AI agent
An AI agent is a program built on a language model that completes tasks on its own: it understands a request, plans steps, calls tools, and responds with a result instead of just text.
Also known as: AI Agent, Autonomer Agent
In detail
Unlike a classic chatbot, an AI agent doesn't just answer questions — it acts. It can book a meeting in a calendar, create a ticket in the CRM, send an email, or look up data in a database.
Technically, an agent has three building blocks:
- Language model (LLM) like Claude or GPT for understanding and generation.
- Tools — functions with a clearly defined API the agent can call.
- State (memory) where conversation and intermediate results are stored.
Good agents are specialized: a support agent, a sales agent, an ops agent — coordinated by a router.
Example
A support agent receives the question 'Where's my order 12345?'. It extracts the order ID, calls the shipping API, fetches the DHL tracking link, and replies: 'Your order shipped yesterday, expected delivery tomorrow. Here's the tracking link.' — all in under 5 seconds, without human help.
Related terms
- HallucinationA hallucination is a plausible-sounding but factually wrong answer from a language model — e.g. a made-up quote or an invented software feature.
- PromptA prompt is the instruction a language model receives — it defines role, tone, allowed actions, and provides the context for the response.
- Context windowThe context window is the maximum amount of text (measured in tokens) a language model can process at once — typically 128k to 1M tokens with current models.