Getting Out of Our Own Way: Reflections on AI at the Government Transformation Summit 2025

On June 25, Valtech had the privilege of returning to the Government Transformation Summit – a space designed not for posturing, but for progress. The event brings together digital, data and technology (DDaT) and operational delivery leaders from across the UK government to step away from the day-to-day and engage in focused, frank conversations with peers who are grappling with many of the same challenges.
This year, we hosted three roundtable discussions on a topic that’s impossible to ignore: AI Use Cases.
We co-hosted our discussions alongside two members of the civil service – Adam Bricknell, Head of Advanced Analytics at HM Treasury, and Euan Slack, Director of Central Digital Platform at the Cabinet Office – who helped lead the discussions and ground them in the reality of public sector delivery today.
No Shortage of Ideas – But What’s Holding Us Back?
The starting point for each conversation was simple: where is AI already making a difference, and what’s getting in the way of doing more of it?
Unsurprisingly, there was no shortage of ideas or potential applications. We heard about powerful, human-centred use cases: from identifying citizens with specific vulnerabilities so they can be proactively fast-tracked for support, to triaging requests more efficiently to better meet public need. In each case, AI isn’t replacing people – it’s being used thoughtfully to direct finite human time and energy to where it’s needed most.
What was striking, however, was that the barriers weren’t technical. Repeatedly, the sentiment around the table was that the biggest obstacle isn’t the technology – it’s the system around it.
Culture. Leadership. Budgets. Governance. The real blockers are all-too-familiar to anyone working in or with the public sector. AI, with all its complexity and potential, simply amplifies them.
“We’re Getting in Our Own Way”
One speaker captured the mood perfectly: “We need to get out of our own way.”
This theme – of internal blockers outweighing external ones – came up time and again. From risk-averse cultures that make experimentation feel dangerous, to siloed ways of working that prevent the cross-pollination of ideas, the challenge isn’t imagining bold AI use cases – it’s creating the conditions in which they can take root and scale.
The summit attendees were definitely not AI sceptics, they were believers – albeit somewhat frustrated ones. Frustrated not by a lack of capability, but by how hard it still is to connect dots between pockets of innovation across government. And frustrated by how difficult it can be to move from a promising pilot to a scalable service.
Getting better at sharing, supported by the right tools and systems rather than red tape and restrictions
One of the most energising conversations of the day centred around the power of sharing: sharing code, sharing data, sharing lessons from use cases that have already been delivered (or have failed, usefully).
There’s a growing sense that this is where real progress can be made. It’s not about standing up a new AI centre or creating a shiny new platform. It’s about removing duplication, speeding up delivery, and making it easier for civil servants to build on each other’s work – not reinvent it.
While many civil service teams are already building brilliant things, we heard a clear call for mechanisms that make it easier to find and reuse what already exists. People want fewer roadblocks, more reuse, and a greater sense of shared purpose.
What Comes Next?
At Valtech, we left Government Transformation Summit energised – not just by the ideas, but by the honesty in the room. These weren’t surface-level conversations. They were deep, practical and – above all – real.
The public sector doesn’t have an AI idea problem. It has a delivery problem. And solving that requires more than investment in tools or skills – it requires investment in ways of working that allow ideas to scale. That means permission to experiment, clearer governance paths, and cross-departmental collaboration by default, not exception.
We’re honoured to be part of the conversation – and privileged to work with public sector teams who are quietly doing the hard work of transforming services every day. For example, we are incredibly proud of our work to deliver AI-powered search on GOV.UK, improving citizen access to essential government information and services. This AI Use Case required our teams to bring to the civil service a new way of thinking – cutting through red tape, and bringing teams and leaders on a journey, whilst delivering a 14% increase in users clicking through to relevant results, an 11% decrease in search refinement, and an 8% reduction in users leaving the site from the search page.
Through our time at the Government Transformation Summit, and our experience delivering AI Use Cases directly in the public sector, we have learnt that the more we create space to learn from each other, the faster we’ll get out of our own way – and realise the full potential of AI to improve outcomes for the people who rely on public services most.
Karl Hampson, Chief Technology Officer, Data & AI, Valtech co-wrote this article.

By Bella Copland
Bella Copland is Principal Strategist, Valtech.Also Read
- Getting Out of Our Own Way: Reflections on AI at the Government Transformation Summit 2025
- DSIT appoints Emran Mian as new permanent secretary
- Government Transformation Summit Insights, Day 2: Data and AI
- "I'd like to see this summit in every country," South African civic tech CEO speaks at Government Transformation Summit