Every vendor's marketing page calls its chatbot an "agent" — so what do AI agents actually do today? The honest answer is smaller and sharper than the hype: a handful of real categories, each with a few leading products that have verifiable capabilities, prices, and limits. Here is the practical map.
Hold products to the engineering definition
Anthropic's field-defining distinction still holds: workflows orchestrate LLMs through predefined code paths, while agents let the LLM dynamically direct its own process and tool usage. The practitioner's shorthand: an agent runs tools in a loop to achieve a goal. Models think. Agents think, decide, and act.
To a vendor's marketing team, "agent" can mean anything with a text box. To an engineer, it means the observe–think–act–observe loop with real tool calls. When you shop, insist on the engineering definition — otherwise you'll buy a chatbot with a new coat of paint.
The 5 categories of AI agents
Products rebrand, merge, and disappear — so learn the categories, not the names:
- General-purpose assistants with agentic abilities — conversational AI that takes multi-step actions: browsing, tool use, code execution, app connectors. ChatGPT agent mode, Claude (with Claude Code and Claude in Chrome), Gemini across Workspace, Perplexity Pro. Best fit: ad-hoc knowledge work, brainstorming, drafting.
- Coding agents — the category with the hardest evidence of real utility. From inline assistants to agents that take a GitHub issue and return a tested PR: Copilot (~$10/mo), Cursor (~$20/mo), Claude Code, OpenAI Codex, Replit Agent (~$25/mo), Windsurf (~$15/mo), Gemini Code Assist, Devin. A tested heuristic: pick by your dominant constraint — Cursor for IDE experience, Claude Code for terminal loops, Copilot for issue-to-PR cycle time.
- Research and information agents — multi-step web investigations that return cited findings (Deep Research-style).
- Workflow and business automation agents — agents operating CRMs, inboxes, spreadsheets, and APIs.
- Specialized and enterprise agents — voice agents, customer-support agents, DevOps incident-response tooling.
A framework for picking the right one
- Hold products to the engineering definition — real tool calls in a loop, not a chat box with a new name.
- Learn the category, not the brand — the five categories stay useful while names churn.
- Match the category to your task — ad-hoc generalist work → assistants; shipping software → coding agents; cited investigations → research agents.
- Check prices and limits against official pages — the 2026 landscape moves fast; verify before budgeting.
Go deeper
The full guide maps every category with leading products, real 2026 prices, and the honest limits of autonomy — read it here:
👉 What AI Agents Are Available in 2026: A Practical Guide to Every Category
Originally published on Agentic Affairs (https://agentic-affairs.com)
Top comments (0)