The next frontier of government is not digital, it’s agentic. Systems that can plan, coordinate, and execute complex workflows with minimal human intervention are beginning to reshape not just service delivery, but the underlying machinery of the state. Yet despite the urgency and potential, most public-sector efforts risk stalling. Governments face unclear priorities, fragmented governance frameworks, and limited institutional readiness to move from experimentation to impact. This session launches the report titled Making Agentic AI Work for Government: A Readiness Framework, introducing the first systematic framework to assess readiness across 70 government functions and to sequence adoption responsibly. The discussion will tackle the critical questions leaders are grappling with now: Where should governments start? What does it take to scale beyond pilots? And how can states unlock the value of agentic AI while preserving trust, accountability, and resilience?