Topics

The vocabulary of governing AI agents

The security industry has spent twenty years building a vocabulary for humans and service accounts: identity, access, permissions, least privilege. AI agents reuse most of those words and quietly change what they mean. An agent with a valid credential and a valid permission can still make a decision nobody authorized it to make.

These pages define the terms one at a time. Each one starts with a plain definition, works through a concrete example, and ends with where the concept stops and what has to sit next to it. Where we are stating a position rather than a settled definition, we say so.

Term map · 16 terms · where BotAris sitshover a term to see its related terms · click to jump

Showing related terms for: cycling through all termsGrouped by where each term sits relative to the authority layer

Authority and governance 3 Authorization at runtime 4 Delegation 3 Accountability 1 Identity and inventory 5

Authority and governance

3 terms

Authority governance

Defining what an agent may decide, enforcing it at runtime, and recording who is accountable.

AI agent governance

The controls that determine what agents may do, enforce it while they run, and name who answers.

AI governance

The broad program, and the part of it that has to run in real time once agents take actions.

Authorization at runtime

4 terms

AI agent authorization

Deciding whether a specific agent action may proceed, beyond what scopes and roles can say.

Runtime authorization

Evaluating each action at the moment it happens, with live context and cumulative state.

AI agent permissions

What an agent can technically do, and the gap between that and what it is allowed to decide.

AI agent least privilege

The minimum authority to complete the task, not just the minimum access.

Delegation

3 terms

Accountability

1 term

AI agent auditability

Showing, for any decision, who authorized it and on what basis. Why action logs are not enough.

Identity and inventory

5 terms

Humans define authority. BotAris enforces it.

Get in touch →See the platform

Topics

The vocabulary of governing AI agents

Authority and governance

3 terms

Authority governanceDefining what an agent may decide, enforcing it at runtime, and recording who is accountable.AI agent governanceThe controls that determine what agents may do, enforce it while they run, and name who answers.AI governanceThe broad program, and the part of it that has to run in real time once agents take actions.

Authorization at runtime

4 terms

AI agent authorizationDeciding whether a specific agent action may proceed, beyond what scopes and roles can say.Runtime authorizationEvaluating each action at the moment it happens, with live context and cumulative state.AI agent permissionsWhat an agent can technically do, and the gap between that and what it is allowed to decide.AI agent least privilegeThe minimum authority to complete the task, not just the minimum access.

Delegation

3 terms

Delegated authorityThe mandate a human grants an agent: which decisions, within what limits, under whose accountability.Human-to-agent authorityThe first delegation in the chain, who should make it, and why the owner field is not enough.Agent-to-agent delegationKeeping authority inside the original grant as agents hand work to other agents.

Accountability

1 term

AI agent auditabilityShowing, for any decision, who authorized it and on what basis. Why action logs are not enough.

Identity and inventory

5 terms

AI agent identityWho the agent is and how it authenticates. Necessary, and not the same as what it may decide.Non-human identity governanceFinding, owning, and reviewing non-human identities. What it does well and where it stops.Workload identityA verifiable identity tied to where software runs. The right foundation for agents, and only the foundation.Agent discoveryFinding the agents in your environment. Useful, and not a control.Shadow AI agentsThe agents nobody registered, and the control that does not depend on finding them first.