Glossary

AI agent platform

An AI agent platform is the software that AI agents run on and are managed through. It handles who each agent is, what it may reach, how it talks to tools and other agents, and what is recorded.

In a run

Every call takes the same path. The agent signs in as itself, gets a credential for this job, and has its policy checked. Inbound content is read at the door, the call reaches the tool, the answer is checked before it leaves, and the decision is recorded.

Our word for it

Ares, the platform.

The parts of Ares

  • Credentials for agents, scoped to each job and limited in time.
  • Access policies, checked on every call.
  • Governed communication between agents, and between agents and tools.
  • Workflows, with human approval where you want it.
  • Content inspection, on the way in and on the way out.
  • The record, which an outside party can check with the open verifier.
  • An agent registry for the agents you already run.
  • Deployment in the cloud, hybrid or on-prem.

They come in one product. Each part supports at least one of the five guarantees.

Build or buy

Your teams can build agents. Credentials per job, policies checked on every call, inspection, and a record an outsider can check are a product in themselves.

What it sits on

Your identity provider stays. It knows who each agent is. Ares adds what each job may do, inspects what passes and keeps the record.

Where it stops

Ares does not replace your identity provider. It does not write your business rules: you approve them.

See the platform and the evaluation steps.

Related terms

  • Control planeAn AI agent control plane sets and enforces the rules for agents, apart from the agents that do the work.
  • AI agentAn AI agent is software that uses a model, usually a language model, to decide its next step and then acts on it.

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.