Definition
A multi-agent system involves multiple AI agents working together, each with specialized roles, to accomplish complex tasks. Agents can collaborate in different patterns: hierarchical (manager delegates to workers), sequential (pipeline), or peer-to-peer (equal collaboration).
Example
A content creation pipeline where a Researcher agent gathers information, a Writer agent drafts the article, an Editor agent reviews and polishes, and a Publisher agent formats and posts, all coordinated automatically.
Why it matters
Single agents work well for focused tasks. Multi-agent systems handle complex business processes that span multiple domains. Research shows that multi-agent architectures reduce task completion time by 40-60% compared to single-agent approaches.