AI Glossary
Multi-agent system
What does “Multi-agent system” mean?
Multi-agent system: Setup in which several agents, each with its own instructions, tools or context, cooperate on one job. Splitting work this way can allow parallel effort and specialization, but adds cost and new ways for errors to spread between agents.
What is an example of “Multi-agent system”?
A research system may send several agents to search different sources at once, then have another merge their findings.