Microsoft is pushing for greater collaboration among artificial intelligence systems from different companies and improving AI memory capabilities, Chief Technology Officer Kevin Scott said Sunday ahead of the company’s annual Build developer conference.
Speaking to reporters and analysts at Microsoft’s headquarters, Scott said the company is working to encourage the adoption of open standards across the tech industry. These standards would allow AI agents, systems designed to complete specific tasks independently, such as fixing software bugs, to interact seamlessly across platforms.
Microsoft is supporting the Model Context Protocol (MCP), an open-source initiative introduced by Anthropic, a company backed by Google. Scott said MCP could play a role in creating an “agentic web,” drawing a parallel to the impact of hypertext protocols that facilitated the spread 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 said.
Microsoft is also working to enhance the memory functions of AI agents. Scott noted that current systems tend to be transactional, with limited ability to remember past interactions.
Improving memory, however, is costly due to the increased computing power required.
To address this, Microsoft is exploring a method called structured retrieval augmentation. This approach allows an AI agent to extract and store key pieces of information from each conversation turn, building a roadmap of the interaction.
“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 said.
The Build conference opens Monday in Seattle, where Microsoft is expected to unveil new tools for developers building AI systems.
