What is AI agent?
Also called: agentic AI · autonomous agent
An AI agent is a model that plans, takes actions with tools (search, code, APIs), observes the results and repeats until a goal is met — instead of answering once.
A plain chat model answers a message. An agent runs a loop: decide what to do, call a tool, read what came back, decide again. That lets it complete multi-step jobs — research a topic across sources, fix a bug and run the tests, book something — with a human checking the important steps.
Agents amplify prompt quality: a vague goal produces expensive wandering. Good agent prompts define the objective, the tools and when to use them, the stopping condition, and what the agent must ask before doing.
Goal: find three venues for a 40-person offsite in Lisbon under €3k. Tools: web search. Rules: stop after 8 searches; never contact a venue; return a table with price, capacity, distance from the airport and source link.
- Set a step or budget limit so loops end.
- List what needs human approval before acting.
- Ask for a plan first and approve it.
Use it right now
Ask our brain anything on the homepage — it remembers the whole conversation — or write a brief in the Studio and see the prompt it compiles to.