AI Agents
Learn how AI agents combine models, tools, memory and workflows to carry out multi-step tasks on their own.
An agent is a language model given the ability to act: to call tools, read results and decide what to do next. This series covers how that loop is built, where it breaks, and how to evaluate it.
Read LLM Basics first. References include the research and engineering guidance each topic draws on.
What You’ll Learn
An agent repeatedly reasons about what to do, takes an action with a tool, observes the result and continues until the goal is met. Anthropic distinguishes workflows, where code defines the path, from agents, where the model directs its own process; many problems are best solved with workflows.
Give an agent a goal and tools, and it decides the steps. Give a workflow a goal, and you decide the steps.
The ReAct paper showed that interleaving reasoning with actions such as search improved results on question-answering and decision tasks.