain formal approval. (NB. It is in this context that the
email from Kevin Bryce to Peter Hall is critical of David Legge's performance: "David lacks urgency in everything he
does" Further, Kevin Bryce acknowledged in the email that CGI had missed the data extract deadline in February
2016, and wanted to set a new date of 18 April 2016, which would still be unlikely to hit but would mean they
could push David Legge: see Peter Hall WS 8.42.3, and JB 1001 1002). This meeting was followed by a meeting
between Ravinder Johal, Stephen Waters, and Derek Brown to discuss the necessary issues around access to live
data, and Derek Brown then puts the document together: see references in JB 1038 and the evidence of Stephen
Waters, Day 14, pages 30 and 31 of the Transcript. This document was ultimately rejected and a Privacy Impact
Assessment had to be submitted: see JB 1144. It is in this context that Peter Hall was asked to get the consent of the
Information Asset Owners within the Council on 26 April 2016, and ultimately only obtained full consent on 19
May 2016: see JB 1241. Although there was some attempt by Stephen Waters to suggest there was some inordinate
delay on the part of Peter Hall, it is submitted that that is to simply disregard: (a) Peter Hall’s explanation that ‘The
delay had been that we could not get CEC to sign off on the PIA. Part of the reason for the delay was that some of the most
senior people within CEC had to sign the PIA and it needed more than one signature’: Para. 6.1 of Peter Hall’s Supp. WS;
Page 190 ⇓
190
and (b) the fact that it is so minor in comparison to the inordinate delay to this point by CGI in even getting to the
point of asking Peter Hall to complete a PIA.
In the course of April 2016, Stephen Waters and David Legge bear to have started to try to make more progress. In
particular, it is in this context that there are discussions about provision of further templates to David Legge,
which ultimately were sent on 22 April 2016: see emails in JB 1030, and JB 1141.
There is the sequence that was explored in cross-examination with Stephen Waters and Colin Brown in relation to
the preparation of the plan provided for in the 22 April Conditions: see JB 1152. This plan is particularly notable
because it utterly failed to meet the deadlines of a first extract cycle being completed by 7 June 2016, and a second
cycle by 8 July 2016. It is also notable in that it identified no dependencies on Agilisys for a data migration strategy
or anything else.
In cross-examination, David Legge accepted that the data migration document he had created in May 2016 was a
combination substantially of material taken from the SDDs joined together particularly with some content written
by himself about anonymisation: see Day 13 at Page 27, Lines 1 7 of the Transcript.
In addition, it is worth also noting in this period that, on 9 March 2016, Julia Jackson sent an updated Migration document
to CEC having been discussed at length with CEC in February: Peter Hall’s WS 8.41, and JB 897. By the end of the month,
they had indicated that CEC would accept the Julia Jackson document’s recommendations if CGI would retain a copy of
Page 191 ⇓
191
Oracle post go live for reference. Niall Sherriffs indicated that their may be a licencing issue but that he would investigate
and requesting a copy of the document: see see Peter Hall WS at para. 8.41 and JB 971. The same explanation is provided
by Ravinder to David Legge on 14 April 2016: see JB1044. Again, the issues in the SDD had been agreed with CEC subject
only to CGI confirming it would retain a copy of Oracle. Accordingly, Colin Brown’s evidence on these meetings and
discussions not having taken place was simply wrong.
June 2016
This period is notable in the evidence for the, almost, complete absence of any evidence of CGI doing anything in this
period in relation to data migration. The evidence bears out Agilisys impression that CGI struggled to obtain resource to
replace David Legge after he left and nothing much seems to have happened in this period: see Ravinder Johal’s WS at
para. 4.15.6. It is not until Graeme King’s involvement in mid-July 2016 that there is anything obviously concrete that
happens.
July 2016
October 2016
The sequence of events in this period of time was explored with Colin Brown in cross-examination on Day 17. It is most
notable for the fact that there is an intensive period of ‘special measures’ that starts to commence near the end of July and
continues throughout August. Despite this, there is no point until the oblique reference to templates in the letter of 29
September 2016 (JB 42) and the more detailed response to RN016 in October 2016 (JB 35). In particular, no complaints of a
lack of a data migration strategy, no complaints that the templates were not fit for purpose, and no complaints that they
could not undertake their task because of something Agilisys had done or not done. That is remarkable if CGI’s case is to
be believed.
Page 192 ⇓
192
In summary, the sequence was as follows:
There was a workshop between CGI and Agilisys on 2 August 2016 to discuss and agree how to deliver data
extracts. Gerard Breukink emailed to advise that they were doing extracts based on templates provided by
Agilisys, but there seemed to be little guidance from Agilisys regarding which specific tables and fields the data
needed to be taken from: cf. Ravinder Johal comments in her Supp. WS at para. 6.3. Graeme King volunteers to be
the ‘technical lead’: JB 1514.
On 3 August 2016, there was an email from Graeme King of CGI. It is plain that, as with all previous data
migration individuals, Mr King has looked at the design document for HR and found that it states what will be
migrated and that this has been agreed by the Council: JB 1519 at page 15425. The CGI team now working on data
migration, therefore, had the matter of the templates, the matter of the scope of the data to be migrated, and data
mapping all in view
On 5 August 2016, there are a series of emailed queries about the templates between CGI and Agilisys. Peter Hall
chased Ravinder Johal to respond quickly, which she does the same day. She, also, sends on the relations values.
This latter point is particularly significant given that CGI later claimed not have seen them or understood them at
all. It was plain from Mr. Brown’s evidence that he had not been aware that they had been shared with the CGI
team at this point: see JB1522, 1524, and 1528
Page 193 ⇓
193
From 6 August 2016 onwards, there is a flurry of special measures such as daily meetings, further data migration
workshops, and daily calls: see JB 1532, 1533, Colin Brown WS at para. 3.26, and 3.27.2.
On 17 August 2016, Colin Brown emailed a plan for data migration to Peter Hall, and the following morning Peter
Hall asked for a call to go through it: JB 1558. There are then further action plans by Stephen Waters following yet
another data migration meeting: JB 1560.
From 30 August 2016, there are daily data migration reports: see first report on JB 1570, and subsequent one on JB
1583. By this point the data migration strategy, data migration plan and templates have all been addressed
intensively over 4 weeks. It defies all credibility that there was some hidden problem that CGI had not identified.
Colin Brown’s evidence on this was strident, but ultimately he gave no credible answer. He appeared to suggest
that CGI had agreed to get on with the method that Agilisys had stipulated. Even if that were true, it would be
incredible if there was an aspect of that method that was so obviously flawed and was actually causing CGI
problems that it had not been identified.
Between August 2016 and September 2016, it is clear that both teams were working very hard at Data Migration:
see Ravinder Johal WS at 4.15.9 4.15.17. However, CGI simply struggled to complete the work in the timescales
they had set themselves and were still working on them into late October 2016: see Ravinder Johal WS at para.
4.15.26.
Page 194 ⇓
194
In October 2016, Agilisys issued RN016 relating to the delay in data extracts: see RN016 at JB 25. CGI responded on 27
Octber 2016: see Response at JB 35. It is worth noting about the response to RN016 that it refers to ‘limited detail within the
Data Migration Strategy’ so that even at that point CGI recognised the SDDs as containing the data migration strategy.
Kevin Sander’s gave evidence explaining precisely why the issues raised in the response to RN015 were correct: see Kevin
Sander WS at para. 5.9.1 5.9.5. The points he makes are manifestly borne out by the chronology. He correctly notes that
all the information required had been provided during discussion and knowledge transfer sessions held with CGI’s David Legge at a
much earlier stage in the project when the templates were originally reviewed.’ The knowledge transfer in August 2016 could be
added to that point as well.
October 2016 There has been limited exploration of the detail of data migration in this period of time at the Proof and it is largely
March 2017
irrelevant as post-dating the last Relief Notice within the scope of the Proof. CGI place weight on it because of the
activities of Kevin Sander in this period, which are said to represent what should have happened originally.
The best evidence in relation to that matter is what Kevin Sander himself says:
In my view, [the work he undertook] became desirable in this case only because by October 2016 the relationship between Agilisys,
CGI and the Council was perceived by the Council to be working so badly, and because so much of the earlier discussion and
knowledge transfer had been eroded by the passage of time, coupled with high turnover of CGI extraction personnel. My motivation
for writing a new document was only to restore confidence with Council stakeholders. I perceived that consolidating all the already
available information into a single document and using that as a basis for review was the quickest and most efficient way to achieve
that.”: Kevin Sander WS at Para. 6.5.
Page 195 ⇓
195
ARCHIVE SOLUTION
In the Notifiable Default Notice served by CGI on Agilisys on 8 November 2016 (within JB 46) and in the pleadings an issue has been taken
about archive solution not being available, the detailed evidence on this point can be found as follows.
First, the strategy for archiving was clearly set out in the SDD and other documents as being a separate instance of Business World that would
hold historic archive data. The strategy for archiving is specifically covered in each of the SDD: see JB 234 241, particularly at pp. 5171 5176,
5289 5293, 5410 5415, 5454 5457, 5615 5620, 5, and 5831 5834. Further, the SDDs set-out in detail the number of years data to be kept and
which is going into live and archive.
Secondly, it is clear from the SDD that archiving was due to be done after the planned Go-Live date: Archiving would be done after Go Live:
(a) the ERP Project Board Slides on 22 October 2015 at JB 73; (b) PID on 12 November 2015 at JB 147 at p. 18; and (c) Solution Design
Documents, above, shows Archive and clearly record as a design decision reached on 20 October 2015 and later signed off as being a separate
BW instance.
Thirdly, as it was always intended to be a separate historic instance of BW, it did not need some sort of bespoke design document. It was
intended to be a further installation of BW and it was for CGI to ‘provide the archiving solution, including hosting and licensing. The archive solution
will be configured by the Supplier and the Supplier will import the archive data provided by CGI…’: see Number (16) of Project Roles for Agresso ERP
in Roles & Responsibilities document in Schedule Part 4.1 (JB 1 at p. 1076).
Page 196 ⇓
196
Fourthly, it is clear that CEC changed its mind on both the levels of historic data and whether the archive was to be available at Go Live or not
in the course of July October 2016. The sequence can be seen as follows: (a) the beginning of the shift in position in July 2016 as per the Project
Board Meeting on 27 July 2016 in JB 96 97, and subsequent email on 2 August 2016 at JB 1535; and (b) it is then a recurring them in September
2016 Project Boards and is properly focused in October and November 2016 Project Boards: see JB 120 at p. 3080, JB 122 at p. 3130, and JB 124 at
p. 3144. It is telling that much of CEC’s change of position in this period was then simply passed back to Agilisys by CGI without any
recognition that it represented a fundamental change and departure from the agreed scope and design decisions reached a year previously.
Page 197 ⇓
Date
September 2015
October 2015
197
APPENDIX 2
ERP ENVIRONMENTS CHRONOLOGY
Chronology
On 15 September 2015, a meeting takes place between CGI and Agilisys ‘running through all the dependencies etc.’ and there
is an immediate concern because CGI ‘are not remotely ready to set up an environment’: email from Tim Pitts to Nigel
Muirhead at JB 449. Claire Conaghan emails Grant Lithgow and Richard Salmond of CGI to suggests that it ‘makes sense
that Unit 4 define the infrastructure. She stresses the timelines are very worrying, and that a development environment
will be needed by start of October 2015. There is no push back from members of CGI and nobody suggests Agilisys
should be doing this on CGI’s behalf: Claire Conaghan’s WS at Para. 15.2.2, Niall Sherriffs’ WS at Para. 3.8, and Emails at
JB 451. On the same day, Agilisys issue a Workstream update that identifies that there is a risk over the timeline for BW
infrastructure build and installation, with an action on Richard Salmond to confirm the timelines and that Unit 4 have
been commissioned: JB 154 at p. 3750.
Between 21 23 September 2015, Niall Sherriffs requests a quote from Unit 4 and indicates that once they have provided
it then they can agree a schedule for installation: see Niall Sherriffs’ WS at Para. 3.11 and Email at JB465. This ultimately
takes until 1 October 2015 to get resolved: see JB 477 at pp. 8116 8118. On 23 September 2015, Niall Sherriffs exchanges
emails with Claire Conaghan to confirm he has progressed the installation quote. There is a clear disagreement about
timescales between Niall and Claire Conaghan, but no issue is raised that Agilisys need to provide CGI with anything or
should be managing Unit 4: see JB 467.
On 29 September 2015, Peter Hall issues a summary plan for ERP to various CEC and CGI staff: see email at JB 470.
On 1 2 October 2015, there is an exchange of emails between Grant Lithgow of CGI, Claire Conaghan, and Niall
Sherriffs. Claire Conaghan explains that ‘Unit 4 are doing the installation, they need to provide the specification so they take full
ownership. We do need an ETA when the environment will be ready’: JB 478 at p. 8122. Niall Sherriffs confirms that Unit 4 will
be doing the installation, and that he is going to catch-up with Raj [Sudra] and Grant [Lithgow] of CGI about the timings
and approach. There is no suggestion that CGI are expecting Agilisys to be involved. The next day Claire Conaghan again
Page 198 ⇓
198
asks for a confirmed ETA and explains that if they do not have the environment by the end of October they are likely to
have significant issues delivering the milestone: JB 478 at p.8121. Niall Sherriffs promises to chase and keep on top of the
team and get a date as soon as possible. Again, there is no suggestion anything is needed from Agilisys at this stage: JB
478 at p. 8120.
In the same period, Niall Sherriffs emails Grant Lithgow and Raj Sudra asking them to get in contact with Unit 4 to ‘find
out what spec is needed and then when and how we can configure the relevant infrastructure’: JB 477 at p.8114. Grant Lithgow
exchanges emails with Unit 4 to arrange a phone call to discuss Unit 4’s ‘requirements for the test and dev enviromments’.
Again, there is no suggestion that CGI need anything from Agilisys. However, Peter Hall emails Unit 4 directly and
indicates to both CGI and Unit 4 that ‘Installation of the dev system is becoming an urgent issue for my delivery team. We have a
drop dead date of end October to have this in place or it will significantly impact our ability to meet our contractual milestones.’: JB
480.
On the following Monday, 5 October 2015, Unit 4 send a series of documents to Grant Lithgow including BW Technical
Specification, and the sizing questionnaire and spreadsheet: see JB 508.
On Wednesday, 7 October 2015, there is an exchange of emails between Grant Lithgow, Ravinder Johal, and John
Davidson about proceeding with a minimum specification for the development environment. Unit 4 ask for an eta from
CGI on when the servers will be ready and whether it will be an onsite or remote installation. There then appears to have
been a call. This is the first mention of internet connection because Unit 4 states that ‘The web client can work over the
internet using standard https though security advice would be needed once sensitive data has been imported into the system.’: JB 496.
On Friday, 9 October 2015, Ravinder Johal chases Grant Lithgow again for an eta on the servers and whether parties are
on track for the end of October. She expressly mentions that this is important for ‘milestone deliveries’. She mentions that
we want to ensure that Unit 4 resource is booked as quickly as possible as we normally have to give several weeks of notice for this.’:
JB 496.
On Tuesday, 13 October 2015, Claire Conaghan and Niall Sherriffs exchange emails about finding a solution that does not
involve using the BT data centre. This leads to discussions with Unit 4 and CGI about possible alternative hosting
methods using Unit4hosting or CGI providing Azure hosting: JB 512.
Page 199 ⇓
199
On 26 October 2015, Raj Sudra and Huw Williams of CGI provide an update on the development environment. Raj Sudra
recognises that this is critical and asks Huw Williams to outline, inter alia, the ‘Dependencies and timescales on Agilisys / Unit
4 / BT / Council etc. that you are dependent on to achieve our deliverables.’ Huw Williams reports on progess, and he does not
identify any dependencies on receiving anything from Agilisys: JB 525.
On 30 October 2015, Huw Williams of CGI sends out an update stating that ‘We have provisioned our VM’s as per the
attached templates’ and that ‘We have requested bandwidth requirements from our third parties.’ Again, there is no indication
they are waiting on or expecting anything from Agilisys: JB 531.
November 2015
On 4 November 2015, there is an internal CGI email chain about the installation of the development environment
following the offer from Unit 4 of installation dates in mid-November 2015. The general subject matter is a concern that
there is no design or blue print for the development environment. Raj Sudra emails to say that ‘I’m not totally sure who is
managing the end to end deliverable - from speaking to Nial [sic] understand this is Agilisys but because of the commercial
relationships we are managing Unit 4 for their ‘specific’ deliverables. However, their statement of work is not clear in respective [sic]
to providing design documents, their requirements, how many servers they would need, software licensing for SQL, OS etc…’: JB
542.
On 5 November 2015, apparently motivated by the internal email chain the day before, Huw Williams emails various
people from CGI, Agilisys and CEC asking whether there has been any work done on the design or requirements by Unit
4 or Agilisys. Huw Williams records that ‘We have been informed that internet access from the dev environment is required, we
will need to discuss how we do this because there is not a shared proxy service in CCC at present.’: JB 543. On the same day, Huw
Williams chases and is sent a copy of Unit 4’s Statement of Work that requires to be completed prior to installation: see JB
617 618. The Statement of Work mentions a number of items that later became controversial and claimed to have never
been required before such as Microsoft Office and Excel (at p. 9764), Supported Web Browser (p. 9769), BIF and the
Report Engine (Excelerator and Worderator) (at p. 9769). It does not appear that CGI ever completed the Statement of
Work as it was sent in its incomplete form to Agilisys later in November 2015: see JB 617.
Page 200 ⇓
200
At this point in November 2015, Sebastian Wood appears to become involved. On 11 November 2015, there is a meeting
between Peter Hall and Sebastian Wood to discuss environments. Peter Hall emails Sebastian Wood a list of the users that
need access to the Development Environment and their levels of access: see JB 556. Peter Hall, again, mentions Excel will
require to be on the server. On 12 November 2015, Sebastian Wood emails out a weekly update for Infrastructure: JB 557.
It does not list anywhere a concern that Agilisys have not provided a design or a specification. There are, however,
planned meetings for the following week. It records that Unit 4 are going to do the installation the following Monday, 15
November 2015. It, also, records that CGI and Agilisys intend to complete the Unit 4 questionnaire together. It, also,
records a plan of meetings with Agilisys to discuss their requirements, which is consistent with the email from Seb Wood
for a meeting on the same date: see emails in JB 558.
In the week commencing 16 November 2015, there are a series of workshops, emails and meetings involving CGI and
Agilisys to discuss and clarify requirements on environments. These are for the purpose of enabling Sebastian Wood to
prepare a High Level Design as mentioned in the previous weeks update report. In particular: (a) a meeting on Tuesday,
17 November 2015, between Ravinder Johal and Sebastian Wood as per the weekly report. This is spoken to by both
Ravinder Johal at para. 7.2.2 of her WS. Paul Hamilton confirmed in evidence that this meeting had taken place: see Day
11, Line 6, of the Transcript. It is, also, clear that Claire Conaghan emailed Sebastian Wood indicating, inter alia, Microsoft
office excel & world on the apps server’ would be needed: JB 577; (b) although Mr. Hamilton claimed it was earlier, there is a
specific email introducing him to Agilisys on 18 November 2015 as coming on board: see JB 574, and cf. Day 11, pages 47
– 50, and on the same day Seb Wood chases for the ‘environments document’ as soon as possible; and (c) on 19 November
2015, Agilisys provided CGI with a simple environments matrix: see JB 573 576, and on the same day Paul Hamilton
emails Peter Hall to ask for a meeting to ‘iron out some of the challenges we are currently facing…with environments’: JB 574.
There is a telling email exchange where, despite the fact that the sizing questionnaire was sent directly to CGI in October
2015, Paul Hamilton is unable to find it and emails Peter Hall to get it: see JB 580.
Although work started on installing the necessary Entrust software on Agilisys laptops towards the end of November
2015 to provide access to the Development Environment, CGI suffered a malware attack and all forms of remote access to
the data centre were severely restricted for a number of days: see Paul Hamilton’s WS at para. 2.15, and see JB 599 about
setting up of accounts.
Page 201 ⇓
December 2015
January 2016
201
In early December 2015, CGI started to provide access.
In that period of time, and contrary to the evidence from a number of CGI’s witnesses, Peter Hall indicated he would
have been willing to accept CGI laptops if that was a way around the problems: see Email of 4 December 2015 at JB 633
Finally, access was provided on or around 9 December 2015. As soon as access was granted, Agilisys mentioned: (a) that
none of the Microsoft products were installed (those having been mentioned on a number of occasions previously in
October and November 2015, as above); (b) a specific part of BW that had been licensed as part of the BW purchase called
Invoice4 Onsite; (c) only 2 concurrent users can access the server (despite the fact that Peter Hall had sent an earlier email
in November 2015 outlining all of the people that needed access); and (d) there were still logins outstanding for the rest of
the Agilisys team to get access (again despite the individuals details having been provided about a month earlier on 11
November 2015): see JB 652. A number of other issues were then identified between 10 December 2015 and 17 December
2015. The main one being internet access (which as pointed out above had actually been raised by Unit 4 in October 2015
and again specifically with CGI in early November 2015, but which seems to have gotten lost sight of by December 2015)
Mr. Hamilton described the internet access as being a new requirement, but it is submitted that that is manifestly wrong,
and when the earlier correspondence was put to him his evidence became very confused and indeed contradictory: see
Day 11, p. 69 72, of the Transcript.
Mr. Hamilton gave evidence in cross-examination that he accepted the need to overcome CGI’s internal security protocols
to enable internet access was a CGI issue and said: “It was, yes. My only additional comment to that would be if we'd known the
requirement was needed earlier, we could have dealt with it a lot sooner.“: Day 11, page 69, Line 20 in the Transcript. It is
submitted that answer is correct the only issue being that CGI did know earlier even if Mr. Hamilton did not.
In early January 2016, there were meetings and discussions about the ongoing difficulties with environments. This led to
the decision that Agilisys would host their own Azure environments to allow the build work to start commencing at
CGI’s expense: see Peter Hall WS at para. 7.10, and emails at JB 709. Peter Hall immediately emailed Unit 4 about setting
up an Azure environment and it was completed in a couple of weeks: see the WS of Peter Hall at para. 7.10 and the
evidence of Peter Hall on Day 3, pages 143 and 144 of the Transcript. The quote was later provided in February 2016 and
paid by CGI: Email on 8 February 2016 at JB 794, page 10874 (Quote on page 10876 is for £39,101 plus VAT).
Page 202 ⇓
202
On 27 January 2016, Agilisys issued RN001 seeking relief as a result of delay of access to the Development Environment:
see RN001 dated 27 January 2016: JB 10. As Peter Hall explained, ‘In this type of project, there is no way that you can lose 6
weeks out of a 3 month window and still deliver to schedule. CGI were well aware that this Relief Notice was coming and had to be
raised.’: para. 7.12.
February 2016
CGI issue response to RN001 on 10 February 2016 accepting CGI Cause, and granting relief in the form of postponing
Milestone 2.2 to 31 March 2016 and recording that the Transformation Plan had been amended to version 12 of the Project
Plan: see Response to RN001 dated 10 February 2016: JB 26
On or around 16 February 2016, the ERP Development Environment is finally provided: see the Lessons Learned slides on
23 February 2016 at JB 842, email of 4 March 2016 at JB 881, and Kevin Bryce’s plan of 14 March 2016 at JB 926.
On or around 23 February 2016, the Lessons Learned exercise is undertaken in relation to the Development Environment.
It does not identify any issues outstanding or required from Agilisys, and portrays a picture that all matters have been
resolved.
In late February 2016, Kevin Bryce is recruited by CGI for the purposes of being an Environments Manager for the ERP
and EI projects. At the point in time he joined, and as recorded in the response to RN002, the relevant date in Version 12
of the Project Plan was 31 March 2016. Mr. Bryce was, plainly, aware of that date when he initially started: see the
reference to these dates in his email of 3 March 2016 at JB 875. He accepted this under cross-examination: see page 48 49
on Day 12 of the Transcript.
It is a notable feature of the evidence that almost nothing seems to have been happening with progressing build of the
Pre-production and Production environments between December 2015 and February 2016. It is not until into March 2016
that Kevin Bryce began to plan and begin construction of the environments.
Page 203 ⇓
March 2016
April 2016
203
At the Partnership Board meeting, following a discussion about the issue of responsibility for managing infrastructure, it
was agreed that ‘Agilisys cannot manage some work streams such as infrastructure’: JB 890, and Robert Price supplementary
WS at para. 2.14.
Mr. Bryce was cross-examined extensively on this period of time onwards. Reference is made generally to his cross-
examination from Pages 50 onwards on Day 12 of the Transcript.
In the course of early March 2016, Mr. Bryce proposed dates of 15 and 29 April 2016: see, in particular, the internal emails
of 8 March 2016 at JB 900 and the Plan on a Page produced on 14 March 2016. This immediately meant the date in the
Project Plan was not going to be viable.
On 15 March 2016, Peter Hall assisted Kevin Bryce in completing the sizing questionnaire for the ERP Environments: see
Peter Hall’s WS at para. 7.7 and JB 931, and 938 – 939
It is submitted that it was clear from the cross-examination of Mr. Bryce that his approach was that, following an initial
meeting with Mr. Hall, he had formed the impression (rightly or wrongly given the issue was never put to Mr. Hall)
that he was not going to obtain a document setting out requirements and that Agilisys were content with what had
previously been provided. It was nerve clear precisely what document, or details he was expecting. He, himself, appeared
content to proceed on that basis. As he said: ‘I had initially expected a detailed document of requirements. When we were pointed
to conversations in different documents I was happy based on the timeline that we could try and push ahead with information we had
and just accept the risks.’: Page 118, lines 11- 16, on Day 12 of the Transcript.
On 18 March 2016, Agilisys issue RN009 seeking relief as a result of delay of access to Production and PPE Environments.
CGI experienced a number of difficulties in setting-up the servers and arranging an installation of the PPE and
Production Environments in this period. In particular, a planned installation for 8 April 2016 was cancelled because the
servers were not ready: see Email chain on 7 11 April 2016 in JB 1018. And then, again, due to difficulties with the CGI
server failures on 21 April 2016 in JB 1222.
Page 204 ⇓
May 2016
June 2016
204
Against this background, the 22 April Conditions were sent to Louise Ah-Wong requesting, inter alia, a detailed plan for
Development, test and Gold Environment provisions including the transfer from Agilisys temporary environment…Pre-production
and production environment delivery and their proposed move (if any) to Wales’: see JB 1069 at p. 12726. Niall Sherriffs allocated
Kevin Bryce to produce this plan and it was produced on 27 April 2016: see emails and plan at JB 1163 1164.
Kevin Bryce was cross-examined on the fact that this plan: (a) records that all requirements had been provided by
Agilisys; and (b) places all of the dependencies for design documents on CGI employees. His evidence came to be that he
did not think at this time there were any outstanding requirements based on his meetings and discussions and said ‘At the
time it was based on what I had been told by Agilisys. I accepted that we had sufficient -- sufficient information to build the
environments, so I didn't labour the point at that stage.’: Day 12, page 111, of the Transcript.
PPE-A was finally installed on 12 May 2016: see email of that date in JB 1163 1164.
As Mr. Bryce eventually accepted in cross-examination, they remained effectively accessible until into June 2016. He did
not attribute any delay in this period to Agilisys either in his Witness Statement of Cross Examination. The progression
can be seen in the email in JB 1246, and the further environments plan in JB 1271 targeting new dates for availability.
The delay by this point because of access issues with the VPN had become significant.
Agilisys issue RN013 seeking relief as a result of delay of access to hosted PPE on 3 June 2016: RN013 dated 3 June 2016:
JB 22. This was a direct result of the ongoing environment difficulties.
CGI accepted in its response CGI cause and agreed to postpone the SAT Milestone of 11 July 2016. They agreed to work
together to understand and agree the extent of the relief granted: see Response to RN013 dated 10 June 2016: JB 33
At the same time, in the beginning of June 2016, Agilisys readily agreed to the suggestion that access issues could be
overcome by the use of CEC Laptops. There is not the slightest hint of reluctance on the part of Agilisys. The email
correspondence is in JB 1281. Peter Hall provided details of what was to go on the laptops in terms of software including
Excelerator, Worderator, BIF and Agresso: see in particular email exchange between Peter Hall and Kevin Bryce on 7 June
2016 at JB 1306.
Page 205 ⇓
205
It took a number of weeks to provide CEC laptops to Agilisys for the purposes of accessing the Environments. There were
still outstanding laptops into July 2016. The process was also disrupted by the data centre move at the end of June 2016:
see WS of Peter Hall at para 10.22, JB 1396 and the evidence of Peter Hall on Day 3, page 145 of the Transcript.
July 2016
Proper access to PPE-A was provided in early July 2016. Agilisys checked the environments and provided a snagging list
to CGI. In his Witness Statement, Kevin Bryce made much of this list as an example of drip feeding of requirements.
However, from his cross-examination, it was clear that: (a) a number of the software issues he identified as new
requirements were the Reporting Engines (Worderator and Excelerator) and BIF being on both the server (client side in
his terminology) and laptops. It was pointed out to him that, again, these had in fact been identified in the Statement of
Works as far back as November 2015, which he ultimately accepted but said he had not seen. As he himself said: “So based
on this document, they certainly had been stated previously.”: Page 140 of Day 12 of the Transcript.
Work on the production environment had been abandoned and it was still not available at this stage. PPE-B became
available early in this month.
Agilisys issue RN014 seeking relief as a result of delay of access to all of the Environments on 26 July 2016: see RN014
dated 26 July 2016 at JB 23.
In the course of July 2016, the major issue was network latency. Mr. Bryce accepted this had nothing to do with Agilisys:
see Day 12, Page 151, of the Transcript. This was not ultimately resolved until into September 2016 by a temporary
workaround.
September 2016
In or around September 2016, the temporary workaround was recommended by Advance 7 to get around the latency
issue: see JB 1608. There, also, continued to be other issues arising on the environments with things not functioning,
largely due to the absence of environment management procedures being in place. There also continued to be issues with
email functionality being rolled back: see JB 1666 and Peter Hall’s WS at Para. 11.5.
Page 206 ⇓
October 2016
206
Agilisys issue RN015 on 13 October 2015 seeking relief as a result of the provision of accessible ERP Environments. It
detailed numerous issues with the Environments, but the main issues were matters within the environments that were
not functioning and the latency issue.
CGI issue RN015 on 27 October 2016 rejecting claim for relief: see RN015 dated 13 October 2016: JB 34, and response to
RN015: JB 34. The response largely accepts the defects and, indeed, records in detail the delays that had occurred in
resolving. The core defence seemed to be the suggestion this was not preventing Agilisys from moving to UAT, albeit CGI
Cause is denied.
That defence was misconceived for two reasons. First, as Kevin Bryce accepted in cross-examination, it would have been
unsuitable for final UAT sign-off: Day 12, page 160 161, of the Transcript. At that stage, there was no obvious or
guaranteed fix. Secondly, CEC had made it clear they would now allow Agilisys to move to UAT using the workaround.
Page 207 ⇓
207
APPENDIX 3
NORTHGATE API & EI ENVIRONMENTS CHRONOLOGY (RN002 AND RN012), AND AD SUITABILITY ISSUES
NORTHGATE API CHRONOLOGY
Date
Chronology
September 2015
At the pre-kick-off meeting on 14 September 2015, Agilisys made it clear they were going to require the Northgate API.
They explained to CGI that it was ‘fundamental’ and ‘critically important’ that access was provided to Northgate iWorld
API: David Buchanan’s WS at para. 8.5.1 and Claire Conaghan’s WS at Para. 12.8.1.
A target date of 30 September 2015 is identified, but missed: see first Weekly Report on 6 October 2015 for EI at JB246 at p.
5948
October 2015
On 9 October 2015, Niall Sherriffs emails Claire Conaghan and David Buchanan about, inter alia, the Northgate API. He
says that ‘iWorld is progressing technically, we will need to do some test data scrambling, but has some commercial issues for the
Council to resolve as it sits in live…They have already completed a significant amount of work, so it should be available soon after an
agreement on commercials.’: JB 494 at p.8204. Accordingly, at this stage, there are no issues identified with Agilisys and the
delay is being caused by agreement on commercials.
On 20 October 2015, the Weekly Report records that Agilisys were still waiting on a confirmed date for the Northgate
iWorld API: JB247 at p 5944.
On 23 October 2015, Niall Sherriffs emails Matt Stiles of Agilisys and others to say, inter alia, ‘Access to iWorld in the BT
centre would be available from mid-Dec. This is a worst case date and we are looking to bring it forward, but it is highly dependent on
availability of key technical resources in short supply.’ The reference to technical resources is, presumably, either at BT or CGI.
Again, at this stage, there are no issues identified with Agilisys and the delay is being caused by lack of sufficient
technical resource to complete the work. This is in slight contrast to the position in Niall’s email on 9 October.
Page 208 ⇓
November 2015
208
The Weekly Reports for EI continue to note delayed access to APIs in line with Niall’s ‘worst case date’. Agilisys make it
clear to CGI in these reports that they have significant concerns with meeting the Milestones as a result of the delay. See,
in particular:
3 November 2015 JB249 p5958
Estimated due date changed to 4/12/15. Commentary: "CGI planning to have a cloned development API with
anonymised data available before 04.12."
10 November 2015 JB250 p5968
Dates as per 3 November. Commentary: "Re-planning has confirmed that access to production iWorld API will be
required 04.12"
17 November 2015 - JB251 p5972 + p5974
Dates as per 3 November. Commentary: "Re-planning and build assumes production API on 04/12/2015. This is
unlikely and may not be until 01/04/2015. Needs impact assessment re go-live implications"
p5974 "Risks"
Risk Description: A date for being able to access the BT hosted iWorld environment has not yet been established with any
confidence.
Impact: The new platform cannot test the interaction with the iWorld APIs, and therefore will need to go live with an interim
access to iWorld
Actions: Ensure the commercial liability is adequately backed off, and BT given approval to make the iWorld test (and
subsequent production) environment available for development and go live
Page 209 ⇓
209
There is, also, a call recorded in relation to this matter as having taken place. CEC are adamant that Agilisys
cannot have access to a production instance of the Northgate API due to liabilities issues, and CGI are attempting
to get an instance for 4 December 2015: see JB 566 at p. 8779.
24 November 2015 JB252 p5975 + p5977
Dates as per 3 November. Commentary: "Re-planning and build assumes production API on 04/12/2015. Impact
assessment requested of AD on build implications"
p5977 Risks as per 17 November.
December 2015
On or around 7 / 8 December 2015, CGI provided Agilisys with a cloned version of the Northgate API: see Para. 8.5.6 of
David Buchanan’s WS, and Para. 12.8.2 of Claire Conaghan’s WS.
January February Once the API had been provided, there were issues first with the level of anonymisation required by CEC and then the
2016
usability of that test data. David Buchanan expanded on the difficulties with the data anonymisation in cross
examination, where he described the issue as follows:
So they would take -- they would take the names, addresses and all the codings in there and change it to something that the council
felt was suitably suitably opaque. But the problem with that, my Lord, was it meant that -- it meant that anyone who was trying to
use the information had real difficulty in understanding whether it was relevant or not. So it was very difficult to configure
something where you would expect to see Mr or Mrs G Smith at such and such address if that information was changed completely,
and it was simply -- you know, words, names, addresses were scrambled, postcodes were switched around into sort of random
statements. You know, you could see the information was there but you couldn't -- it was not in a format that was usable to allow us
to configure the system.”: Day 6 at Page 53, lines 1 – 16, of the Transcript.
The Weekly Reports for EI record a progressive resolution of the issues in January February 2016:
Page 210 ⇓
210
5 January 2016 JB 256 p. 5988 89
Lowlights: “Data in iWorld still to be anonyised before any testing can complete – City of Edinburgh CIO requesting
changes.
Commentary: “05.01 Initial connectivity confirmed. Test data WIP (extra steps requested by CEC)
David Buchanan was re-examined on this on Day 6 at page 81 84. He explained the anonymisation issue as
follows: “it effectively says that the whole anonymisation process still had to be fully resolved before any further progress
with respect to testing could -- could complete.”: Day 6 at page 83, line 3 of the Transcript.
12 January 2016 JB 257 p.5993 5994
Highlights: iWorld connectivity (development) and anonymised now confirmed CGI 12.01
Commentary: “12.01 Test data now resolved. There may be a v. 6.12 issue that Northgae need to fix.
A usability issue was raised by David Buchanan with Niall Sherriffs by email of 22 January 2016: see JB 741. This
records that the issue was that the test data was causing limited to no testing capability. David Buchanan was re-
examined on this issue on Day 6 at page 83 84. He explained the nature of the practical problem that the lack of a
usable API was causing for configuration and the Conference Room Pilot session.
27 January 2016 JB 258 -
Commentary: “27.01 Test data anonymisation resolved, usability issue. Further test extracts required
In re-examination, David Buchanan expanded on this by saying: “That there were not sufficient test information. So
the -- the test anonymisation process, if you like is described there, the usability issue had been resolved, but there were
further test extracts, and we required more test information to be provided to us. And this comes back to the issue with the
clone. So the clone only contained a certain extract at a point in time and wasn't the same as AD would normally have used.
If they had access to the full system they would know what they wanted to pull.” At Day 6, page 85, of the Transcript.
Page 211 ⇓
211
9 February 2016 JB 259 p.6008
Commentary: “09.02 Test data anonymization issue being resolved with support by Service areas
And it is shown as green.
David Buchanan gave evidence that around this time the issue had been resolved, but that both it and the issue of
environments had caused delay and the initial Milestone would not have been required to move: see para. 8.5.10 of David
Bucnhanan’s WS. Mr. Buchanan was cross-examined on this evidence. The line of cross did not seek to dispute this and
tended only to suggest that it was both of the issues of environments and the Northgate API operating: see page 54 55 of
Day 6 of the Transcript.
Page 212 ⇓
212
Date
September 2015
October 2015
EI - ENVIRONMENT CHRONOLOGY
Chronology
On Monday 14 September 2015, there was a pre-kick-off meeting between CGI and Agilisys. It immediately became clear
that CGI was not going to be able to provide us with the necessary infrastructure in the timescales required: Claire
Conaghan’s WS at para. 12.7.1. This is the same concern manifested in relation to the ERP Environments, namely that CGI
are not remotely ready to set up an environment’: email from Tim Pitts to Nigel Muirhead at JB 449.
On Wednesday 16 September 2015, Claire Conaghan emailed Grant Lithgow and Raj Sudra two documents specifying the
details on connectivity requirements for the infrastructure hosting on Agilisys Digital: see JB 460 – 463. Claire states ‘As
stated, in order to ensure we meet our tight deadlines, we need to have infrastructure in place by the end of September for the team to
start installation. I appreciate this is a challenge and thank you for your support in helping us find a resolution’ This is, entirely,
consistent with the Implementation Plan in the Subcontract that had building activities commencing at the end of
September. This is discussed at para. 12.7.6 of Claire Conaghan’s WS. She states that: “the real issue appeared to be that CGI
had not given any thought to their responsibility to host the AD Platform and where they were going to do this." at para. 12.7.3 of
her WS. These documents covered the specifications for both the AD product itself and the Biztalk environment needed
to support it.
On 21 September 2015, Raj Sudra emailed Claire Conaghan back to say “Further to our conversation today it’s clear that there
are a number of challenges we have to set-up the test/dev environment you need. Whilst we progress with setting up meetings with
BT and looking at alternatives can I please ask the following from yourselves…’ and goes on to request that Agilisys look at
alternatives and also pull together an interface document such as a system context diagram.
By the beginning of October 2015, there was no committed timeframe that had been supplied by BT to CGI for setting up
the Development and Test Environments. Claire Conaghan, accordingly, issued a draft Relief Notice to Niall Sherriffs
saying that ‘any feedback would be appreciated before it goes to the commercial team.’: see email at JB 484 and draft RN at JB 485.
The principal comments of Niall Sherriffs on this RN were that 14 September 2015 had been “the first date this was identified
as far as I am aware, and the timeline for delivery by 30 September was not agreed by CGI’, and also criticising Agilisys for not
Page 213 ⇓
November 2015
213
having called it out. This production is important in understanding the background and context to the response to RN002
referring to a draft Relief Notice and referring to ‘the original environment requirements not being requested with sufficient
notice’: see JB 27. It is obvious from the context that Niall Sherriffs was referring to the fact that CGI had not realised and
was not prepared to setup an environment as quickly as Agilisys needed. It had nothing to do with requirements not
having been provided by Agilisys.
On 9 October 2015, Niall Sherriffs emailed Claire Conaghan and others to say that he had just had a call with BT and that
the development environment, this should be available Friday 16th am (the current servers need to be upgraded to Win 2008).’: JB
494. There is absolutely no indication that CGI or BT were waiting on, or needed, anything from Agilisys. This deadline
was, subsequently, missed.
On 22 October 2015, it is plain that the Agilisys development environment is in the process of going through the BT
request process. See JB 521.
On 23 October 2015, there had been no progress. Niall Sherriffs emailed Agilisys to say that ‘We had an update from the
latest discussion about when we could arrange environments and more importantly connectivity with BT, and the answer I got was
not going to be acceptable. So could you please go ahead with the set of environments. We have a route to resolve the commercial issue
which was holding us up, but the technical delivery of connectivity meant we would not be able to use this approach.’: JB 523.
By 2 November 2015, Agilisys had setup the two development and test environments themselves: para. 12.7.15 of Claire
Conaghan’s WS and Weekly Update for EI 3 November 2015 JB249 p5958 (Action 4 EI Infrastructure to be built by
Agilisys Digital showing as Blue (complete) with an estimated due date of 02/11/2015). CGI agreed to pay Agilisys for
these: see Invoice in JB 2127. Again, as with the ERP project, it is striking that Agilisys were able to setup environments
within a matter of weeks, whereas CGI had been unable to do so at all for the prior 6 weeks.
However, insofar as Biztalk environments to support EI, and the more permanent Environments were concerned, they
remained with CGI. They followed essentially the same path as the ERP Development environment. This can be seen by
comparing the narrative of RN003 and RN001: see JB 10 12. Reference is made to that chronology, but for summary
purposes it was as follows.
Page 214 ⇓
December 2015
January 2016
214
As is recorded on 5 November 2015, CGI held an internal meeting appointing Maciej Olejnik as project manager and
Sebastian Wood as the primary architect for both ERP and EI Environments (including Biztalk).
Around 11 November 2015 Sebastian Wood held meetings with Shepherd Chengeta of Agilisys, and planned to engage in
an EI and Biztalk environment scoping meeting on 16 November 2015: see weekly update for Infrastructure: JB 557. The
same potential issue about internet access being required for access to the EI Biztalk environment is raised, specifically by
the Biztalk developers employed by CGI for the purposes of doing the Biztalk redevelopment: see JB 543 at p. 8573.
On 8 December 2015, Shepherd Chengeta emailed Richard Salmond saying that he was ‘greatly concerned by the pace at
which the provisioning of the Development Environment is going. Even the revised date for the availability of this environment (27th
November 2015) has come and gone without success. Its my understanding that outstanding issues are: 1. No internet access on the
actual Biztalk servers, which will allow service calls to be made 2. Not enough storage on one of the terminal servers…3. One of the
developers (Jim McLay) is yet to gain access to the servers. I have a provisional estimate of over 3 months for two developers.’: see JB
648 at p. 9960. Paul Hamilton replies and apologises for the issues they are experiencing. Shepherd replies to say, among
other things, that the delay may mean that they need a further Biztalk developer resource: see JB 648.
These issues are captured by Paul Hamilton in his snagging list, as per the discussion in the ERP Environments
chronology. Again, it will be noted that despite the fact that Agilisys had indicated that the EI Environments, including
Biztalk, would require internet access from as long ago as October 2015 the issue appears to come as a surprise to Mr.
Hamilton.
On 21 December 2015, there is an exchange of emails due to the fact that, although internet access was initially activated
in December 2015 following the requests above, it was then switched off due to CGI internal security protocols: see JB 704.
Shepherd Chengeta advises that, in relation to the EI Biztalk environment, ‘Internet is a critical piece for the EI work and the
developers cannot progress until this is resolved.’: see p. 10276 of JB 704. This same internet connectivity issue appears to have
stalled any progress from this point until late January 2016 / early February 2016, as per the ERP Development
Environment.
On 28 January 2016, and one day after RN001 in relation to the ERP Development Environment, Agilisys issued RN003 in
relation to delays to the Biztalk server: see JB 12.
Page 215 ⇓
February 2016
March 2016
215
On 11 February 2016, CGI issued its initial response to RN003: JB 11. It denied responsibility as it said the issues had not
been escalated early enough.
The Weekly Update for 12 February 2016 records that “Access to Biztalk environments now available, issues still to be resolved
(refer to daily tracker for detail)
On 17 February 2016, Robert Price replied to CGI about RN003. He provided a detailed timeline summarising much of the
above. And he said ‘It is simply nonsense to suggest that we haven’t escalated this before 19 Jan: see JB 821. It is submitted that
that was plainly correct given the history outlined above.
The access to the Biztalk Development Environment continued not to be available until around mid-February 2016: it
disappears from the Weekly Highlights in the Report for the week ending 22 February 2016 at JB 261, and see generally
David Buchanan’s WS at para. 8.6.2 – 8.6.10.
There was, at this stage, still no Biztalk Test Environment in use though, as that did not become available till much later:
see Para. 8.7.14 8.7.24 David Buchanan's WS.
At the end of February 2016, Agilisys raised concerns over the quality of the Biztalk development resource: see para.
12.9.8 of Claire Conaghan’s WS. Although this is denied by Richard Salmond, he had no very clear recollection of events
during the project. And this would be consistent with the fact that access had only just been enabled to the Biztalk
infrastructure and so would have been the first opportunity for an issue of this nature to emerge.
In late February 2016, Kevin Bryce is recruited by CGI for the purposes of being an Environments Manager for the ERP
and EI projects.
On 3 March 2016, Kevin Bryce emailed a number of internal CGI people to record the fact that he was working to a
deadline of 31 March 2016 for the Test and Production Environments for EI: see JB 875. It is a striking feature of his
evidence in his witness statement that he makes no specific statements about the progress of EI or his involvement in
those environments, particularly as detailed documents were provided at the outset and the issues experienced were
different.
Page 216 ⇓
April 2016
May 2016
216
In the course of early March 2016, Mr. Bryce proposed dates of 1 April 2016 for the EI Test and PPE Environments and 29
April 2016 for the Production Environments: see, in particular, the internal emails of 8 March 2016 at JB 900 and the Plan
on a Page produced and sent to Agilisys on 14 March 2016 at JB 926, and discussed by David Buchanan at para. 8.7.11 of
his WS. Agilisys relied on the plan in planning its activities, although it was also being asked by CGI to explore the
possibility of it hosting the AD product: see Niall Sherriffs’ WS at para. 4.62. Agilisys’ witnesses were consistent in
speaking to the fact that they considered they were waiting on information from CGI in order to explore whether that was
possible, and there is no evidence that Agilisys consented to defer all work whilst this option was explored: see David
Buchanan WS at para. 8.7.12.
On 22 March 2016, the EI-Web Project Board formally agreed a re-plan to move the milestone date for Drop 1 to 16 June
2016: see David Buchanan’s WS at para. 6.2.
CGI missed the deadline of 1 April 2016 for the EI Environments. Accordingly, on 6 April 2016, Agilisys issued RN010
relating to a delay in access to the EI & Biztalk test environment: see JB 19.
No clear evidence was given on this point but in April 2016 at the Partnership Board meeting stated that ‘work has stopped
on this [i.e. the EI Environments] while waiting for the quote [from Agilisys]’: see JB 2104 at p.20885. It was discussion on this
point between Bob Price and Niall Sherriffs that caused the latter to leave the meeting.
On 16 April 2016, Kevin Bryce provided an update, inter alia, indicating that the EI Test environment was being targeted
for 22 April 2016, and PPE and Production for 29 April 2016. These deadlines were, again, missed. There is nothing, in the
witness statements or elsewhere, that suggest CGI were remotely waiting on Agilisys for anything.
Against this background, the 22 April Conditions were sent by Louise Ah-Wong requesting, inter alia, that CGI agree to
RN003: see JB 1069 at p. 12726.
On 1 May 2016, CGI issued a composite response re-addressing RN003 and addressing RN010. They granted relief and
formally agreed to push the Milestones back consistently with the agreement reached in March 2016.
Page 217 ⇓
217
On 6 May 2016, CGI issued an Environments update indicating that they were aiming to have the EI Test, PPE and
Production Environments built by 11 May 2016: JB 1202, and see Claire Conaghan’s WS at para. 13.3.4.
On 10 May 2016, David Buchanan recorded that in his discussions this morning with Keith B [Bolton] CGI were not
confident they would hit the deadline of the following day: JB 1202.
On 11 May 2016, Martin McNeney and Stephen Waters had a call. Martin stated ‘In summary The EI SIT environment is
NOT YET ready. This is still a work in progress (i.e. the CGI target handover COB Tue 10th May has now come and gone.’: JB 1206.
Against this background, on 12 May 2016, Agilisys issued RN012 seeking relief from delay of access to the EI Test, PPE,
and Production Environments: see JB 21. The narrative contained therein is clearly referable to the chronology above, and
is entirely accurate. The main point being that the absence of a test environment meant that there was simply no prospect
of System Integration Testing being able to start on 16 June 2016.
The Weekly Report demonstrates that shortly after this access to the Test Environment was provided but with significant
access issues. In particular, the Weekly Report for 16 May 2016 at JB273 p6075 and p6078 record:
p6075 "EI Project / Overall Summary…EI Test environment handed over but with restricted connectivity to allow
commencement of SIT to proceed 13/05. Critical connectivity issues remain WIP and need to be resolved for SIT"
p6078 Risks: "The remaining environmental build elements are required for EI SIT…Daily environmental review calls to
progress agreed solutions and planned delivered dates" Owner Stephen Waters
Although this allowed SIT to commence, it was very shortly thereafter suspended: see David Buchanan Supp. WS at
paras. 8.7.14 8.7.24.
On 20 May 2016, CGI issued its response re-addressing RN010 and addressing RN012. The notice simply stated ‘CGI agree
to this delay as being CGI Cause’, and envisaged working together to quantify the precise relief.
Page 218 ⇓
218
June - August 2016 On 7 June 2016, the Weekly Report records that the connectivity issues with the Environments had caused SIT to be
suspended:
“EI Project / Overall Summary…
EI environments (SIT, Pre-Production and Production) including completion of any subsystem connectivity remains
problematic. A revised infrastructure (and connectivity) plan has been issued that will not work. This has initiated
further re-planning to mitigate impact on EI go-live dates.
SIT has temporarily been suspended because of incomplete installation / connectivity of subsystems associated with the
SIT environment.”: Weekly Report for 7 June 2016 in JB276 at p6090
The issues with the Test Environment continued until around 27 July 2016, and SIT re-commenced with ‘smoke testing’
on around 2 August 2016: see David Buchanan WS at para. 8.7.17. The cumulative effect of the delay meant that SIT was
not completed until November 2016, which represented a delay of 5 months from the original planned completion date:
see David Buchanan’s WS at para. 8.7.21.
All of the witnesses confirmed that the remaining environments of PPE and Production were not eventually produced
until around mid-late August 2016: see Niall Sherriff’s WS at para. 4.85, and David Buchanan’s WS at para. 8.7.19
Page 219 ⇓
219
APPENDIX 4
1. A preliminary review of the submissions for the pursuer has identified the following
productions which do not appear to have been spoken to by a witness to fact:
Joint bundle number
459
676
402
356
450
617
577
575
619
712
1312
1378
1372
569
587
589
621-632
677
682
1037
1153
920
1524
1389
1616
247
249
252
450
503
520
249
545
649
700
740
Reference in submissions for the pursuer
§4.3.4
§6.2.2
§7.3.2
§8.10.5
Appendix A, p 1
Appendix A, p 4
Appendix A, p 4
Appendix A, p 4
Appendix A, p 5
Appendix A, p 6 (mistakenly identified as
702).
Appendix A, p 9
Appendix A, p 9
Appendix A, p 9
Appendix B, p 1
Appendix B, p 1
Appendix B, p 2
Appendix B, p 3
Appendix B, p 3
Appendix B, p 4
Appendix B, p 8
Appendix B, p 8 and p 9
Appendix B, p 10
Appendix B, p 12
Appendix B, p 16
Appendix B, p 16
Appendix C, p 1
Appendix C, p 2
Appendix C, p 3
Appendix C, p 6
Appendix C, p 7
Appendix C, p 7
Appendix C, p 8
Appendix C, p 8
Appendix C, p 8
Appendix C, p 9
Appendix C, p 9
Page 220 ⇓
220
261
276
2049
Appendix C, p 10
Appendix C, p 13
Appendix C, p 21
2. In addition, the following productions are referred to by Dr Hunt in her report without
having been spoken to by a factual witness:
a. JB 29;
b. JB 244;
c. JB 261;
d. JB 402;
e. JB430;
f. HB441;
g. JB443;
h. JB446
i. JB448
j. JB498;
k. JB499;
l. JB515;
m. JB550;
n. JB589;
o. JB676;
p. JB811;
q. JB886;
r. JB901;
s. JB903;
t. JB906;
u. JB920;
v. JB932;
w. JB1171;
x. JB1280;
y. JB1304;
z. JB1336;
Page 221 ⇓
221
aa. JB1381;
bb. JB1402
cc. JB1486;
dd. JB1485;
ee. JB1561;
ff. JB1596;
gg. JB1597;
hh. JB1687;
ii. JB1861;
jj. JB2015;
kk. JB2074;
ll. JB2111;
mm. JB2112
nn. JB2135;
oo. JB2156;
pp. JB2160;
qq. JB2185;
rr. JB2187-2192.
ss. JB2230
Page 222 ⇓
222
APPENDIX 5
1. A preliminary review of the submissions for the pursuer has identified the following
productions which do not appear to have been spoken to by a witness to fact:
Joint bundle number
459
676
402
356
450
617
577
575
619
Reference in submissions for Response for the pursuer
the pursuer
§4.3.4
§6.2.2
§7.3.2
§8.10.5
Appendix A, p 1
Appendix A, p 4
Appendix A, p 4
Appendix A, p 4
Appendix A, p 5
This reference has been
removed.
JB 676 was referred to in
the Expert Report of Dr
Hunt at para 9.3.4
(footnote 224).
JB 402 was referred to in
the Expert Report of Dr
Hunt at para 3.2.4
(footnote 14).
We will change the
reference to JB 651, which
is referred to in the WS of
Paul Neary at para 3.2
We will change the
reference to JB 449 at pg.
7985. Referred to in the
WS of David Buchanan at
para 5.1.3.
JB 618 and 617 (pg. 9745)
put to Kevin Bryce Day
12 page 139 of the
transcript.
JB 577 put to Peter Hall
Day 3 page 137 to 138 of
the transcript.
JB 575 (pp. 8850 8851) in
JB 577 (pp. 8859 8860),
which was spoken to by
Peter Hall as above.
This reference has been
removed.
Page 223 ⇓
712
1312
1378
1372
569
587
589
621-632
677
682
223
Appendix A, p 6 (mistakenly
identified as 702).
Appendix A, p 9
Appendix A, p 9
Appendix A, p 9
Appendix B, p 1
Appendix B, p 1
Appendix B, p 2
Appendix B, p 3
Appendix B, p 3
Appendix B, p 4
We will change the
reference to the WS of
Peter Hall at para 7.10 and
the evidence of Peter Hall
on Day 3, pages 143 and
144 of the transcript.
Reference should be to JB
1306, which was put to
Kevin Bryce Day 12
page 136 of the transcript.
We will change these
references to the WS of
Peter Hall at para 10.22, JB
1396 (referred to in that
paragraph), and the
evidence of Peter Hall on
Day 3 page 145 of the
transcript.
Email is also contained in
JB 1753. Referred to in the
WS of Ravinder Johal at
para 4.2.5
This reference has been
removed.
JB 589 is a cover email.
The reference is to JB 590,
which is document 46 in
the appendix to the
supplementary WS of
David Buchanan. It is also
referred to at para 3.2.2 of
the Joint Statement of
Experts.
Referred to in the WS of
Ravinder Johal at para
4.15.1
JB 677 is not cited as a
reference. In any event
the email is contained
within JB 683, which was
put to David Legge Day
13 page 35 of the
transcript.
We will change this
reference to the WS of
Niall Sheriffs at para 3.64
Page 224 ⇓
1037
1153
920
1524
1389
1616
247
249
252
450
224
Appendix B, p 8
Appendix B, p 8 and p 9
Appendix B, p 10
Appendix B, p 12
Appendix B, p 16
Appendix B, p 16
Appendix C, p 1
Appendix C, p 2
Appendix C, p 3
Appendix C, p 6
We will change this
reference to the WS of
Ravinder Johal at para
4.15.6 (d) and the evidence
of Stephen Waters Day
14 pages 30 31 of the
transcript.
On page 8 we will change
the reference to JB 1241,
which is referred to at
para 2.12.2 of the WS of
Stephen Waters.
On page 9 we will change
the reference to JB 1152,
which was put to Colin
Brown Day 17 page 78
of the transcript.
This reference has been
removed.
JB 1524 was put to Colin
Brown Day 17 page
127 of the transcript.
This reference has been
removed.
This reference has been
removed.
This production was part
of a section in the joint
bundle (EI Highlight
Reports and Weekly
Reports) that was put to
David Buchanan who
confirmed that he had
reviewed the items, that
they were correctly and
accurately described
Day 6 pages 6 11 of the
transcript.
As above see the
response to JB 247.
As above see the
response to JB 247.
The response in relation to
JB 450 is set out above.
Page 225 ⇓
503
520
249
545
649
700
740
261
276
2049
225
Appendix C, p 7
Appendix C, p 7
Appendix C, p 8
Appendix C, p 8
Appendix C, p 8
Appendix C, p 9
Appendix C, p 9
Appendix C, p 10
Appendix C, p 13
Appendix C, p 21
This reference has been
removed.
We will change the
reference to JB 521, which
is referred to in the WS of
David Buchanan at para
8.5.3.
As above see the
response to JB 247.
We will change the
reference to JB 543 at pg.
8573. This production is
referred to in the WS of
Niall Sheriffs at para 3.38.
We will change the
reference to JB 648 at pg.
9960. This production is
referred to in the WS of
David Buchanan at para
8.6.2.
We will change the
reference to JB 704 at pg.
10276. This production is
referred to in the WS of
David Buchanan at para
8.6.2.
This reference has been
removed.
As above see the
response to JB 247.
As above see the
response to JB 247.
We will change the
reference to JB 2028.
This production is
referred to in the WS of
Paul Neary at para 3.3.
We also refer to the oral
evidence of Paul Neary on
Day 13 pages 97 99 of
the transcript.
Page 226 ⇓
226
Part 2 of the Note
With reference to Part 2 of the Note, it appears to be correct that most (but not all) of the
productions listed in this part of the Note were only referred to by Dr Hunt. We understand
from a review we have undertaken this week, that it appears that the same position applies to
a material number of the documents referred to by Mr Coyne in his report, i.e. they were only
referred to by him. We do not see an issue with this as both experts are referring to
documents which they have found as a result of their investigations. The experts were
entitled to give evidence in this respect and the reference to the relevant documents by the
experts are sufficient for those documents to be in evidence. On this basis, we have not
included a document by document response to Part 2 of the Note.



BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: http://www.bailii.org/scot/cases/ScotCS/2018/[2018]_CSOH_112.html