AI, data, and the delivery gap in government

The NAO's March 2024 report on AI in government makes a useful read: around 70% of government bodies are piloting or planning AI, and only about a third have actually deployed it so far.

I see these findings reflected in real delivery settings across government and public services. From experience, the barriers aren't usually about the algorithms — they're about the environment.

• Legacy systems that limit integration and scalability
• Fragmented or poor-quality data
• Skills gaps and capability constraints
• Inconsistent governance and unclear accountability

So where does the PMO come in?

• Align AI projects with business strategy and policy goals
• Enforce governance, standards, and benefits tracking
• Coordinate dependencies across teams and departments
• Keep efficiency, outcomes, and value for money at the centre

“AI is powerful — but without delivery discipline, it can easily amplify existing problems.”

A well-structured PMO helps governments move from experimentation to execution — turning innovation into measurable, lasting public value.