Artificial intelligence has quickly moved onto the agenda for public sector leaders. And the potential is significant: reducing administrative workloads, making better use of organisational data, supporting employees with everyday tasks and, ultimately, helping public services operate more efficiently.
But there is a significant difference between experimenting with AI and embedding it successfully across an organisation.
The Government’s own findings demonstrate the challenge. At the time of the National Audit Office’s March 2024 report, just over a third (37%) of 87 government bodies surveyed reported that they had deployed AI. In contrast, 70% were piloting or planning its use. The Public Accounts Committee subsequently concluded that there remained little evidence of successful adoption at scale.
This reflects something we increasingly see when discussing AI adoption with organisations at Akita: The question is shifting from “What could we do with AI?” to “How do we introduce AI in a way that delivers practical value, remains secure and can actually scale?”
For public sector leaders, answering that question requires looking beyond AI technology itself.
Start With The Problem, Not The AI
The accessibility of generative AI has made it relatively easy for organisations to start experimenting.
But experimentation shouldn’t be confused with an AI strategy.
A successful AI rollout must start with addressing an organisation’s operational challenges and objectives, rather than a particular AI product.
Where are teams losing time to repetitive administration? Which processes rely heavily on manual input? Where is valuable organisational knowledge difficult to access? Which activities could potentially be streamlined through automation? And where could better access to information support faster decision-making?
From there, organisations can start identifying where AI could genuinely make a difference.
This principle sits at the heart of Akita’s approach to AI consultancy. We help organisations identify potential use cases and consider their value, feasibility and associated risk before moving towards deployment.
It helps prevent a common problem: investing in AI simply because the technology is available.
For public sector leaders, the ambition should not be more AI. It should be finding the right applications of AI to deliver measurable organisational and public value.
Make Sure Your Organisation Is Ready For AI
Not every barrier to AI is actually an AI problem.
Often, it is the technology, data and processes already sitting underneath it that determine whether an organisation is ready to proceed.
The Public Accounts Committee’s Use of AI in Government report highlights the scale of this issue across government. DSIT estimated that 28% of central government systems met its definition of legacy systems in 2024. The report also found that 21 of the 72 highest-risk legacy digital systems prioritised through the government’s 2022-25 roadmap still lacked remediation funding.
Data presents another challenge. 62% of the 87 government bodies responding to the NAO survey identified access to good-quality data as a barrier to implementing AI.
This matters because AI doesn’t operate separately from your existing digital environment.
Information quality, storage, permissions, governance, security and existing platforms all influence what’s possible and how safely AI can be introduced.
Akita therefore takes a readiness-first approach to AI adoption. Our consultancy considers business objectives alongside data readiness, information governance, security, user adoption and appropriate use cases.
For Microsoft 365 Copilot specifically, Akita’s readiness assessment examines the Microsoft 365 environment, data landscape and governance posture before creating a prioritised action plan and roadmap for responsible deployment.
The important point is simple: don’t discover problems with your AI foundations after rollout has started.
Treat Governance As An Enabler, Not A Blocker
Public sector AI inevitably raises questions around sensitive information, accountability, appropriate use and public trust.
The government recognises this challenge. The Public Accounts Committee reported that 67% of government bodies responding to the NAO survey regarded support from central government in fostering public trust in AI as very important. Its report specifically highlights fairness, accuracy, security, safety and transparency as important to successful public-sector adoption.
For leaders, however, this shouldn’t translate into avoiding AI because the governance questions appear difficult.
Good governance should make responsible adoption easier.
An organisation needs clarity over which AI technologies employees can use, what information can be processed, who has access to that information, where human oversight is required and who remains accountable for decisions.
Akita’s AI consultancy and readiness approach looks at these questions before deployment. For Microsoft Copilot environments, for example, this includes reviewing areas that could affect security, suitability and deployment success, with Akita bringing together expertise across Microsoft 365, cyber security and data governance.
That creates guardrails within which people can explore productive uses of AI with greater confidence.
Don’t Overlook Your Data And Permissions
For organisations considering Microsoft Copilot, data permissions deserve particular attention.
Microsoft 365 environments can contain years of accumulated content across SharePoint, Teams and OneDrive. Akita’s guidance on Copilot rollout therefore recommends reviewing permissions and access before wider deployment, alongside retention, classification and data loss prevention controls.
This is where an AI initiative can expose broader information-management issues that may previously have gone unnoticed.
If employees have access to information they no longer require, for example, introducing AI doesn’t solve the underlying permissions problem. It makes addressing that existing problem an important part of getting the environment ready for AI.
This is why Akita’s approach combines AI expertise with wider Microsoft 365, cyber security and data governance capabilities. AI readiness is rarely an isolated technology exercise.
Build Skills And Confidence Around AI
Technology readiness is only part of the equation. Public sector organisations also need people who understand where AI can help, where its limitations sit and how it should be used responsibly.
There is evidence that skills represent a substantial obstacle.
The Public Accounts Committee reported that 70% of government bodies responding to the NAO survey identified difficulties recruiting and retaining staff with AI skills as a barrier to adoption. It also reported that around 50% of roles advertised through Civil Service digital and data campaigns in 2024 went unfilled.
However, successful adoption doesn’t mean turning everybody into an AI specialist. It means developing sufficient understanding at different levels of the organisation.
Leadership teams need the confidence to evaluate investment and determine priorities. Managers need to understand where AI could improve their processes. Employees need practical guidance to use approved AI tools safely and productively.
That’s why user readiness should be considered alongside technical readiness. Akita’s existing AI proposition takes this broader view, considering employee readiness alongside data quality, governance, security, technology platforms and organisational processes.
Ultimately, giving people access to AI is relatively straightforward. Embedding AI into how an organisation works requires adoption.
Pilot AI For The Public Sector – With A Plan To Scale
There is nothing wrong with starting small.
In fact, phased adoption can allow organisations to evaluate an AI use case, discover issues and measure its value before committing to wider implementation.
The problem arises when pilot activity becomes the strategy.
The Public Accounts Committee found that government had no systematic mechanism for bringing together learning from AI pilots and warned of the potential duplication of effort and cost from siloed pilot activity. It recommended identifying common, scalable products and supporting their wider rollout where appropriate.
For individual public sector organisations, there is an important lesson here: A pilot should have a route forwards.
At Akita, our AI readiness approach considers practical use cases and their potential value, feasibility and risk before helping shape a controlled pilot and, where appropriate, a longer-term roadmap for adoption, governance and optimisation.
That means establishing what success looks like before deployment begins.
If the aim is to reduce administrative workloads, establish how that will be assessed. If the objective is faster access to knowledge, establish the current challenges and intended improvement. If an AI-supported process is being introduced, determine how its quality and effectiveness will be reviewed.
Not every pilot needs to progress.
Finding that a proposed use case isn’t sufficiently valuable or practical is itself a useful outcome. The purpose of piloting should be to create evidence for the next decision, not simply demonstrate that an AI tool works.
Measure Outcomes Rather Than AI Usage
This distinction becomes particularly important as adoption grows.
Licence numbers, logins and user activity can indicate whether employees are engaging with AI. They don’t necessarily demonstrate that the organisation is benefiting from it.
A better question is:
What has improved because we introduced AI?
Depending on the use case, this could mean looking at administrative time saved, processes streamlined, response times improved, employee capacity released or the quality and consistency of an activity improving.
The measures will differ between organisations and individual use cases. What matters is agreeing the required outcome beforehand.
This is also why Akita’s AI consultancy approach includes creating phased adoption roadmaps and measuring ROI and business outcomes, rather than treating technology deployment as the end of the project.
For public sector organisations, the same principle applies with an additional consideration: technology investment ultimately needs to support the delivery of organisational or public value.
Moving From AI Ambition To Practical Adoption
The challenge facing public sector leaders isn’t a lack of AI possibilities.
If anything, the abundance of possibilities can make determining where to start more difficult.
Government’s own assessment of public sector AI adoption points towards the areas that need attention: legacy technology, data quality, transparency, skills, learning from pilots and leadership. The Public Accounts Committee concluded that realising the benefits of AI would require clear accountability, delivery milestones and metrics to track progress and spending.
Ready To Move AI Beyond The Pilot Stage?
Turn AI ambition into a practical plan. Akita can help you assess your organisation’s AI readiness, identify high-value use cases and understand the data, security and governance considerations that could affect rollout.
Contact Us
