Communication to Completion: Modeling Collaborative Workflows with Intelligent Multi-Agent Communication
Yiming Lu, Xun Wang, Simin Ma, Shujian Liu, Sathish Reddy Indurthi, Song Wang, Haoyun Deng, Fei Liu, Kaiqiang Song
2025-10-24
Summary
This paper introduces a new way for multiple AI agents to work together on complicated tasks, specifically focusing on how they communicate to get things done efficiently.
What's the problem?
When AI agents try to collaborate on complex projects, they often struggle because there isn't a good system for deciding *when* and *what* to communicate to each other. Existing systems don't really measure how well agents understand each other's goals, leading to wasted effort and slower progress.
What's the solution?
The researchers created a framework called 'Communication to Completion' or C2C. It has two main parts: first, a way to measure how well the agents' goals line up, called the 'Alignment Factor'. This helps them understand if they're on the same page. Second, it uses a step-by-step approach where agents intelligently decide when to communicate based on how much it will cost in terms of time and effort, and how much it will improve understanding. They tested this on coding tasks with teams of 5 to 17 AI agents.
Why it matters?
This work is important because it provides both a way to *measure* how effective communication is between AI agents, and a practical system for making them communicate better. The tests showed a significant speedup – about a 40% reduction in task completion time – meaning this could be a big step towards building AI teams that can tackle really complex problems.
Abstract
Teamwork in workspace for complex tasks requires diverse communication strategies, but current multi-agent LLM systems lack systematic frameworks for task oriented communication. We introduce Communication to Completion (C2C), a scalable framework that addresses this gap through two key innovations: (1) the Alignment Factor (AF), a novel metric quantifying agent task alignment that directly impacts work efficiency, and (2) a Sequential Action Framework that integrates stepwise execution with intelligent communication decisions. C2C enables agents to make cost aware communication choices, dynamically improving task understanding through targeted interactions. We evaluated C2C on realistic coding workflows across three complexity tiers and team sizes from 5 to 17 agents, comparing against no communication and fixed steps baselines. The results show that C2C reduces the task completion time by about 40% with acceptable communication costs. The framework completes all tasks successfully in standard configurations and maintains effectiveness at scale. C2C establishes both a theoretical foundation for measuring communication effectiveness in multi-agent systems and a practical framework for complex collaborative tasks.