Case Study: Unraveling the Software Disaster of the Scottish Police’s i6

Section 2 Group 3

Muhammad Ali

Avneet Grewal

Shaheer Khan

Ashmit Sharma

Lucas Sudol
Table of Contents
Introduction 3
Project Analysis 3
I. Project Budget 3
II. Schedule of Events 4
III. Contingencies 4
IV. Performance 5
V. Client acceptance and satisfaction 5
VI. Project Analysis Conclusions 6
Project Timeline 6
Conclusions and Lessons Learned 7
Team Contribution Table 7
References 7
Appendix A: 8
ChatGPT Evaluation Appendix 9
Introduction
The i6 software project, launched in 2013 by the Scottish Police Authority (SPA) and Police Scotland, contracted
Accenture for £46.11 million to develop a comprehensive computer system. Initially promising with the aim to replace
130 existing systems and generate £200 million in savings, the project ended in 2016 with wasted resources, time, and a
setback for Scotland's police forces.

Project Basics:
● Purpose of the Software System: The i6 system aimed to modernize and streamline police operations, covering
80% of processes related to recording crime and missing persons. It intended to provide a centralized platform to
replace outdated electronic and paper-based systems.
● Replacement of Previous System: The i6 project was designed to replace 130 existing electronic and
paper-based systems, representing 80% of police processes related to recording crime and missing persons.
● System Creator: Professional services company Accenture was awarded the £46.11 million contract to create the
i6 computer system for Police Scotland.
● Client: The Scottish Police Authority (SPA) and Police Scotland were the clients overseeing the i6 project,
responsible for ensuring its successful implementation.
● Intended Users: The intended users of the i6 system included a broad spectrum within the police force, covering
frontline officers, investigators, and administrative staff. The system was intended to support real-time data
access, case analysis, and overall operational efficiency.

Main Point Summary:

The i6 software project, initiated with high expectations and a goal to revolutionize police operations, ended in
2016 as a costly failure. Created by Accenture for the Scottish Police Authority and Police Scotland, the project aimed to
replace numerous outdated systems and generate significant savings. However, issues such as disagreements over
functionality, delays, and a loss of trust between the involved parties led to the termination of the contract in 2016. The
aftermath involved a financial settlement of £24.65 million, reflecting the refund of the amount paid by the SPA and
additional compensation. The case underscores the challenges of large-scale IT projects, emphasizing the importance of
clear communication, realistic planning, and adaptability in ensuring successful outcomes.
Generated By ChatGPT
Project Analysis

I. Project Budget
Key financial details of the project:
● The original budget for the i6 software project was £46.11 million. This contract was price fixed and awarded in
June 2013 to the company Accenture.
● Even with the fixed price contract, changes to the original budgets occurred as a result of disagreement and
difficulties within the project. This resulted in a contract variation agreement being signed.
● This resulted in a revised delivery and requirement schedule. This change in agreement also signaled a breakdown
of trust between the parties.
● As a result of the changes within the contract variation agreement, additional cost and delays were incurred.

Post Release Costs:

● Fundamental flaws and serious errors were discovered after the software was given to the Scottish Police for
testing in August of 2015.
● Accenture re-estimate that completing the requirements of the contract would take an additional two and a half
years
 ● The SPA and Accenture mutually agreed to terminate the project contract in 2016. This resulted in Accenture
refunding the amount the Scottish Police have paid for the project, which at the time was £11.09 million. They
also agreed to pay an additional amount of £13.56 million in order to cover staff, maintenance, and software
licensing fees.

Key sources of cost changes and overruns:

Negotiations with Disagreements
● Disagreements early on led to a breakdown of relationship and trust between the parties. This led to
further disagreements upon requirement negotiations and the overall interpretation of the system.
Underestimating Project Scope and Complexity
● The i6 was originally planned to be based off an existing system Accenture had delivered. As
development proceeded, this proved impossible as the system's complexity was underestimated by
developers.
● This lack of understanding of project scope and complexity resulted in additional costs, as the software
system had to be built from scratch.
Development Methodology
● This project adopted the “waterfall methodology”. This made the change of requirements extremely
difficult within the development cycle. This amplified the original underestimation of scope and
complexity, as it made it nearly impossible to reevaluate requirements in order to proceed with the
project.
Lack of Proper Assurance
● Despite their underestimation and problems faced in systems development, Accenture frequently provided
assurance to the SPA about the successful progress within their software system. This lack of
communication made it extremely difficult for the police academy to gauge their own project
requirements, as there was a lack of transparency with the project status.
Template Generated By ChatGPT
II. Schedule of Events
Refer to Appendix B.

Outcome: The program was terminated in July 2016, resulting in a settlement agreement between the SPA and Accenture.
Accenture refunded £11.09 million and made an additional payment of £13.56 million. The project's collapse was
attributed to a damaging loss of trust and fundamental disagreements about project deliverables.
Generated By ChatGPT
III. Contingencies
Contingencies and plans for ensuring a successful product release are integral to any software project. In the case
of the i6, many contingencies were set forth by all parties.

The i6 program team demonstrated a proactive approach to addressing expertise gaps by appointing external
specialists to ensure comprehensive coverage. Notably, Deloitte was enlisted as external experts on procurement and
managing commercial contracts, emphasizing the commitment to robust procurement practices. The inclusion of
Eversheds as legal advisors contributed to legal diligence, ensuring that the contractual framework remained sound.
Furthermore, Exception UK was appointed as technical advisors, bringing technical expertise into the fold. These strategic
appointments reflected a concerted effort to bolster the program team's capabilities and navigate the intricate landscape of
IT system development and implementation.
Above Paragraph Generated By ChatGPT
The final “contract was robust” [3] which in turn acted as a safeguard that allowed for ease of termination for the
SPA down the line. After Accenture continuously faced difficulties in submitting deliverables through stages 2 & 3, the
program board requested Accenture “to provide future program board meetings with a quantitative assessment of its
confidence in delivering i6”. [3] In response to identified gaps and disagreements, a Contract Variation Agreement was
initiated in March 2014, demonstrating a commitment to addressing issues and aligning the IT system with specified
requirements. The agreement aimed to "reset the relationship between organizations and, temporarily, improved levels of
trust." [3] Additionally, the program board opted for a flexible payment schedule in 2014, deviating from the original
contractual terms. The report notes, "While this could be considered pragmatic, given the multiple challenges and
problems with the i6’s development, this is not good practice." [3]

Despite political pressures and collaboration challenges, there were ongoing efforts between the SPA, Police
Scotland, and Accenture to navigate complexities and deliver on project expectations. However, these efforts ultimately
faced significant hurdles, leading to the mutual agreement to terminate the contract in 2016. An external review in 2017
by Audit Scotland identified a loss of trust and fundamental disagreements as key contributors to the project's collapse,
emphasizing the importance of effective contingencies in large-scale IT endeavors.
Above Paragraph Generated By ChatGPT
IV. Performance
The i6 Program encountered several critical failures during its development. This prompted an urgent
reassessment by the SPA and Police Scotland, emphasizing the need to determine the next steps in developing and
delivering a much-needed police IT system. The I6 Program's failures left police officers and staff grappling with
out-of-date, inefficient, and poorly integrated systems. Specific examples of errors, such as those in the search and audit
modules, impeded data retrieval and system monitoring. Limitations in the administration module further hindered
effective system management. The consequences of these performance issues shaped an ongoing dialogue about the
strategy and implementation of police information technology in Scotland. Accenture's misjudgment of the system's
complexity and resource requirements contributed to project delays and operational inefficiencies. This misjudgment had
cascading effects on resource allocation and project management, ultimately causing the system to fall short of the
contractual standards. These problems were found late because of the way the project was managed, which was using the
waterfall method. This linear method of development led to a lot of time and money being wasted for bug fixing and
revision highlighting the need for a more continuous testing and evaluation approach during the development of such
complex systems.

The testing for the i6 system involved multiple phases. Initially, there was a product testing phase conducted by
Accenture, which found various technical problems. Following this, the system was handed over to Police Scotland for
User Acceptance Testing (UAT). During UAT, Police Scotland discovered critical errors, unresolved defects from
Accenture's product testing phase, and other major issues such as non-compliance with criminal justice data standards and
problems in search, audit, and administration modules. The testing strategy employed by Accenture was challenged by
Police Scotland, indicating a possible misalignment in testing standards and methodologies between the two entities. The
discovery of these errors and issues during the testing phases brought to light the serious shortcomings in the system's
performance and functionality, contributing to the eventual failure of the project.
Generated By ChatGPT
V. Client acceptance and satisfaction
Miscommunication and Misalignment in Requirements:
The Scottish i6 program case underscores the consequences of miscommunication and misalignment in project
requirements. The Scottish Police were dissatisfied with the delivered product due to misunderstandings regarding the
project's scope and requirements interpretation.

Inadequate Requirement Gathering and Documentation:

The case exemplifies the failure to establish clear, unambiguous, and well-documented requirements at the
project's outset. The Scottish Police believed their requirements were more extensive than what Accenture had planned to
deliver. This points to the critical software engineering best practice of comprehensive and detailed requirement gathering
and validation, which was not adequately followed.
Deviation from Agile Principles:
When it became apparent that the delivered product did not align with the Scottish Police's expectations,
Accenture informed them that adding the desired features would extend the project timeline by more than two years. This
situation deviated from the agile software development principle of responding to change over following a plan. It
highlights the importance of adapting to evolving client needs and avoiding rigid adherence to the original plan.

Lack of Early and Continuous Client Collaboration:

Effective client involvement during the development process can help identify issues and discrepancies early on,
making it easier to address them and align the project with the client's expectations. Despite initial efforts to split
requirements into various work examples and inform design decisions, there were still disagreements and
misunderstandings. This reveals a deficiency in early and continuous client collaboration, a critical software engineering
best practice.

In summary, the Scottish i6 program case study emphasizes the significance of comprehensive requirement gathering,
effective client collaboration, and flexibility in adapting to changing client needs. Neglecting these software engineering
best practices can lead to costly project terminations and client dissatisfaction, as demonstrated by this incident.
Generated By ChatGPT
VI. Project Analysis Conclusions
The i6 system failure presents a multifaceted challenge for Accenture, starting with substantial financial setbacks.
The company is obligated to refund £11.09 million and make an additional payment of £13.56 million as part of the
settlement agreement, extending the impact beyond the initial project budget. In addition to financial ramifications, the
erosion of trust with Police Scotland raises concerns about Accenture's future client engagements, particularly in the
public sector. The fallout from this high-profile project failure may induce hesitancy among prospective clients, affecting
the company's ability to secure contracts. Internally, the failure prompts Accenture to reassess its operational strategies,
including project management methodologies and quality assurance processes. Operational changes become imperative to
prevent recurrence and demonstrate a commitment to improvement. Navigating these challenges requires a comprehensive
recovery plan that addresses financial, contractual, and reputational aspects, emphasizing Accenture's dedication to
overcoming the aftermath of the i6 project failure.
Generated By ChatGPT
Project Timeline

For increased clarity, visit here or refer to Appendix A

Conclusions and Lessons Learned
The i6 project for Police Scotland exemplifies key lessons in large-scale software project management. Essential
points include the need for accurate budgeting, realistic timelines, and clear, agreed-upon requirements to prevent
financial overruns, delays, and scope creep. The late discovery of critical errors underscores the importance of early and
thorough testing, potentially using Agile methodologies for prompt feedback. Effective communication and trust are vital
for a positive client-vendor relationship, as are the right expertise and a realistic assessment of project complexity.
Involving users in testing ensures the system meets real-world needs. A robust contingency plan and risk mitigation
strategies are crucial, as is a supportive environment that fosters transparent reporting and constructive issue handling,
avoiding optimistic biases and high-pressure pitfalls evident in the i6 project.
Generated By ChatGPT

In conclusion performance, scheduling, budgeting, and client satisfaction are all important categories in making a
successful software project. These different areas are highly interdependent because messing up in a single area such as
scheduling can have effect on how the team works this is because if a mistake in scheduling is made it will affect the way
the code is written because enough time might not be allocated for such a task which can cause a rush which occurred in
this project as well. Therefore, having a strong foundation and methodology can help in staying on task while also
retaining quality.

Team Contribution Table

Name Contributions

Ashmit ● Performance Analysis - Specifically the performance and impact of the software
● Project Timeline Visual

Avneet ● Schedule of Events

● Contingencies
● Project Timeline Visual

Lucas ● Introduction
● Budget Analysis
● Project Analysis Conclusion

Muhammad ● Performance analysis - specifically the testing methods and the errors that were found
before the product was released
● Conclusion and lessons learnt

Shaheer ● Client Acceptance and Satisfaction

● Reformatting

References
[1] R. Alderson, “‘lost’ £200m in savings after police IT scheme collapses,” BBC News,
https://www.bbc.com/news/uk-scotland-39208096 (accessed Nov. 13, 2023).

[2] A. Scotland, “I6: A review - audit scotland,” Audit Scotland,

https://www.audit-scotland.gov.uk/uploads/docs/report/2017/nr_170309_i6_review.pdf (accessed Nov. 13, 2023).

[3] H. Dolfing, “Case study 15: How the scottish police got £25 million back but lost 3 years on i6,” Henrico
Dolfing - Interim Management and Project Recovery,
https://www.henricodolfing.com/2021/09/case-study-scottish-police-i6-accenture.html (accessed Nov. 13, 2023).
Appendix A:

Appendix B:

Schedule of Events
● 2010 - 2011: In 2010, the procurement process for the i6 computer system began. A competitive dialogue process
took place, and a dedicated program team developed the business case for i6. The business case aimed to cover
80% of core police services and generate efficiency savings of around £200 million over ten years.
● 2012: Ninety suppliers expressed interest in the procurement process. Accenture was selected as the preferred
bidder in November 2012.
● 2013: The SPA approved the final business case in June 2013, awarding a £46.11 million contract to Accenture.
The i6 program faced immediate difficulties, with a disagreement emerging over the search function.
● 2014: In March of 2014 the Justice Sub-Committee on Policing expressed frustration at the lack of information
about ongoing issues. In November a contract variation agreement was made, resetting the relationship and
temporarily improving trust.
● 2015: January 2015: The product testing phase began, four months behind the original schedule. June 2015: The
program board raised concerns about unresolved defects, and the relationship between Police Scotland and
Accenture became extremely fragile. August 2015: Critical errors and higher-than-projected flaws were identified
in the system.
● 2016: May 2016: The SPA decided that the revised plan was not viable. July 2016: The SPA and Accenture
mutually agreed to terminate the contract, signing a settlement agreement of £24.65 million.
ChatGPT Evaluation Appendix

Muhammad
- Accuracy: The content was aligned with the source provided, accurately summarizing the key events and
outcomes of the i6 project. There were no significant inaccuracies, omissions, or misinterpretations in the
information relayed. However sometimes when prompted with the same question multiple times it would give
different outputs sometimes even giving contradictory results.

- Completeness: The responses covered essential aspects of the disaster, including the project's initiation,
challenges, and eventual termination. However, for an exhaustive understanding, additional details from other
sources might be beneficial.

- Clarity and Coherence: The information was presented in a coherent and structured manner, ensuring ease of
understanding. The sequence of events and the project's challenges were laid out systematically.

- Insights and Analysis: While the summary provided a clear depiction of events, it mainly focused on
summarization. Deeper analysis or critical reflection on the project's implications and lessons learned was not
extensively covered.

- Consistency: The responses were consistent in summarizing the content from the provided source. They
maintained a balance in detailing the project's timeline and the issues encountered, though more in-depth analysis
varied.

Shaheer
- Accuracy: The content was fairly accurate as I provided it notes in advance and told it to only write about
information that it was provided.

- Completeness: The responses I received were complete based on my prompts, and included all details that were
specified.

- Clarity and Coherence: The outputs were quite clear, but were lengthy before tweaking the prompt to have it
write shorter sections.

- Insights and Analysis: The insights and analysis were largely my own, as my notes provided my own
conclusions based on the information, however it did correctly link several failures of the i6 to actual software
engineering practices.

- Consistency: The outputs were consistent with each other, but this is most likely due to the prompts and provided
information that I included with the first prompt.

Ashmit
- Accuracy: The content was relevant to the source that was provided and it accurately summarized the key events
and outcomes of the i6 project, however sometimes it tended to repeat a lot of information along with copying
specific lines from the article, which led to me double checking what I had to read. It also included information
that I did not want, which led me to editing my paragraphs several times.

- Completeness: ChatGPT was able to complete the prompts that I asked for. However it struggled when I asked it
be more specific and go deeper into the content, it struggled.
 - Clarity and Coherence: At times, ChatGPT was not really coherent as it was really specific and provided
information which would be hard for the user to read and understand. The outputs were also really long which
meant that we had to condense certain stuff that it outputted and make it shorter.

- Insights and Analysis: I believe ChatGPT did a mediocre job of analyzing the information provided because it
provided several inaccuracies. However, it was able to better analyze the information once I provided more
context along with my notes and website.

- Consistency: I asked ChatGPT to be as consistent as possible, however at times it did make several mistakes
which led me to double checking my work. I just had to ensure that the prompts that I gave it were accurate
enough.

Lucas
- Accuracy: Although the content generated through AI was very careful to stay accurate, it struggled to perform
research as this software disaster wasn't significantly popular within the media. This resulted in the need to
provide ChatGPT with a text article that it could use within its dataset. Given a well written article, the AI was
able to interpret the text and come to accurate conclusions.

- Completeness: As this disaster wasn't overly publicized, ChatGPT struggled to provide complete results without
being given information to use within its prompt. Instead, it referred to providing templates with whatever
knowledge it had in its database. Given the right prompts, including excerpts from accurate articles, the AI was
able to build off of the given information in order to build complete results.

- Clarity and Coherence: Responses for the most part were given clear and concise, there were occasions where
the AI model used an extreme degree of elevated vocabulary to describe concepts, which made the information
hard to follow.

- Insights and Analysis: Given accurate and consistent information, ChatGPT was able to extrapolate the
information with a high degree of success and provide useful insights into the data. It was able to provide accurate
conclusions and interpret the ideas of the given article.

- Consistency: When provided with an article or information excerpt, the AI was consistent with its prompts.
Otherwise, the prompts would have a wide degree of variability depending on what research the AI decides to use
within the dataset for prompt generation.

Avneet
- Accuracy: Content and research was provided but sometimes ChatGPT struggled with interpreting information in
relevant ways.

- Completeness: ChatGPT was able to interpret sources and take away key points but when asked to delve deeper
or provide more nuance, it struggled.

- Clarity and Coherence: ChatGPT when replying in bullet form, tended to be more direct with its information but
when asked to format points into paragraphs, it tended to add a lot of unnecessary fluff which sometimes led to
the main arguments being lost or hard to follow.
- Insights and Analysis: At times, ChatGPT struggled with providing the full extent of the information which
loops back to its struggles with accuracy of interpretation. By providing more centralized information, its ability
to interpret content increased.

- Consistency: ChatGPT did provide consistent information when prompted in different ways. It was better at
formatting in bullet points rather than paragraphs.