Introducing A2H: A Protocol for Agent-to-Human Communication

A2H (Agent-to-Human) is an open-source protocol introduced by Twilio to facilitate standardized communication between AI agents and humans, addressing the need for a unified, channel-agnostic interface in the rapidly evolving AI landscape. This protocol defines five core intent types—INFORM, COLLECT, AUTHORIZE, ESCALATE, and RESULT—enabling agents to execute complex workflows while ensuring each interaction is simple and auditable. By abstracting the complexities of multi-channel communication, A2H allows agents to focus on their tasks without worrying about the specifics of reaching a human, while built-in security features like authentication flexibility and cryptographic evidence ensure trust and auditability. The protocol is designed to integrate seamlessly with existing frameworks like MCP, enhancing agent capabilities in human-centric interactions. Twilio invites the community to explore and refine A2H, offering a roadmap that includes formalizing JSON Schemas and integrating additional agent frameworks, with the ultimate goal of establishing a comprehensive communication model for autonomous agents.