← All glossary terms
AI fundamentals

AI agent

An AI agent is a system that uses a language model to plan and take multi-step actions toward a goal, calling tools and reacting to results along the way.

Also known as: autonomous agent, agentic system

Last updated 2026-07-16

An AI agent wraps a language model in a loop that can plan, act, and observe. Rather than returning one reply, it breaks a goal into steps, decides which tool to use for each, runs it, reads the result, and repeats until the job is done or it gives up. The model supplies the judgment; the surrounding code supplies the hands.

Tools are what make an agent useful. Give it the ability to search the web, query a database, send an email, or call your own API, and it can chain those actions together. You define the tools and their boundaries; the agent decides when to reach for each. An agent asked to research a lead might search, read several pages, and compile a summary without a human directing each hop.

The trade-off is control. Every extra step adds latency, cost, and a fresh chance to go off track, and errors compound as the chain grows. Agents shine on open-ended tasks where the path is not known in advance, but a fixed workflow you could script yourself is usually cheaper and more predictable.

If you deploy one, put guardrails around it. Cap the number of steps, require confirmation before irreversible actions like charging a card, and log every decision so you can trace what happened. Start narrow and widen scope only as trust builds, because the autonomy that makes an agent powerful is also exactly why it needs limits.

Related terms

Looking for tools in this space? Browse the LaunchHive directory or explore launches.