Glossary · The AI Index

AI Agent

An AI agent is an autonomous AI system that can plan, reason, take actions, and use tools to accomplish a goal — going beyond a chatbot’s single-turn response to execute multi-step tasks on a user’s behalf. Examples include coding agents that write and test code, browsing agents that complete forms, and operations agents that resolve customer-support tickets end-to-end.

How it works

An agent typically wraps a large language model with three things: a memory of the task so far, a set of tools it can call (APIs, code execution, web browsers), and a planning loop that decides which tool to call next. Each step uses model inference; combined, the steps form an autonomous workflow. “Agentic AI” refers to systems built around this loop.

Thank you for reading this post, don't forget to subscribe!

Why it matters

Agents are the defining enterprise AI trend of 2026. McKinsey finds 23% of organizations already scaling agentic AI with another 39% experimenting, and Gartner expects 33% of enterprise software applications to embed agentic AI by 2028 — though it also predicts 40%+ of agentic projects will be canceled by 2027. See the full data in our Enterprise AI Statistics 2026.