Microsoft is advocating for a more interconnected approach in artificial intelligence known as the ‘agentic web’, where AI systems can interact seamlessly across various platforms. This concept was presented by Kevin Scott, Microsoft’s Chief Technology Officer, prior to the upcoming Build conference.

The Vision for Agentic AI

Scott envisions a collaborative environment for AI agents that operate on a standard protocol, the Model Context Protocol (MCP), developed by Anthropic, a burgeoning AI company. This open-source framework aims to facilitate better cooperation among diverse AI systems, thus enhancing their performance through shared training and knowledge.

“MCP has the potential to create an ‘agentic web’ similar to how hypertext protocols helped spread the internet in the 1990s,” Scott stated.

In essence, agentic AI refers to systems that autonomously carry out specific tasks, making decisions based on training data. This approach is expected to revolutionize fields such as cybersecurity and customer support. Reinforcement learning will play a pivotal role in refining their output and decision-making.

Microsoft intends to leverage this technology to advance its AI capabilities and sustain its competitive edge in the rapidly evolving tech landscape.