The podcast discusses the emergence of AI agents as a new type of internet user capable of browsing and making purchases online, highlighting that current web infrastructure is not optimized for such interactions. To address this, Google and Microsoft have introduced WebMCP, an experimental framework that enables websites to provide structured, actionable interfaces for AI agents. Unlike older methods that required agents to parse entire web pages (DOM), WebMCP allows direct, efficient interaction by leveraging existing browser session states - eliminating the need for API keys or complex authentication - and enabling agents to perform tasks like searching, booking, or buying seamlessly.
WebMCP is positioned as a significant evolution beyond traditional SEO and even AI Experience Optimization (AEO), shifting focus from mere information retrieval to enabling task completion. It offers advantages over alternative agent interaction models such as headless APIs, UI-based automation, or vendor-locked in-app agents by supporting user-owned agents with cross-platform compatibility. The framework opens opportunities across industries, especially in e-commerce, regulated sectors, and internal business tools, where secure, efficient agent access is valuable. Use cases include helping users select compatible products (e.g., espresso machines or car parts), streamlining complex purchases, and improving customer support. Additionally, early adoption presents business opportunities such as WebMCP conversion services, agent-based testing ("mystery shopper" models), and reporting on technical implementation quality, with potential for new startup models and competitive advantage.