Hey, Government! workshop put agentic government’s political problem at the center
In Berlin, a workshop anchored by a Mira demonstration moved quickly beyond AI’s technical promise to the harder questions of trust, process reform, public procurement, and political will.
The Hey, Government! workshop of the Government Tomorrow Summit-2026 opened with a direct challenge: what must be true for governments to move from digital services to genuinely agentic government—and how can readiness be assessed?
A demonstration of Metafore.ai’s Mira framed the discussion in practical terms. The service was presented as a consent-based interface that could help citizens complete tasks such as renewing a driving licence, while connecting responsible institutions behind a single conversation. Participants focused less on the interface itself than on the conditions required to make such systems work: interoperable but secure data, legacy technology, regulatory constraints, accountable decision-making and the political difficulty of breaking institutional silos.
Ukraine’s participants described Diia’s step-by-step approach: redesign services, change the relevant legal framework and build digital infrastructure before adding agentic capabilities. They argued that transformation must come before digitization, with mobile-first services, clear citizen value and senior digital leaders inside agencies helping to drive change.
Trust became the session’s central theme. Speakers raised concerns about imperfect or exclusionary data, cybersecurity, transparency, consent, algorithmic opacity and the risk that automation could reproduce dysfunctional processes faster. One recurring formulation captured the caution: digitizing a broken process produces a faster broken process.
The workshop also widened the frame beyond efficiency. Participants debated whether AI-enabled government should be measured by reduced administrative friction, better-quality public services, greater inclusion, citizen autonomy or democratic values. Alexandre Pointier closed by grouping the challenge into three tasks: make AI-enabled public transformation a political project, make it workable inside administrations, and make it desirable for citizens.
