The UK government’s partnership with OpenAI has brought a familiar public sector challenge into sharper focus. As public bodies explore how AI can improve productivity and support service delivery, attention is increasingly turning to what it takes to implement these tools effectively in practice. Adopting new technology is one thing. Turning it into meaningful organisational change is far harder, particularly in public services shaped by years of operational pressure and workarounds.
For public bodies, the real barrier is rarely the technology itself. More often, it is whether the organisation is ready to absorb change and use new tools in ways that genuinely improve outcomes. AI ambition may be high, but ambition alone does not change how decisions are made, how services are delivered or the way people experience public services.
That gap often becomes visible in implementation. When leaders are under pressure to demonstrate progress, that’s when pilots launch, systems go live, and new tools are rolled out quickly. But if processes stick to the old ways of working, staff will remain unclear on what is changing and what success looks like, meaning momentum can fade fast.
Rollout does not guarantee progress
Public services need new ways to improve productivity and make better use of capacity. Recent government AI pilots have shown why the opportunity is attracting attention, with one consultation analysis tool estimated to save officials as much as 75,000 days of work each year if used across around 500 annual consultations. While the scale of that potential is significant, the long-term value depends on whether organisations are able to use these tools effectively.
That potential is significant, but AI adoption only delivers value when the surrounding organisation is ready for it. Introduced into unclear processes or weak governance, it may simply move existing problems faster. The result can be activity without impact, where rollout is mistaken for progress.
One area where this challenge becomes particularly visible is data quality. Most organisations, including those in the public sector, already know their data is not where it needs to be. They have been able to live with it, work around it and, in some cases, avoid the harder questions because people understand the gaps. Teams know which records need checking, where local knowledge fills the gaps and when reports require interpretation.
AI changes the consequence of leaving those issues unresolved because it can act on information at scale. Poor data does not simply reduce performance. It can accelerate the wrong outcome. In public services, where decisions shape budgets, service access and people’s lives, that risk cannot be treated as a technical inconvenience.
The people closest to the data are often furthest from the decision making. Someone entering or maintaining information may never see how it later informs a service decision or funding case. When that connection is invisible, improving data quality can feel like administration rather than shared ownership of better outcomes. Leaders need to make that link clearer.
Adoption must be built in early
Readiness should be treated as a central part of transformation from the start, not something to address once deployment is already under way. Leaders need to ask whether the organisation understands the quality of its data, if processes are still fit for purpose and whether staff have the support to work differently.
A survey of 938 UK public service professionals found that 45% were aware of generative AI being used in their area of work, while only 32% felt there was clear guidance in place. That gap shows how quickly informal adoption can move ahead of the structures needed to use AI safely and confidently.
Guidance must be practical enough to apply within day-to-day work, helping people to understand how AI should support the services they deliver and where human oversight is still needed. Basic training alone is unlikely to be enough, particularly when organisations are asking staff to adapt established ways of working while also building confidence in new tools and processes.
Leadership alignment is just as important. Decisions about what data matters, how it is defined and who owns it end to end are key. As a result, AI is only increasing the need for clarity from strategy through to delivery.
Many initiatives stall because adoption is treated as an afterthought. Staff are brought in once decisions have already been made, then expected to adapt at speed. Resistance in that context is often a sign that the organisation has moved too quickly into implementation without creating the conditions for people to develop with it.
For public sector organisations, the opportunity remains significant. AI can help teams make better use of information and reduce avoidable pressure on services. Yet technology cannot resolve unclear ownership, fragmented processes or a lack of confidence in the way change is being led.
Lasting value will come from organisations willing to confront the realities of implementation before it gathers pace. That means understanding whether their foundations are strong enough to support change in practice, not just in principle.
If those foundations are not in place, AI ambition will not be enough. Organisational readiness will determine whether projects deliver sustainable value or simply create the appearance of progress without the deeper change public services need. For organisations under pressure to modernise quickly, readiness may ultimately prove the difference between isolated AI experiments and lasting public service transformation.
