TeamFusion: Supporting Open-ended Teamwork with Multi-Agent Systems

TL;DR

TeamFusion is a multi-agent system that enhances open-ended teamwork by facilitating structured discussions, surfacing disagreements, and synthesizing consensus-oriented deliverables, outperforming traditional aggregation methods.

Contribution

It introduces a novel multi-agent framework that supports open-ended teamwork through proxy agents, structured discussions, and iterative synthesis, addressing limitations of existing answer aggregation approaches.

Findings

Outperforms baseline aggregation methods across multiple metrics.

Effectively surfaces disagreements and consensus in team discussions.

Improves representation of individual views in team decisions.

Abstract

In open-ended domains, teams must reconcile diverse viewpoints to produce strong deliverables. Answer aggregation approaches commonly used in closed domains are ill-suited to this setting, as they tend to suppress minority perspectives rather than resolve underlying disagreements. We present TeamFusion, a multi-agent system designed to support teamwork in open-ended domains by: 1. Instantiating a proxy agent for each team member conditioned on their expressed preferences; 2. Conducting a structured discussion to surface agreements and disagreements; and 3. Synthesizing more consensus-oriented deliverables that feed into new iterations of discussion and refinement. We evaluate TeamFusion on two teamwork tasks where team members can assess how well their individual views are represented in team decisions and how consensually strong the final deliverables are, finding that it outperforms direct aggregation baselines across metrics, tasks, and team configurations.