This paper presents a practical, enforceable governance model for AI agents operating in enterprise environments, spanning RBAC, SSO, groups, and more.