# Multi-agent system

Several AI agents working together, each handling part of a problem.

The term's main use here is in [How AI agents fail](/research/ai-agent-failure-modes). There, a UC Berkeley-led study catalogs 14 failure modes found in 1,642 recorded runs of multi-agent systems built on seven frameworks. Systems with a verification step failed less, and clearer instructions for each agent's role raised success rates by 9.4 percent. In [How long does it take to build an AI agent?](/research/how-long-to-build-an-ai-agent), one agency quotes three to six months as the minimum for a multi-agent system.

Wikipedia adds: "Multi-agent systems can solve problems that are difficult or impossible for an individual agent or a monolithic system to solve." The article adds that with large language models, "LLM-based multi-agent systems have emerged as a new area of research".

---

Published 2026-11-05. Web page: https://www.mackenziebowes.com/reference/glossary/multi-agent-system
