Multi user agent teams perform worse than single coordinators
Worse Together: How Performance Breaks Down in Multi-User Multi-Agent Teams
Artificial Intelligence
Summary
When many AI agents each serve different users with different goals and share resources like calendars or budgets, their ability to work together often breaks down. The authors show that teams of agents perform worse than a single coordinator agent acting for everyone, even when the agents can communicate. Problems arise like agents stalling, overriding each other, or making false claims. The authors also explore fixes tailored to specific environments, such as appointing a team lead or requiring agents to check others' messages before acting.
What this means in practice
- •For software development teams: Manage AI agents interacting over shared codebases to reduce conflicts and improve team productivity by implementing coordination strategies identified in the paper.
- •For personal assistant developers: Design group ordering and booking assistants that improve task success rates by preventing conflicting actions and incorporating communication checks between agents.
Authors
Sahan Paliskara, Nattaput Namchittai, Andrew Lampinen
Abstract
People are increasingly delegating tasks to AI agents, and those agents are increasingly encountering other people's agents over shared resources such as a codebase, a calendar, or a budget. When each agent acts for a different user with different goals, coordination often fails, and the group ends up worse off than if a single agent had acted for everyone. We study this multi-user, multi-agent setting across five frontier models and 77 scenarios in four environments: an API key environment in which agents share a compute budget, a clinic in which they share a calendar, a personal assistant environment in which they share a group order or booking, and a merge queue in which they share a release cutoff. In each scenario, we compare a single agent that serves every user (a coordinator) to a team in which each agent serves one user, with and without a communication channel between the agents. Teams deliver worse group outcomes than the coordinator in every environment: without a channel, they completely collapse in two environments, and even with one, coordination overhead creates substantial gaps. For example, in the personal assistant environment, the coordinator fulfills a targeted user request about twice as often as teams. We identify distinct behaviors associated with this poor group-level performance, including stalling as teams grow, overriding each other's actions, and fabricating claims. We find effective but environment-specific mitigations, such as a team lead, explicit procedural instructions, and a platform check that makes an agent read its peers' messages before committing. We will release the API key, clinic, and personal assistant environments as MAMUBench, comprising 74 scenarios for evaluating multi-user, multi-agent coordination.