From blueprint to backlog: Closing the gap between AI policy and public sector delivery
Shashin Shah
A year on from the launch of the AI Opportunities Action Plan, and with the Blueprint for a Modern Digital Government setting the direction of travel, the UK has rarely had a clearer policy mandate for digital transformation. The intent is unambiguous: faster public services, better outcomes for citizens, and smarter use of data and AI across every part of the state.
What is less clear is what happens next. In conversations with public sector leaders, the same question keeps coming up — not “should we be doing AI?”, but “where, exactly, do we start, and how do we ensure we are investing in the right things?”
That gap — between policy ambition and prioritised delivery — is where most digital transformation stalls.
The translation gap
Translating policy into practice in a public sector context is harder than it looks. Departments are sitting on hundreds of potential AI use cases — casework triage, fraud detection, citizen-facing services, internal automation, predictive maintenance of legacy estates, to name a few. The art of the possible is enormous. The art of the achievable, given regulatory constraints, technical debt, workforce capacity, and public-trust expectations, is far smaller.
Yet the way most organisations bridge that gap is still surprisingly informal. Use cases get pitched by enthusiastic teams or by suppliers with something to sell. Pilots are funded based on energy and influence rather than evidence. Governance and risk assessment arrive late — often after a vendor is already on the hook. The result is a portfolio of disconnected experiments, very few of which scale, and even fewer of which deliver the citizen outcomes the original policy promised.
What is missing is not appetite. It is a structured, repeatable way to go from “everything is possible” to “these are the right ten things to do next.”
Three things that have to be true
For policy to translate reliably into public sector delivery, three things have to be true.
First, discovery has to be systematic. Identifying AI opportunities should not depend on which directorate shouts loudest or which supplier got the meeting. The full surface area of a department’s operations, regulations, and strategic priorities needs to be examined, and use cases generated against that context — not against a generic playbook. A typical large public body will surface well over a hundred candidate use cases when this is done properly. That breadth matters: it forces a comparison, and comparison is where prioritisation begins.
Second, governance has to come first, not last. Every candidate use case should be assessed against compliance (UK GDPR, ICO guidance, sector-specific rules, AI guardrails), risk frameworks (operational, ethical, public trust), and technical readiness — before it ever gets near a business case. Doing this upstream is dramatically cheaper than discovering, eighteen months in, that a flagship pilot cannot go live because the data lineage was never sound or the procurement route was wrong.
Third, the output has to be a ranked portfolio, not a wish list. Senior decision-makers do not need another long list. They need a defensible, value-ranked, risk-adjusted shortlist with clear, quick wins, realistic effort estimates, and an honest read on which use cases need a platform investment and which can be delivered standalone. That is the document that unlocks funding, procurement, and delivery momentum.
What we are learning in practice
In our work with clients across regulated sectors — including transport, financial services, and increasingly public sector bodies — we have seen what happens when these three principles are applied with discipline. We have built our own accelerator, Onepoint Differential™, to do exactly this: rapidly discover, evaluate, and prioritise AI use cases against an organisation’s strategic, regulatory, and technical context. Differential routinely identifies more than 150 candidate use cases for a large organisation, then narrows that down to a governance-assessed, risk-adjusted shortlist with the financial impact and effort attached.
The accelerator is the easy part. The harder, more valuable work happens around it: structured workshops with operational leaders, rapid validation of priority use cases through our Onepoint Valuepath™ method, and crucially, vendor-independent recommendations. In a public sector context, that last point is non-negotiable. Departments need an honest read on whether to build, buy, partner, or wait — not a steer toward whichever vendor is footing the bill (some of it).
The pattern we see: prioritised portfolios delivered in eight to twelve weeks rather than nine to twelve months, with a clear line of sight from policy intent to delivery backlog.
What good looks like
A smarter state is not measured by the number of AI pilots it has running. It is measured by how reliably it turns strategic intent into outcomes that citizens actually experience — and by how honestly it manages the risks along the way.
That requires a few unfashionable things: pragmatism over hype, evidence over enthusiasm, governance designed in rather than bolted on, and a portfolio mindset rather than a parade of pilots. It also requires a willingness to say no to use cases that do not earn their place, however politically attractive they might be.
If you are wrestling with how to translate the AI Opportunities Action Plan, or your own departmental strategy, into a credible delivery backlog, come and find us at Building the Smarter State Week. We would be glad to compare notes.
How to discover, evaluate, and prioritise high-impact AI use cases tailored to your organisation is unpacked at onepointltd.com/differential.
techUK - Transforming Public Services
techUK members are transforming public services in the UK. Our community help to shape a smarter, digitally empowered public sector.
techUK drives public sector digital transformation by uniting the public sector and tech industry. Through early market engagement, efficient procurement, and innovative technology adoption, we help to modernise legacy IT, and enable efficient, secure, and personalised services.
Get involved: We run a busy calendar of activity including events, reports, and insights that demonstrate some of the most significant digital transformation opportunities for the sector. Our Transforming Public Services Hub is where you will find details of all upcoming activities. We also send a monthly public services newsletter to which you can subscribe here.
Upcoming 'Transforming Public Services' events
Latest news and insights
More resources
Authors
