What Microsoft’s NLWeb means for the future of the internet

Microsoft‘s new NLWeb protocol marks a pivotal shift in how websites will interact with AI, offering enterprises a standardized way to make their digital content accessible to AI systems. Similar to how RSS transformed content syndication in the early 2000s, NLWeb creates a framework for websites to become AI-ready through natural language interfaces. This protocol represents a significant advancement in the evolution of the web, potentially changing how users and AI agents alike access and interact with online information.

The big picture: Microsoft has launched NLWeb, an open-source protocol designed to AI-enable websites by creating natural language interfaces that work for both human users and AI agents.

• The protocol was conceived by RV Guha, who previously helped create RSS, RDF, and schema.org, linking NLWeb to earlier web standardization efforts.
• NLWeb builds upon existing protocols rather than competing with them, functioning as a Model Control Protocol server while utilizing familiar structured data formats like RSS.

How it works: NLWeb transforms websites into AI applications through a four-step process that leverages existing data structures and enhances them with artificial intelligence.

• The system begins by using structured data already published by websites, including markup and RSS feeds.
• Special tools process this data into vector databases, enabling efficient semantic search capabilities.
• Large language models enhance the stored data with external knowledge and contextual understanding.
• The final output is a universal natural language interface serving both human users and AI agents.

Early adoption: Several major digital platforms have already implemented NLWeb, signaling growing interest in AI-enabling web content.

• Early adopters include Chicago Public Media, Allrecipes, Eventbrite, Hearst (Delish), O’Reilly Media, Tripadvisor and Shopify.
• Microsoft CTO Kevin Scott described NLWeb as “a little bit like HTML for the agentic web,” highlighting its potential for standardization.

Market implications: Experts suggest a measured approach to NLWeb implementation, with timeline expectations varying by industry.

• Constellation Research projects substantial adoption will take 2-3 years to materialize.
• Companies with specific use cases are positioned to benefit from early pilot programs.
• Highly regulated sectors like insurance, banking, and healthcare are advised to proceed with caution.

Why this matters: For enterprise AI leaders, NLWeb represents a significant opportunity to transform websites into AI-ready platforms, potentially becoming as essential to the AI era as RSS was to the early web.