Governing AI agents through organisational structure rather than bolted-on oversight. Covers role assignment, domain boundaries, policy constraints, structured meetings for evolving agent scope, and the audit trails that make agent decisions traceable and accountable.