Very pleased to have had the opportunity to present at Services Week 2025, an annual event hosted by the Government Digital and Data community. Services Week brings together professionals from across the UK public sector to share best practices and learn from others, all with the aim of transforming lives through high-quality, technology-powered services.

The 2025 theme, people-centred services, emphasised designing effective public services through innovation, a strong understanding of user needs, and close collaboration between different teams and specialisms.

My session, titled “Designing at the Speed of Trust: Building Services in the AI Age,” focused on the critical challenge of adopting and designing AI-powered services that are not only efficient but also trustworthy, transparent, and grounded in real public value, examining the pace and purpose of how we’re adopting artificial intelligence in government.

In this post, I’ll share the full presentation and key takeaways, exploring how we can integrate policy, service design, IT, data, and frontline experience into a collaborative process that understands both the risks and promises of AI. Whether you’re a technologist, service designer, or policymaker, I hope it provides valuable insights into building services that genuinely work for people.

Key Takeaways

Why We’re Here
AI is already weaving its way into public service delivery—but there’s a fundamental problem: we’re either too hesitant, or too reckless.

Some organisations move too slowly, paralysed by uncertainty, lack of clarity, or fear of missteps.

Others charge ahead, lured by the buzzwords and promises of AI as a “silver bullet” solution.

Neither extreme serves the public well. In reality, AI is inevitable in the future of public services. But it’s how we design and deploy it that will determine whether it works for people—or against them.

Trust vs Ethics vs Security vs Compliance

All of these elements are critical not interchangeable.
Trust sits at the top of this pyramid.

A Framework for Trust in AI
To bridge this trust gap, I proposed a framework rooted in five key principles:

Solve the Right Problem
Don’t start with AI. Start with the service problem. Use technology only when it’s the best-fit solution.

Engage Stakeholders
Cross-disciplinary teams—policy, tech, service design, data science, frontline staff—must all be involved from day one.

Develop & Test Transparently
AI must be explainable and auditable throughout its lifecycle. If it can’t be challenged, it can’t be trusted.

Embed Governance
Ownership, accountability, and oversight aren’t optional. They’re essential.

Communicate & Iterate
Trust is built through clarity, honesty, and adaptation over time. Don’t “set and forget” AI systems.

Trust-Designed AI vs. Reactive AI

Problem Definition Clear problem first, AI second AI used for its own sake
Team Structure Cross-disciplinary collaboration Tech-led silos
Governance Robust oversight and auditability Minimal accountability
User Communication Transparent explanations Black-box decisions
Outcome Trusted, effective services Likely public rejection

When trust is designed in from the outset, AI enhances public service delivery. When it’s bolted on afterward—or ignored entirely—we risk waste, harm, or disillusionment.

Breaking the Silo Mentality

Too often, public sector AI is designed in silos. The reality is we need five distinct perspectives in the room from the start:

Bringing these voices together isn’t just good practice. It’s non-negotiable if we want outcomes that work in the real world.

Designing for the Future

Communication
In particular, language plays a powerful role in either deepening trust or damaging it. Sensationalist language around AI—“whiplash adoption,” “injecting AI into the veins of government”—may excite some, but risks alienating the public. We need calm, clear, human conversations.

What Comes Next?
We stand at a crossroads. AI will become central to how we govern, serve, and solve problems. But speed alone isn’t enough. Without trust, we risk rejection, failure, and harm.

So I’ll end with the same question I posed in my talk:
What will you do to ensure AI is trusted in your services?

📨 Want to talk about AI governance or service design in your organisation? Get in touch at r@govfutures.uk or connect on LinkedIn.