An AI agent is more than an LLM with tools. It's a system architecture: observe → plan → act → observe. The agent loop continues until the task is complete or a termination condition is met.
Modern agent frameworks (Claude Agent SDK, LangGraph, CrewAI) provide the scaffolding: tool registries, memory management, planning strategies, and error handling. The LLM provides the reasoning that decides which action to take at each step.