AI Systems

Agent, explained

An AI agent is a system that uses a model to choose actions, call tools, and make progress toward a goal over multiple steps.
Back to Glossary

Part of the Semantic Notion technical glossary.

Short Definition

An AI agent is a system that uses a model to choose actions, call tools, and make progress toward a goal over multiple steps.

Intuition

An agent is not just a chatbot response. It can plan, inspect results, use tools, and adapt the next action.

Technical Definition

An agent couples a policy or language model with state, tools, control flow, memory, and stopping criteria to execute tasks in an environment.

Example

A coding agent reads files, edits code, runs tests, and iterates on failures.

Common Misunderstandings

An agent is a system design pattern, not only a model capability.
More autonomy increases the need for permissions, monitoring, and rollback.

Start here

Need the broader concept map?

Return to the glossary index to move from this definition into adjacent AI, ML, mathematics, and computer science terms.