John Mueller, Google’s Senior Search Analyst, has poured cold water on one of 2026’s most talked-about SEO trends.
According to him, a small text file called llms.txt that many website owners have started adding to their sites, hoping it will convince AI chatbots like ChatGPT or Gemini to recommend their website over competitors.
Speaking on Google’s official Search Off the Record podcast, Mueller explained why this simply does not work the way people hope.
Think of llms.txt as a website owner writing their own reference letter and handing it directly to an AI system, saying “trust me, I’m the best, send your users to my site.” The problem is obvious once you think about it: every website can write the exact same glowing letter about itself. If everyone claims to be the best, an AI chatbot has no real way to tell which claim is actually true.
In his own words:
“It’s basically you’re telling these systems, like, I have the best website ever. And here are all of the pages that everyone must go to. And you must buy all of my products or whatever you put in there. So in LLM system, it basically, by design, can’t trust what is here as a way of differentiating between different websites.”
That said, Mueller did find one place where llms.txt might still help. He compared it to a shop directory inside a building, useless for getting customers through the front door, but genuinely helpful once someone has already walked in and needs directions to a specific counter. In the same way, if an AI assistant has already landed on a specific website to complete a task, like buying a product, the llms.txt file might give it useful instructions for finishing that task.
This is not the first time something similar has happened in search engine history. Decades ago, websites stuffed a different hidden file, called the meta keywords tag, with exaggerated claims to trick search engines into ranking them higher.
Search engines eventually stopped paying attention to it entirely, because it became meaningless once everyone gamed it. Mueller is warning that llms.txt is heading toward the exact same fate. Independent research already backs this up: a study of 300,000 websites found no real connection between having an llms.txt file and getting mentioned more often by AI tools.
For now, Mueller says the basics still matter most: well-organized web pages and clear internal links between pages remain what genuinely helps both search engines and AI systems understand and recommend a website.
