The Agentic AI Starter Map
A starter map of the ideas behind agentic AI. It's not comprehensive yet — more topics are on the way. Start anywhere below, or explore how the ideas connect on the interactive map further down.
foundations
- What is an AI agent?An AI agent is a system where a language model decides its own sequence of actions — which tools to call, what to check, when it is done — rather than following a hand-coded script.
- Tool use (function calling)Tool use (also called function calling) lets a model request that an external function be run — a web search, a database lookup, a calculator, an API call — and get the result back to reason over, instead of relying only on facts memorized during training.
- The agentic loop (plan, act, observe)An agentic loop is the repeated cycle an agent runs through: gather context, decide on an action, execute it (often via a tool), observe the result, and decide whether to continue or stop.
- Context managementContext management is the discipline of curating what information — instructions, tool results, prior conversation, retrieved documents — actually sits in a model's finite context window at each step, since a model can only reason well over what's currently in front of it.