Learn More
← Back to C1 Academy
What is an Agent?
Jess Starr·Product Manager, AI Access Management
Jess, Product Manager for AI Access Management at C1, defines what an AI agent actually is and breaks down the two access models, autonomous and on-behalf-of-human, that determine how agents show up in your systems.
Main Takeaways
- An AI agent is a new kind of actor. Instead of waiting for prompts, agents do real work through model loops: they plan steps, call tools, read results, and go again.
- That loop means agents reach into and act on production systems with little to no human intervention.
- There are two access models. As an autonomous worker, the agent acts independently and needs its own identity so your systems show clearly that the agent did the acting.
- When an agent acts on behalf of a human, it must be tied to that human so the relationship is reflected in downstream access.
- Organizations need both models in play to give agents enough access to be productive while keeping access at a level security teams can stand behind.