Microsoft’s chief technologist stated on Sunday, ahead of the company’s annual software developer conference, that Microsoft envisions a future where artificial intelligence agents from any company can collaborate with agents from other firms and possess improved memories of their interactions.
Microsoft is hosting its Build conference in Seattle on May 19, where analysts anticipate the company will unveil its latest tools for developers building AI systems.
Speaking at Microsoft’s headquarters in Redmond, Washington, prior to the conference, Chief Technology Officer Kevin Scott informed reporters and analysts that the company is focused on facilitating the adoption of industry-wide standards that will enable agents from different developers to work together. Agents are AI systems capable of independently executing specific tasks, such as resolving a software bug.
Scott mentioned that Microsoft is supporting a technology called Model Context Protocol (MCP), an open-source protocol introduced by Anthropic, which is backed by Google. Scott believes MCP has the potential to create an “agentic web” similar to how hypertext protocols contributed to the proliferation of the internet in the 1990s.
“It means that your imagination gets to drive what the agentic web becomes, not just a handful of companies that happen to see some of these problems first,” Scott explained.
Scott also indicated that Microsoft is working to enhance the memory capabilities of AI agents regarding tasks users have requested, noting that currently, “most of what we’re building feels very transactional.”
However, improving an AI agent’s memory is costly due to the increased computing power required. Microsoft is concentrating on a novel approach called structured retrieval augmentation, where an agent extracts concise segments from each turn of a conversation with a user, thereby creating a roadmap of the discussed topics.
“This is a core part of how you train a biological brain – you don’t brute force everything in your head every time you need to solve a particular problem,” Scott concluded.
