Guide
What is an AI agent?
A plain-English answer, with real agents you can try.
The short answer
An AI agent is software that takes a goal and does the work to reach it. It plans the steps, uses tools like a browser, your files, your code or the phone, checks what happened, and keeps going until the job is done or it needs you.
A chatbot answers. An agent acts. That one difference is the line AIcensus draws: we only list agents.
Agent, chatbot, model, tool
A model is the engine, like GPT, Claude or Gemini. A chatbot is a chat window on top of a model that answers questions. A tool does one fixed job when you click it. An agent combines a model with tools and the freedom to decide the next step, so you can hand it a task instead of a question.
The same company can make all four. ChatGPT is a chatbot; ChatGPT Work, its agent mode that completes tasks on websites, is an agent.
The kinds of agents
Personal assistants book, call, shop and keep your calendar, often from a text message (AI assistants you can text).
Coding agents read a codebase, edit files, run tests and open pull requests.
Browser and computer-use agents click, type and fill in forms on real websites.
Customer support agents and sales agents answer customers and run outreach for companies.
What they can do today
Coding agents like Claude Code and Codex take a task and come back with working changes. Browser agents like Comet and Manus finish multi-step jobs on the web. Personal agents like Tab and Poke work over text message, and open-source ones like OpenClaw run on your own computer.
They still make mistakes, so good agents ask before anything you can't undo, like paying or sending.
How AIcensus decides what counts
A listing has to do the work, not just talk about it, and every fact on it must come from the maker's own pages. The full rules are on How we verify.
A short glossary
- Agentic browser
- A web browser with an agent built in that can act on the pages you visit.
- Computer use
- An agent controlling a computer the way a person does: looking at the screen, moving the pointer and typing.
- Coding agent
- An agent that reads and changes code, runs tests and opens pull requests.
- MCP
- Model Context Protocol: an open standard that lets agents connect to apps and data in the same way everywhere.
- Human in the loop
- An agent that pauses for your OK before important steps.
Updated October 14, 2026 by AIcensus.