---
title: "AI Agent: Definition and Examples"
description: "An AI agent is a software program that uses an AI model to work toward a goal on its own: it decides which steps to take, uses tools such as search, APIs or..."
path: /glossary/ai-agent
family: glossary
updated: 2026-10-17
html: /glossary/ai-agent
sources:
  - "https://aws.amazon.com/what-is/ai-agents/"
  - "https://www.ibm.com/think/topics/ai-agents"
  - "https://www.anthropic.com/engineering/building-effective-agents"
facts:
  term: "AI agent"
  also_called: "AI agents, autonomous agent, intelligent agent, LLM agent"
  category: "AI and Agents"
---

# AI agent

An AI agent is a software program that uses an AI model to work toward a goal on its own: it decides which steps to take, uses tools such as search, APIs or a browser, and acts on the results.

## What Is AI Agent?

**In short:** Most AI agents combine four parts: a model that reasons, usually a large language model; instructions that set the goal and limits; tools the model may call, such as search, a CRM, email or code; and memory of what it has done.

The agent runs a loop of plan, act and check until the task is done or it hits a limit.

IBM sorts agents into five types by how they decide, from simple reflex agents that follow if-then rules to model-based, goal-based, utility-based and learning agents. In practice the useful split is narrow versus broad. A narrow agent does one job, such as classifying replies as interested, not now or unsubscribe. A broad agent takes a goal like research these 50 accounts and picks its own steps. Narrow agents are easier to test and trust.

## AI Agent Example

For example, a reply-triage agent reads the 300 replies a campaign gets in a week. It tags 40 as interested, 90 as out of office and 25 as unsubscribe requests, removes the 25 from future steps, and drafts answers to the 40 interested replies for a rep to review before anything is sent.

## AI Agent in Practice

Sales teams use agents for account research, list building, enrichment, reply triage, meeting prep and CRM updates. The useful question about any AI agent product is which actions it may take without approval, and how you can see what it did.

## FAQ

### What are the 5 types of AI agents?

IBM lists simple reflex, model-based reflex, goal-based, utility-based and learning agents, ordered from simplest to most capable.

### What is the difference between an AI agent and a chatbot?

A chatbot answers inside a conversation and waits for your next message. An AI agent works toward a goal on its own, choosing actions and calling tools, and can finish a multi-step task without a person prompting each step.

### What is the difference between an AI agent and a workflow?

In a workflow, code fixes the order of steps. In an agent, the model decides which steps and tools to use. Anthropic groups both under agentic systems.

### What can AI agents do in sales?

Research accounts, build and enrich lead lists, draft personalized outreach, sort replies, prepare meeting briefs and update the CRM. Most teams keep a person approving anything that reaches a customer.

## Do this in Ken

### Do This in Ken

Point your own AI agent at the Ken MCP server to give it Ken as a set of tools.

- Link: [Do This in Ken](/mcp)

## Related

- [GTM Glossary](/glossary): Every GTM term, grouped by topic.
- [Agentic AI](/glossary/agentic-ai)
- [AI agent builder](/glossary/ai-agent-builder)
- [LLM (large language model)](/glossary/large-language-model)
- [Model Context Protocol (MCP)](/glossary/model-context-protocol)
- [AI SDR](/glossary/ai-sdr)
- [Prompt engineering](/glossary/prompt-engineering)
- [Ken MCP server](/mcp)

## Sources

| Source | URL | Checked |
| --- | --- | --- |
| AWS: What are AI agents? | https://aws.amazon.com/what-is/ai-agents/ | 2026-10-17 |
| IBM: What are AI agents? | https://www.ibm.com/think/topics/ai-agents | 2026-10-17 |
| Anthropic: Building effective agents | https://www.anthropic.com/engineering/building-effective-agents | 2026-10-17 |

## Changelog

- 2026-10-17: Page published.