← Concept Index

Agents — core

The agent loop

Also called: perceive-plan-act-observe, ReAct loop

DEFINITION

An agent wraps a model in a loop: perceive the situation, plan a next step, act by calling a tool, observe the result, and repeat until the task is done or a stop condition is hit.

WHY IT MATTERS

The loop is what makes an agent able to do things — send email, browse, run code — instead of only describing them. It's also where the risk concentrates: the model's guesses now have real-world effects.

COMMONLY CONFUSED WITH

A chatbot. A chatbot answers; an agent takes actions in the world and reads back what happened.