What Should Agents Say? Action-state Communication for Efficient Multi-Agent Systems

2026-06-03 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Expert, quick

Summary

Multi-agent systems (MAS) leveraging large language models often suffer from inefficient, unconstrained natural language communication, which inflates token usage and context window consumption, impacting performance and inference cost. An analysis of five common inter-agent communication strategies across two MAS topologies revealed no universally optimal fixed approach. Instead, effective messages consistently preserve action-centered information crucial for downstream agents. To address this, researchers propose PACT (Protocolized Action-state Communication and Transmission), a method that treats inter-agent communication as a public state-update problem. PACT projects raw agent outputs into compact action-state records before they enter shared history. This protocol consistently improves the performance-cost trade-off across different MAS topologies, delivering comparable or stronger task performance with substantially fewer tokens. For instance, PACT enhances OpenHands' resolve rate at -10% tokens-per-resolved and maintains SWE-agent's resolve rate while halving input tokens.

Key takeaway

For AI Engineers building multi-agent systems, inefficient communication patterns can significantly increase operational costs and degrade performance. You should consider implementing structured communication protocols like PACT to optimize inter-agent messaging. By compacting raw agent outputs into action-state records, you can achieve comparable or better task performance with substantially fewer tokens, improving your system's efficiency and cost-effectiveness.

Key insights

Unconstrained natural language communication in MAS inflates costs; structured action-state records improve efficiency and performance.

Principles

• Effective inter-agent messages preserve action-centered information.
• No single fixed communication strategy is universally optimal.
• Communication can be modeled as a public state-update problem.

Method

PACT (Protocolized Action-state Communication and Transmission) projects raw agent outputs into compact action-state records for shared history, treating communication as a public state-update problem.

In practice

• Implement PACT to reduce token usage in LLM-based MAS.
• Apply action-state record compression for inter-agent messages.
• Evaluate communication strategies for specific MAS topologies.

Topics

• Multi-agent Systems
• Large Language Models
• Inter-agent Communication
• Token Efficiency
• PACT Protocol
• OpenHands
• SWE-agent

Code references

• iNLP-Lab/PACT

Best for: Research Scientist, AI Architect, NLP Engineer, AI Scientist, Machine Learning Engineer, AI Engineer

Related on AIssential

• What Should Agents Say? Action-state Communication for Efficient Multi-Agent Systems — Structured, action-state communication significantly reduces token usage and improves multi-agent system efficiency.
• Toward Goal-Oriented Communication in Multi-Agent Systems: An overview — Goal-oriented communication optimizes MAS efficiency by prioritizing task-relevant information over raw data fidelity under resource constraints.
• Talk is Cheap, Communication is Hard: Dynamic Grounding Failures and Repair in Multi-Agent Negotiation — Dynamic grounding, not just static understanding, is critical for effective multi-agent LLM coordination and negotiation.
• New Stanford study reveals when teaming up AI agents is worth the compute — Multi-agent AI systems' performance advantage often stems from increased compute, not inherent teamwork superiority.
• 🥇Top AI Papers of the Week — Advancements in AI agent systems focus on structured evolution, efficient reasoning, and dynamic collaboration for improved performance and scalability.
• the layer openclaw builders should care about next — Prioritize structured state for inter-agent communication to reduce cost and improve workflow clarity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.