Presenting at the Public Service Data & AI Conference was a fantastic opportunity to share a real-world account of digital transformation in the public sector and celebrate some of the progress we have made over the last year. I’m grateful to the Global Government Forum for giving practitioners and SMEs a platform to share honest experiences and practical lessons learned.
My session, “From Data Silos to Interoperability: Measuring the Business Value of AI”, focused on a challenge that many organisations continue to face: AI success is rarely limited by technology. More often, it is constrained by the quality, accessibility, trustworthiness and governance of the underlying data.
Throughout 2026, both I and Mercator Digital have been privileged to work at the forefront of AI and data transformation within government. This has included supporting one of central government’s first agentic AI monitoring initiatives, for which I authored the Data Protection Impact Assessment (DPIA), alongside the development of multiple AI proof-of-concepts focused on fraud detection, compliance and operational improvement.
While technology often attracts the attention, one of the most significant transformations we observed was cultural. Many of the users we worked with began the journey frustrated by existing systems and struggled to articulate their requirements. Over time, those same users became confident advocates for new data capabilities, actively shaping future solutions and contributing to strategic conversations about data and AI. For me, that shift represented one of the clearest indicators of success.
A topic that generated considerable discussion during the conference was AI measurement and success metrics. Too often, AI initiatives are treated as experiments without clearly defined outcomes. My view is that organisations should establish success measures far earlier in the delivery lifecycle. Measuring adoption, behavioural change, operational outcomes and business value should not be an afterthought. In many respects, the ability to measure success is itself a critical capability that accelerates transformation.
Another consistent theme throughout the event was the appetite for innovation across the public sector. Delegates from a wide range of professions were asking thoughtful questions about how to accelerate their data journeys and realise value from emerging technologies. What became clear is that organisations are ready for innovation, but many remain overwhelmed by the complexity of implementation.
This is where experience matters. One of the key lessons from our work has been that capabilities alone do not create value. Adoption creates value. Organisations can invest in cutting-edge platforms, AI models and analytical tools, but without user engagement, skills development and organisational buy-in, those capabilities struggle to deliver meaningful outcomes.
The final message of the session was one that continues to shape our work today. Rather than creating more application-specific solutions, organisations should focus on building reusable, governed data capabilities. By investing in data products, interoperability, governance and sustainable architecture, public-sector organisations can reduce technical debt, improve reuse, enable automation and create the foundations required for future AI initiatives.
The technologies may evolve rapidly, but the principle remains constant: successful AI transformation starts with trusted data, clear outcomes and people who are empowered to adopt change.
