AI news ·
Agentic AI Could Boost Government Productivity by Up to 30%, Former Lead NED Michael Jary Says
Agentic AI could boost government productivity by 20-30%, enabling faster, more transparent services. Success depends on agile teams, new skills, decentralization, and strong leadership.

Agentic AI Could Boost Government Productivity by 20-30%, Former Lead NED Says
Artificial intelligence systems capable of operating with limited human supervision have the potential to increase government productivity by 20-30% or more, according to Michael Jary, former lead non-executive director (NED) in government. Speaking at an Institute for Government event, Jary highlighted how "agentic AI" might enable faster, more accessible, and transparent government services.
Jary envisions a future where government is less siloed and more networked. He suggested that common data sets and shared components, connected through rules and APIs, could transform fragmented services into seamless user journeys. However, he stressed that imposing monolithic structures should be avoided, given the varied nature of government services.
Agile Teams and New Ways of Working
For this future to materialize, Jary said government must adopt agile development methods focused on real user needs. He emphasized shorter feedback loops between policy and design, noting that policy is essentially design. Testing solutions with stakeholders and rapidly moving to production should become standard practice—approaches already common in successful private sector companies.
Jary pointed to India’s Aadhaar digital ID system as a powerful example. Serving 1.4 billion people with a core team of just 100 demonstrated the scale and efficiency achievable with the right approach. He acknowledged that compliance and accountability could be built into such systems but noted ongoing debates about what infrastructure should remain publicly owned versus privately managed.
Skills and Talent for the Future Civil Service
The future civil service will demand a smaller headcount but a broader range of skills, including designers, product managers, engineers, data scientists, and citizen-insight specialists. Jary expressed confidence that government can attract top talent by highlighting the opportunity to tackle significant societal challenges rather than less impactful work.
He also called for a shift away from departmental boundaries toward empowered, multidisciplinary teams. While the current government’s mission-based organisation is promising, Jary warned it risks becoming bogged down by existing silos and coordination challenges.
Decentralisation and Devolution
During his time as a NED at the Ministry of Housing, Communities and Local Government, Jary observed the negative effects of extreme centralisation. He criticized funding models that force local authorities to compete for resources through central government bids. Instead, he advocated for deeper devolution to major cities and empowering local government to deliver services, supported by simplifying central government functions.
Leadership and Civil Service Reform
Transforming the civil service requires exceptional leadership, according to Jary. He supports proposals to split the cabinet secretary role and appoint a full-time chief executive of the civil service but believes this alone is insufficient. Instead, he calls for a mix of “insider/outsider” leaders—those with government knowledge and experience from other sectors.
These views echo Lord Maude’s 2023 civil service reform review, which recommended appointing a "head of civil service" from outside the civil service for at least a decade to drive change. The 2021 Declaration on Government Reform by Michael Gove also committed to creating new entry routes for professionals from outside government, including time-limited project roles.
Recently, Tom Adeyoola, the government’s preferred executive chair of Innovate UK, expressed strong support for "tours of duty" in the public sector. He highlighted the importance of creating an environment where external talent can thrive and contribute meaningfully during limited tenures.
Conclusion
Agentic AI presents a clear opportunity to boost government productivity and improve service delivery. Achieving this will require agile working methods, a reshaped workforce with new skills, decentralised decision-making, and visionary leadership embracing diverse experience. As these changes unfold, government departments can become faster, more connected, and better aligned with citizen needs.
For government professionals interested in AI and its applications, exploring relevant training can be valuable. Resources such as Complete AI Training’s latest AI courses offer practical knowledge to help navigate this evolving landscape.