Microsoft is adding support for the open Agent2Agent (A2A) protocol to Azure AI Foundry and Copilot Studio, aiming to enable AI agents to work together across different platforms.
With A2A support, Copilot Studio agents can call on external agents—even those outside the Microsoft ecosystem or built with tools like LangChain or Semantic Kernel. Microsoft says over 230,000 organizations already use Copilot Studio, including 90 percent of the Fortune 500. Developers can access sample applications, such as automated meeting scheduling between two agents.
A2A defines how a client agent formulates tasks and a remote agent executes them. Tasks can be handled synchronously or over longer periods, with results and status updates—known as artifacts—exchanged via the protocol. Microsoft promises full integration with existing security and governance systems, including Microsoft Entra and audit logging.
Google Launched A2A in April
Google introduced the A2A protocol in April alongside more than 50 technology partners. A2A is designed to let agents work together using standardized interfaces like HTTP and JSON-RPC, regardless of framework or vendor. It also supports audio, video, and interactive user interfaces. Microsoft is contributing to the specification work on GitHub and plans to help drive further development.
A public preview of A2A in Azure Foundry and Copilot Studio is set to launch soon. Microsoft sees protocols like A2A as the foundation for a new kind of software architecture, where connected agents automate daily workflows and collaborate across platforms—without vendor lock-in, but with auditability and control.