What are AI Agents?

An AI agent is a system where a language model operates in a loop: it reasons about a goal, takes actions using tools (web browsing, code execution, APIs), observes the results, and continues until the task is done.

Simple agents answer questions with tool help. Advanced agents can book travel, debug codebases, or run multi-step research — planning, recovering from errors, and asking for clarification.

Key building blocks: tool/function calling, memory across steps, planning and reflection loops, and guardrails. Frameworks and protocols like MCP standardize how agents connect to tools.