In the past year, we have witnessed AI agents evolve from experimental tools into essential components of enterprise systems. From simple prompt and response bots to agents that act autonomously on your behalf, this shift marks a new era in software design, where intelligence is no longer confined to static interfaces or single applications.
At Microsoft, we have experienced this transformation firsthand. Azure AI Foundry is now used by developers at over 70,000 enterprises and digital native companies, including Atomicwork, Epic, Fujitsu, Gainsight, H&R Block, and LG Electronics, to design, customize, and manage AI applications and agents. In just four months, over 10,000 organizations have adopted our new agent services to build, deploy, and scale agent systems. More than 230,000 organizations, including 90% of the Fortune 500, are already using Microsoft Copilot Studio.
As agents take on more complex functions, they need access not only to various models and tools but also to each other. That’s why we are committed to advancing open protocols like Agent2Agent (A2A), which is coming soon to Azure AI Foundry and Copilot Studio, that will enable agents to collaborate across clouds, platforms, and organizational boundaries.
We are aligning with the broader industry push for shared agent protocols—doing what we’ve always done: embracing openness, supporting real-world developers, and turning experiments into enterprise-grade platforms. Our goal is simple: to empower professional and citizen developers to build agents that interoperate across clouds and frameworks.
We believe that Microsoft Copilot will empower every employee, and serve as the “AI user interface” to connect agents and agent systems – networks of agents that can reason, act, and adapt across boundaries. As customers scale these systems, interoperability is no longer optional. They want their agents to coordinate tasks that span vendors, clouds, and data silos. They want control, visibility, and trust—without being locked in.
A2A enables structured agent communication – securely and observably exchanging goals, managing state, invoking actions, and returning results. Developers can use the tools they know and love, like Semantic Kernel or LangChain, and still interoperate. Every call is secured with enterprise-grade security measures: Microsoft Entra, mutual TLS, Azure AI Content Safety, and full audit logs. Azure AI Foundry is built with trust by default, and as the agent ecosystem becomes more open and distributed, security, compliance, and accountability remain paramount.
By supporting A2A:
- Azure AI Foundry customers can build complex, multi-agent workflows that span internal Copilots, partner tools, and production infrastructure – while maintaining governance and service level agreements (SLAs).
- Copilot Studio agents will be able to securely call external agents, including those built on other platforms or hosted outside of Microsoft.
- Enterprises gain a path to composable, intelligent systems that can scale across organizational and cloud boundaries.
- Microsoft’s contributions will accelerate the development and adoption of open A2A protocols across the industry.
This is just one step in a long journey. As we’ve done with innovations like Autogen, Semantic Kernel, our contributions to the Model Context Protocol (MCP), and our open model catalog, we will continue to evolve the platform to support the protocols, models, and frameworks that matter most to developers and enterprises. We see protocols like A2A and MCP as important steps toward realizing our vision for the future of agents.
Agentic computing is not a trend – it is a fundamental shift. It changes how software is built, how decisions are made, and how value is created.
We’ve joined the A2A working group on GitHub, contributing to specifications and tooling. Public previews of A2A in Foundry and Copilot Studio are coming soon.
By supporting A2A and building on our open orchestration platform, we are laying the foundation for the next generation of software – collaborative, observable, and adaptive by design. The best agents won’t live in one app or cloud; they’ll operate in workflows that span models, domains, and ecosystems. We are building for that future with openness at the core – because agents shouldn’t be islands, and intelligence should work across boundaries, just like the world it serves.
We’ve added a .NET and Python sample in Semantic Kernel that shows two local agents scheduling a meeting and drafting an email via A2A. Clone the repo, install, run, and watch a real workflow with zero custom code.
The Rise of A2A: A New Era of Cross-Platform Agent Collaboration
The field of artificial intelligence (AI) is undergoing a profound transformation, moving beyond traditional single applications and towards a world composed of interconnected and collaborative agents. At the heart of this shift lies the emergence of open Agent2Agent (A2A) protocols, which promise to break down platform and organizational barriers, thereby unleashing unprecedented potential for agent collaboration.
In the past year, we have witnessed AI agents evolve from their initial experimental stages to become integral components of enterprise-level systems. They are no longer merely simple prompt-and-response bots, but have evolved into intelligent assistants capable of autonomously executing complex tasks on behalf of users. This evolution marks a fundamental paradigm shift in software design, where intelligence is no longer confined within the narrow frameworks of static interfaces or single applications.
Microsoft has been at the forefront of this transformation, deeply recognizing the importance of agent interoperability. Azure AI Foundry, Microsoft’s flagship AI platform, has attracted over 70,000 enterprises and digital-native companies, who leverage the platform to design, customize, and manage various AI applications and agents. These enterprises include industry leaders such as Atomicwork, Epic, Fujitsu, Gainsight, H&R Block, and LG Electronics.
Notably, in just four months, over 10,000 organizations have adopted Microsoft’s new agent services to build, deploy, and scale their agent systems. Furthermore, over 230,000 organizations, including 90% of the Fortune 500, have begun using Microsoft Copilot Studio to build and manage their intelligent assistants.
As agents play increasingly complex and important roles within enterprises, they not only require access to a wide range of AI models and tools, but also need to be able to seamlessly interact and collaborate with other agents. It is precisely out of this need that Microsoft is committed to promoting the development of open protocols like Agent2Agent (A2A). The A2A protocol is poised to land on the Azure AI Foundry and Copilot Studio platforms, enabling agents to collaborate across different cloud environments, platforms, and organizational boundaries, thereby breaking down information silos and enabling more efficient and intelligent workflows.
Microsoft actively responds to the industry’s call for shared agent protocols and consistently adheres to the principles of openness, supporting developers, and transforming experiments into enterprise-grade platforms. Microsoft’s goal is very clear: to empower professional developers and non-professional developers (i.e., citizen developers) to build agents that can interoperate across different cloud environments and frameworks.
Microsoft firmly believes that Microsoft Copilot will empower every employee and become the “AI user interface” connecting agents and agent systems. These agent systems are networks of multiple agents that can reason, act, and adapt across organizational boundaries. As customers continue to expand these agent systems, interoperability is no longer an option but a necessity. Customers want their agents to coordinate tasks across different vendors, cloud environments, and data silos, while maintaining control, visibility, and trust in the system, without being locked into specific platforms or technologies.
The A2A protocol, by providing a structured agent communication mechanism, enables agents to securely and observably exchange goals, manage state, invoke actions, and return results. Developers can continue to use the tools they are familiar with, such as Semantic Kernel or LangChain, without worrying about interoperability issues. The A2A protocol ensures that each call is protected by enterprise-grade security measures, including Microsoft Entra, mutual TLS, Azure AI Content Safety, and complete audit logs. Azure AI Foundry is built with trust as a core principle from the outset, and as the agent ecosystem becomes more open and distributed, security, compliance, and responsibility remain the platform’s top priorities.
The Advantages of A2A: Unleashing the Potential of Agent Collaboration
By supporting the A2A protocol, Azure AI Foundry customers can build complex multi-agent workflows that span internal Copilots, partner tools, and production infrastructure, while maintaining governance and service level agreements (SLAs). Copilot Studio agents can also securely call external agents, including those built on other platforms or hosted outside of Microsoft.
More importantly, the A2A protocol provides enterprises with a path to composable, intelligent systems that can scale across organizational and cloud boundaries. Through the A2A protocol, enterprises can build more flexible, adaptable, and efficient intelligent solutions, thereby gaining a greater advantage in competitive markets.
Microsoft’s contribution to the A2A protocol will accelerate the development and adoption of the protocol across the industry. Microsoft will continue to work with the industry to jointly promote the standardization and improvement of the A2A protocol, thereby laying a solid foundation for the future of agent collaboration.
Towards the Future of Agents: Microsoft’s Open Platform Strategy
Supporting the A2A protocol is just the first step in Microsoft’s journey towards its vision for the future of agents. As Microsoft has done in innovations such as Autogen, Semantic Kernel, contributions to the Model Context Protocol (MCP), and the open model catalog, Microsoft will continue to evolve its platform to support the protocols, models, and frameworks that are most important to developers and enterprises. Microsoft believes that protocols like A2A and MCP are important steps towards realizing its vision for the future of agents.
Agentic computing is more than just a trend; it is a fundamental shift. It is changing how software is built, how decisions are made, and how value is created. Agentic computing will drive innovation across various industries and bring unprecedented opportunities for businesses.
To actively participate in the development and promotion of the A2A protocol, Microsoft has joined the A2A working group on GitHub and is contributing to the protocol’s specifications and tools. Public previews of A2A in Foundry and Copilot Studio are coming soon, which will enable developers to experience the powerful features of the A2A protocol early on.
By supporting the A2A protocol and building on an open orchestration platform, Microsoft is laying the foundation for the next generation of software - collaborative, observable, and adaptive by design. Microsoft believes that the best agents will not reside in one application or cloud, but will operate in workflows, spanning models, domains, and ecosystems. Microsoft is building for that future with openness at the core - because agents should not be islands, and intelligence should work across boundaries, just like the world it serves.
To demonstrate the practical application of the A2A protocol, Microsoft has added a .NET and Python sample in Semantic Kernel that demonstrates how two local agents can schedule meetings and draft emails via the A2A protocol. Developers can clone the repository, install, run, and watch a real workflow with zero custom code.
Embracing Openness, Co-Creating an Intelligent Future
Microsoft’s commitment to the A2A protocol and its continued investment in open platforms reflect its unwavering belief in the future vision of agents. Microsoft believes that by embracing open standards and protocols, and by working together with the industry, we can jointly create a more intelligent, collaborative, and efficient future. In this future, agents will be ubiquitous and bring revolutionary changes to our lives and work.
Artificial Intelligence (AI) agents are rapidly evolving from simple tools to key components of sophisticated enterprise systems. Central to this evolution is their growing need for interoperability, which the Agent2Agent (A2A) protocol aims to address. A2A is emerging as a bridge for communication and collaboration between agents, regardless of the platform, cloud, or organization they reside within.
Microsoft recognizes the significance of this trend and actively supports the development of the A2A protocol. By integrating A2A into Azure AI Foundry and Copilot Studio, Microsoft is empowering developers to build more robust, flexible, and interconnected agent systems.
As the A2A protocol gains traction, we can anticipate several key trends:
- Cross-Platform Collaboration: Agents will be able to collaborate seamlessly across different platforms and cloud environments, breaking down information silos and enabling more efficient workflows.
- Intelligent Automation: Agents will be able to autonomously execute complex tasks, boosting productivity and reducing human error.
- Personalized Experiences: Agents will be customizable to individual user needs and preferences, delivering more tailored experiences.
- Innovative Applications: The A2A protocol will unlock a wide range of innovative applications, such as smart homes, autonomous vehicles, and intelligent healthcare.
The A2A protocol is changing how software is built, how decisions are made, and how value is created. It represents the future of agentic computing, and Microsoft is leading the way.