AI Agent Governance Part 2 - What Good Looks Like: Governing AI Agents in Practice

**AI agent governance must move from principles to operational control.** The paper argues organizations should treat AI agents as formal organizational actors by issuing agent charters (boundaries, authority, escalation), embedding roles for sponsorship, stewardship, risk/compliance, and operations control, and enforcing runtime monitoring, dynamic permissioning, and intervention. It advocates a maturity-based autonomy approach (assist → supervised → conditional → full), staged scaling with gated oversight, and Agent Risk Management (ARM) to ensure accountability, separation of powers, and resilience to leadership failure.