AI agent

An AI agent is a system where a language model does not just answer, but plans and executes multi-step tasks: it calls tools and APIs, reads the results and decides the next action toward a goal. Unlike a chatbot, an agent acts in a loop until the task is done.

An AI agent wraps a language model in a loop: the model receives a goal, decides which action to take, calls a tool — a search engine, a database, a browser, code execution — observes the result, and plans the next step. The loop continues until the goal is reached or the agent gives up and reports back.

Tools are what separate agents from chatbots. A chatbot can only describe how to book a meeting; an agent checks the calendar API, finds a free slot and sends the invite. Frameworks standardize this pattern with function calling, tool schemas, and protocols like MCP that let one agent discover and use many external services.

The hard problems are reliability and control. Agents compound errors across steps, so production systems add guardrails: constrained tool permissions, human approval for destructive actions, verification loops, and budgets on steps and spend. The rule of thumb: the more autonomy, the more observability you need.