SE Ranking finds no link between llms.txt and AI citations
SE Ranking compared nearly 300,000 domains with how often AI answers cited them. Having an llms.txt made no difference, and leaving it out of the model made the model more accurate.
Last updated
What happened?
SE Ranking first published the study on 7 November 2025; it was written by Yulia Deda and reviewed by Svitlana Tomko. The article has since been republished with a new date, 15 October 2026.
How was it done?
The team took nearly 300,000 domains, of which 10.13% had an llms.txt, and compared them with how often each was cited in a sample of AI-generated answers. It used three methods: Spearman correlation, an XGBoost regression model and SHAP analysis. The article does not name the AI systems sampled or the time window.
What did it find?
No correlation between having an llms.txt and being cited. In SE Ranking’s words, “Removing this variable from our XGBoost model actually improved its accuracy”: the file behaved like noise.
What does it mean for your site?
There is no evidence that an llms.txt makes AI assistants cite you more. Pages that answer questions clearly and specifically still do that work.
Publish the file because it helps tools that read it describe you accurately, and measure your AI visibility by watching the answers themselves, not by whether the file exists.
Sources
Source: SE Ranking: llms.txt study
Coverage: Search Engine Journal, 20 November 2025