Last year, DSIT shared that i.AI built 22 AI prototypes in 12 months, eleven of which reached Alpha or Beta stage. Over 1,000 probation officers now use bespoke AI transcription in their daily meetings with offenders, freeing them to focus on rehabilitation rather than paperwork. An AI-powered consultation analysis tool slashed the time to review a single response from hours to a median of 23 seconds. These are not theoretical projections - they are operational realities reshaping frontline public services right now.
Yet the National Audit Office found that only 37% of government bodies had deployed AI, while 70% were still piloting or planning. The gap between ambition and delivery remains the central challenge. The Prime Minister's AI Exemplars Programme is designed to close it, taking a portfolio approach that spans healthcare, justice, education and planning - testing AI where it can reduce bureaucracy, improve consistency and return time to the people who deliver services.
Co-hosted by Jeremy Gould, Deputy Director, Public Sector AI Delivery at Government Digital Service, and Sumitra Varma, Deputy Director, Probations Data, at the Ministry of Justice, the AI Use Cases discussion table at Government Transformation Summit (25 June, Westminster), supported by RedRock, puts DDaT Civil Servants in direct conversation with two leaders driving that shift.
Jeremy Gould brings over 25 years in digital and 20 in public service, including Chief Digital Officer roles at Homes England and leadership positions at the Ministry of Defence and the Foreign Office. His focus at GDS is operationalising the "scan, pilot, scale" model - building common enablers around procurement, recruitment, security assessments and responsible AI frameworks so that successful prototypes don't stall at the pilot stage.
Sumitra Varma leads data management for The Probation Service, a statutory criminal justice service supervising offenders on community sentences or released from prison. With over two decades across private and public sector software roles - from senior developer at the Home Office to engineering lead at the Medicines and Healthcare products Regulatory Agency - Sumitra understands the practical realities of making complex operational data AI-ready while maintaining the integrity that justice systems demand.
Together, they represent both the centre of government driving AI adoption and the departmental frontline implementing it. The questions they confront daily are the questions every public sector leader will face: How do you move from a promising prototype to a production-grade service used by thousands? How do you build the data foundations that AI requires in environments shaped by legacy systems and statutory obligations? How do you maintain responsible innovation at speed?
The AI Opportunities Action Plan committed to 50 recommendations and has delivered on 38 within 12 months. The government's target of 90% digital interactions by 2030 demands that AI becomes embedded infrastructure, not isolated experimentation. With one million AI training courses already delivered - ahead of schedule on the path to 10 million by 2030 - the workforce readiness pipeline is accelerating.
This is a working session for government leaders ready to move beyond strategy documents and into delivery. Bring your use cases, your blockers and your ambitions. Leave with practical approaches tested by peers who are already building at scale. Find out more below.