Control plane
An AI agent control plane sets and enforces the rules for agents, apart from the agents that do the work. It decides who each agent is, what it may call, and what is recorded.
In a run
The agent does its job in its own code. Ares sits on its path: each call is checked against the policy before it reaches a tool or another agent, and each outcome is recorded.
Our word for it
Limits set from outside the agent.
Why keep it apart from the agent
Rules written into an agent’s prompt or code change whenever someone changes the agent, and a model can be talked out of them. Rules set outside the agent apply whatever the agent decides.
With Ares there is no framework to adopt and nothing to fork. Your team keeps building agents the way it does now.
What it decides on each call
- Which agent is calling, under which credential, and for whom.
- Whether the policy allows this call to this tool or agent.
- Whether the content passing in or out needs a filter.
- Whether the step waits for a person.
No route around it
An agent can connect through a private tunnel, so it has no route around Ares. A public endpoint is also possible, when you accept that trade.
Where it stops
It covers only the agents you enrol. It does not replace your identity provider. It sits on top of it. It also does not write your rules: you approve them.
See how one call takes the same path, from sign-in to the record.
Related terms
- AI agent platformAn AI agent platform is the software that AI agents run on and are managed through.
- Policy as codePolicy as code means writing rules, such as access rules, as text a machine can check, under the same review and change control as software.
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.