AGNTCY (pronounced “agency”) is an open-source protocol initiative aimed at creating a universal language and infrastructure for AI agents to discover, communicate, and collaborate across different platforms and organizations. As AI agents become more central to business operations—handling tasks like data analysis, customer service, or even autonomous decision-making—the need for them to work together in a secure, scalable way has grown significantly. AGNTCY addresses this by offering a shared set of standards and tools to ensure agents built by different developers or companies can interoperate smoothly, regardless of the underlying technologies they use.

The concept behind AGNTCY is inspired by the early internet’s unifying protocols like TCP/IP, which enabled computers to connect and share data across global networks. In much the same way, AGNTCY envisions an “Internet of Agents,” where intelligent systems can seamlessly link up, exchange information, and act together in complex workflows. This is especially important as businesses begin integrating multiple AI agents across departments, tools, and data sources. Without a common protocol, these agents would remain isolated, limiting their usefulness.

For leaders exploring agentic AI, the Agentic AI and AI Agents for Leaders course offers practical guidance on designing and deploying AI agents using tools like custom GPTs. It’s a great next step if you’re interested in protocols like AGNTCY and AI interoperability.*

To bring this vision to life, AGNTCY includes several technical components. The Agent Gateway Protocol (AGP) is designed for secure communication between agents, using patterns like request-response and real-time streaming. Meanwhile, the Agent Connect Protocol (ACP) defines how agents can be started, configured, and managed remotely—much like launching an app in a cloud environment. Evaluation tools are also part of the package, allowing developers and teams to test and compare how well their agents perform under different conditions.

AGNTCY is backed by a coalition of influential tech players including Cisco, LangChain, LlamaIndex, Galileo, and Glean. It’s an open-source, community-driven project, inviting contributions from across the AI landscape. By doing so, it aims not just to solve a technical problem, but to set the foundation for a more open, efficient, and cooperative AI future—one where intelligent systems can work together as easily as humans exchanging emails or apps sharing data today.