Glossary · 6 · Quality, risk and governance
Agent identity
Also known as: Non-human identity (AI agent), AI agent identity, Workload identity (AI agent)
German: Agentenidentität
An agent identity is a distinct, authenticated identity under which one AI agent acts in other systems — with its own credentials, roles and audit records — so that its actions can be told apart from those of the people and other agents it works for, limited and revoked on their own.
- Expert
- Developers
- Technical project managers
In one sentence
An agent identity lets systems tell an AI agent’s actions apart from people’s — with its own credentials, roles and logs that can be revoked alone.
Example
The documentation agent signs in to the CCMS as “agent-docs-01”, not with the writer’s account; every change it makes is marked as its own and its access can be withdrawn without locking the writer out.
Explained in context
This term is a station of a mini course
- When an AI agent crosses the line: detect, contain and restore operationsStation: Scope, allowlist and own identityStart the course
Context cards connect this term with others to answer one question. Also in British English and German.
- Permission, capability, mandate: the three parts of an AI agent’s boundaryWhich boundary does an AI agent cross when it “goes too far” — and why does it matter which one?UKDeutsch
- When an AI agent crosses the line: detect, contain and restore operationsWhat has to be in place when an AI agent does something it isn’t allowed to do — in IT systems, IoT fleets, OT and on machines — and how is it detected, stopped and released again?UKDeutsch
How it applies
- Not the user's account: An agent that works with a borrowed human account inherits all its rights and makes every log ambiguous. OWASP recommends a separate identity per agent and short-lived, scoped credentials.
- Delegation is explicit: When an agent acts on behalf of a person, the record should show both: who delegated and which agent acted. Standards for agent delegation are still being drafted (several IETF Internet-Drafts as of 2026); established building blocks are service accounts, OAuth 2.0 and token exchange.
- Revocable on its own: A separate identity is what makes credential revocation during containment fast and precise.
- In OT and IoT: Agents that reach controllers, gateways or device clouds need identities registered like any other remote user or service, inside the access model of IEC 62443 where it applies.
Why it matters on your learning path
- Technical writers: Logs and change histories become readable when agent and human are named separately.
- Technical marketers: “Every agent action is attributable” answers a question security buyers always ask.
- Technical project managers: Require an identity per agent in the procurement criteria.
- Developers: Never reuse human credentials for an agent; issue short-lived tokens per task.