Agent identity
Agent identity is what makes an AI agent one named party that every system can recognise, separate from the people and other agents around it. It says which agent this is and which group in your organisation it belongs to.
In a run
The record names the agent on every line: which agent acted, for whom, and under which credential. When one agent hands work to another, both names appear, so any step traces back to the agent that took it.
Our word for it
Each agent signs in as itself.
Identity first, then authentication
Agent identity is who the agent is. Agent authentication is how it proves that on each call. Ares keeps the two apart: the identity lasts while the agent is enrolled, and each credential lasts for one job.
What Ares holds for each agent
- Its own identity, enrolled from your directory. No agent shares an account with another.
- An owning group in your directory. A named owner and a review date for every agent are planned.
- Its entry in the agent registry, whether we built the agent or your team did.
Revoking an identity
Decommission an agent and its credentials stop working, in the middle of a job if needed. Every call after that is refused, and the revocation goes on the record.
Where it stops
An identity says who the agent is. It does not make the agent’s work correct, and it does not reach agents that run outside Ares.
See how a permission is issued, narrowed and revoked on the guarantees page.
Related terms
- Agent authenticationAgent authentication is how a system knows which AI agent is calling it.
- Non-human identityA non-human identity is an account or credential used by software rather than a person, such as a service, a script or an AI agent.
Tell us about one workflow.
We will say whether an agent should do it, and what the first step costs.
A senior member of the team replies within two business days.