AI Glossary

AI agent

autonomous agent, AI agent

An AI agent is a system that, given a goal, plans its own steps, uses tools (search, APIs, code, and so on) and carries out tasks in a loop — rather than simply answering a single question.

Works in a loop: plan → act → check.
Draws on external tools and data.
Suited to multi-step tasks, not just answering questions.

An ordinary model answers once. An agent is given a goal and decides for itself which steps to take to reach it — it can look up data, call an API, run code, then assess the result and adjust its approach.

The key to a safe agent is oversight: clear boundaries on its tools, checkpoints, and the ability to stop the loop.

Related terms

In guides