How Whitehall delivery units, AI and devolution work

If you have ever wondered why ministers announce big plans that seem to vanish into delay, the GOV.UK speech 'Remaking the State' is really about that problem. Strip away the conference setting and the argument is simple: Whitehall is often better at discussing change than carrying it through, and the state needs clearer priorities, faster decisions and more trust in the people doing the work. That matters well beyond Westminster. When government is slow, you feel it in late infrastructure, clunky public services, confusing websites and policies that sound impressive but never quite reach your street. **What this means:** this is less a story about one minister and more a lesson in how the machinery of government either helps delivery or gets in its way.

The speech says the system is weighed down by inertia: too many layers of sign-off, too many people invited to comment, and not enough people allowed to decide. It also argues that departments often chase too many priorities at once, which makes it harder to tell what success even looks like. There is a useful media-literacy point here. Speeches like this are not neutral descriptions; they are arguments. Even so, the diagnosis will sound familiar to anyone who has used a public service and wondered why a simple task takes so long. The criticism is not that civil servants do not work hard. It is that the structure around them can reward caution, paperwork and repeated checking over action.

One answer offered in the GOV.UK speech is to cut the number of top priorities and measure them more clearly. The minister says each department was given only two or three top-tier prime ministerial priorities, backed by shared data, agreed metrics and an integrated delivery dashboard so departments and the centre were looking at the same picture. That may sound dry, but it matters. If one team thinks a project is going well and another thinks it is slipping, arguments can drag on without anything changing. A shared dashboard is meant to force a harder conversation: is this genuinely on track, or has everyone simply become used to warning signs? In plain English, the reform tries to replace vague optimism with evidence.

The speech also makes a strong case for performance management inside the Senior Civil Service. According to the minister, some departments did not have key performance indicators in place when he arrived, and in some cases those measures had not been shaped clearly enough by ministers' priorities. The response was to insist on clearer KPIs, tougher bonus thresholds and performance-based pay progression across the Senior Civil Service. You do not have to agree with every part of that approach to see the logic. Public institutions need fairness, but they also need a way to tell the difference between routine administration and genuinely strong delivery. **What this means:** if government wants faster results, it is trying to make senior leaders more directly accountable for them.

Where the speech becomes most vivid is in its attack on what officials nicknamed 'sludge'. That is the pile-up of reviews, approvals and cross-Whitehall checks that can trap both major schemes and smaller projects. The minister points to projects such as HS2 as the kind of case that shows what happens when everyone has a say but nobody can act: costs rise, timescales stretch and public trust falls away. The proposed fix was structural. A new National Infrastructure and Service Transformation Authority was created in the Treasury, and Project RESET was launched to raise departments' delegated limits and cut low-risk approvals. In the speech, the minister says one commercial checkpoint that had reviewed around 120 cases a quarter later fell to zero. The promise here is clear: spend less time asking permission for minor decisions, and more time focusing expert scrutiny on the projects that really carry risk.

Technology is presented as the area where reform could move from theory to something you can actually see. The speech highlights the expansion of the No10 Innovation Fellowship, a scheme bringing highly skilled coders, data scientists and AI specialists into departments to build tools around real working practices rather than forcing staff to fit a rigid central template. Two examples do a lot of work in the speech. One fellow embedded at HMP Wandsworth and built AI tools that automated parts of prison officers' admin, freeing time for frontline work. Another created a childcare website so parents could check free-hours entitlement, estimate costs and compare nearby providers in one place. These are useful case studies because they show what good public technology should do: remove friction, save time and make the service clearer for the person using it.

Another big idea is the return of delivery units. The criticism the minister was answering is easy to understand: surely every department is supposed to deliver? In theory, yes. In practice, the speech argues that not every department has a dedicated team tracking progress, spotting blockages and pushing problems upwards before they turn into failure. The speech points to departments such as education and health, which already have stronger delivery machinery, as evidence that dedicated tracking can help. The model would not have to look identical everywhere. But the shared aim is to stop delivery from depending on whether a particular minister remembers to chase the same officials every single week. Re-establishing the No10 Delivery Unit is presented as part of that wider shift.

The speech then widens out to communications, and this is where the argument becomes as much about trust as management. The minister says government estimates suggest only around 2 per cent of the public engage with official communications, while Whitehall still spreads its messages across a confusing patchwork of brands, departments and campaigns. The response has been to build a new Government Media Unit at the centre, with the job of helping departments tell clearer stories on the platforms people actually use. There is a fair challenge here, though. Better storytelling can help people understand services, but trust is not built by branding alone. If delivery is weak, the slickest video in the world will not fix it. The speech is strongest when it treats communication as support for real change, not a substitute for it.

The future-looking part of the speech begins with skills. A new National School of Government and Public Services was presented as a way to train civil servants more consistently, including AI training across Whitehall, while also offering structured support to ministers who often enter office with huge responsibility and very little formal preparation. That last point is more important than it sounds. We usually expect teachers, nurses, engineers and managers to train for their roles. Ministers make decisions affecting millions of people, yet are often expected to learn entirely on the job. **What this means:** the speech is arguing that a better state is not only about new rules but about better-prepared people.

The final lesson is about devolution. The speech ties weak growth and squeezed living standards to a wider economic problem that has dragged on since the 2008 banking crisis: Britain has struggled to grow strongly enough, and too much decision-making still sits in Westminster. If cities and regions are supposed to drive jobs, skills and investment, then local leaders need real authority, not a system where they still have to filter major decisions through Whitehall. This is where the phrase 'devolution must mean devolution, not duplication' hits home. Passing powers outwards while keeping the same checks, headcount and habits at the centre does not produce freedom; it produces two layers of bureaucracy instead of one. For students of politics, that is the clearest takeaway from the GOV.UK speech. Remaking the state is really about three questions: who gets to decide, how progress is measured, and whether power is trusted to sit closer to the people it affects.