Agent governance
GobiernoAI agent governance
Agent governance is the set of policies, permissions, limits and audit that decides what an autonomous agent may do: which data and systems it can reach, which actions it executes on its own and which need human approval, and how everything is logged so it can be reviewed later.
When an agent stops suggesting and starts acting (paying, sending, changing records), the question is no longer technical but one of accountability: who answers when it gets it wrong. Agent governance answers with concrete rules: agent identity and permissions, spend and scope limits, approval points (human-in-the-loop), an immutable action log and a stop button. It is what separates 'we have an agent' from 'we have an agent we can defend to an auditor, a client or a regulator'.
How it differs from
- Human-in-the-loop
- Human-in-the-loop is one mechanism inside governance; governance is the full framework of policy, permissions and audit.
Examples
- Requiring human approval before an agent issues a refund above a threshold is agent governance.
- Giving each agent an identity and a permissions cap so it cannot touch systems outside its task.
FAQ
- Do I need governance if my agent only reads data?
- Less, but not zero: even a read-only agent touches data that may be sensitive. The level of governance should scale with what the agent can do and break.