DigiGov 2026: AI ambition is growing. Are the foundations ready?

After two days at DigiGov 2026, I left with the sense that the conversation about artificial intelligence in government has moved on. The question is no longer whether it has a role in the public sector, but how public bodies can use it safely and show that it actually improves outcomes. 

The risks around security, privacy, transparency and accountability haven’t gone away. But neither uncontrolled adoption nor organisational paralysis is sustainable. 

Across the event sessions and conversations on the TXP stand, the same point kept surfacing: AI ambition is growing faster than some of the services, processes, data and systems needed to support it. Five conversations stood out.

1. The AI conversation is moving from pilots to outcomes

Government has largely moved past the question of whether AI has a role to play. The focus for them is now on identifying where AI can deliver value and what it takes to turn pilots into change. 

Many public bodies have shown that they can develop proofs of concept. The harder task is demonstrating value and establishing a credible path to adoption. That means linking each initiative to a defined service problem, workflow and meaningful measures: reduced effort, improved timeliness or quality, released capacity, stronger decisions or a better user experience. 

This matters when funding is constrained. Leaders cannot justify emerging technology on novelty alone; they need a clear outcome, the cost and risk of the current position, and measurable benefits. For a regulator or arm’s-length body, AI might support triage, evidence review, case handling or regulatory insight. Its value, however, depends on reliable data, auditable decisions, clear accountability and professional judgement at the right points. 

A better place to start is with the problem and not the technology itself. What services need improving, who is this effecting and how will we know if we’ve made a difference?

2. People experience services, not organisational boundaries 

One point that repeatedly came up during DigiGov was that service users don’t think about departments or organisational structures, they simply want their issues resolved effectively.  

Someone applying for support, renewing a licence, submitting regulatory information or accessing a public service should not need to understand which team owns the data, where one system ends and another begins, or why information must be supplied repeatedly. 

Many services cross boundaries users cannot see. Staff reconcile records, move information manually, interpret inconsistent rules and bridge system gaps, creating backlog, rework and cost internally, and delay, repetition and lost confidence externally. 

The UK Government’s Blueprint for Modern Digital Government calls for services that cross institutional boundaries. Its 2026 roadmap develops that ambition through joined-up services, responsible AI, stronger infrastructure and talent, outcome-based funding and greater accountability. 

Leaders should therefore judge transformation from the perspective of the citizen, service user, business or regulated community, not the convenience of the existing structure. The most successful programmes focus on making services easier to use, with technology supporting the outcome rather than driving it.

3. Workarounds reveal where services are under pressure

A recurring theme at the event was the gap between documented processes and the reality of how services operate day to day. Teams maintain local spreadsheets, transfer information between systems and rely on knowledge held by a small number of experienced colleagues. These workarounds may keep the service moving, but they can also hide failure points, weaken auditability and create dependencies that only become visible under pressure. 

That is why legacy modernisation should not begin with a list of old systems. Start with the services that matter most, then trace the processes, people, data, technology, suppliers and workarounds they depend on. 

This gives leaders a better basis for investment decisions. A newer system may still present serious operational risk, while an older platform may be stable, supported and appropriately controlled. Priority should follow service impact, the likelihood of failure and the ability to recover. 

Modernisation isn’t always about replacing an outdated system. Sometimes, the best option is to improve what’s already there by simplifying processes or retiring features that no longer add value. Decisions should be led by data and what you already know.

4. Trust has to be designed into adoption

Every conversation about AI in public services eventually returns to trust: trust in the technology, the data, the decision and the organisation using it. 

But the challenge is not only public trust. It is also confidence inside public sector organisations about whether, where and how AI can be used. 

Fear and uncertainty remain barriers. Teams may see valuable uses but hesitate because governance, legal, security, data protection, commercial and assurance processes feel unclear or disproportionate. Others avoid the subject while informal experimentation grows outside approved environments. 

The challenge is finding a practical balance between proper oversight and giving teams the space to explore new ideas. It is to create a safe and proportionate route between unmanaged adoption and organisational paralysis. 

The UK Government’s Artificial Intelligence Playbook for the UK Government provides a sound basis: understand the technology and its limitations; use it lawfully, ethically and securely; retain meaningful human control; manage the full lifecycle; involve commercial colleagues early; build the necessary skills; and apply the right assurance. 

The harder task is putting those principles into practice. That means deciding where AI can genuinely add value, agreeing the right safeguards and being clear about what success looks like. 

This is where external support can help. It should not provide another strategy document but instead help teams move from uncertainty to a controlled proof of value. Done well, small and measurable interventions build confidence because they show what works, what does not and what must be true for the organisation to proceed responsibly.

5. Capability determines whether change lasts 

Perhaps the strongest theme at DigiGov was not technology at all. It was people. Frontline teams often hold the clearest evidence of service breakdowns, manual interventions, practical controls and informally managed risks. Involving them early improves the solution and its adoption. 

Leaders also need enough confidence in digital, data and AI to ask the right questions. They do not need to become technical specialists, but they do need to connect technology choices to service outcomes, operational risk, value for money and organisational accountability. 

Sustainable change requires more than tool training. People need to understand why it matters, how roles will evolve, where judgement remains essential and how to raise concerns. Knowledge should stay within the organisation, not with a supplier or a few individuals. 

The government’s modern digital roadmap recognises this through its emphasis on stronger digital leadership, learning pathways and skills across the public sector. That focus is well placed. Platforms can be implemented; lasting change depends on whether people have the capability and confidence to use them well.

From technology-led activity to evidence-led action 

DigiGov 2026 left me optimistic. This is not because these challenges have been solved, but because the conversation is becoming more practical. Public sector leaders are asking harder questions about outcomes, governance, service design, resilience and value. They want evidence that an intervention solves a real problem, delivers measurable benefit and has a credible route to adoption. 

Whether the challenge is responsible AI or legacy modernisation, the immediate need is a safe, proportionate route from uncertainty to evidence. That starts by identifying the services with the greatest risk or potential impact, understanding the foundations beneath them and choosing a focused next step with clear controls and measures. 

The future of digital government won’t be determined by who adopts AI first or replaces the oldest systems. It will be shaped by organisations that can make confident decisions about where change will create public value and build the capability to sustain it.

Leaders should start by identifying the services that matter most to their organisations and then focus on understanding what’s holding those services back and whether the foundations are in place to support change.

How we can help

If your organisation is working through these questions, TXP can help turn uncertainty into evidence and prioritised action. 

We work with public bodies to: 

• identify where AI could create measurable service or operational value, with a safe and proportionate route through governance and assurance; 
• design focused proofs of value with clear measures and a credible path to adoption; 
• identify the critical services most exposed to legacy and operational risk, including their process, data, technology, supplier and people dependencies; and 
• build an evidence-led roadmap that prioritises investment according to public impact, resilience and value for money.

If these challenges are familiar, we would welcome a conversation about where a focused discovery or proof of value could help your organisation move forward with confidence.