Google's John Mueller said LLMs.txt, the file many publishers have rushed to adopt hoping to get their content surfaced by AI, was never built for that and won't deliver it. Mueller said on Google's Search Off the Record podcast that one of the standard's creators told him it was meant to help an AI that already knows a site find more of its pages, not to make a site discoverable. He also called the format inherently untrustworthy, since it lets a site owner simply assert what their content is about, something an LLM has no reason to believe, and stressed that discovery and ranking still run through a site's HTML. Mueller pointed to WebMCP as a better fit for e-commerce, since it gives AI agents concrete abilities like filtering products, comparing them, and adding items to a cart, though he said no agentic standard has won out yet.
