For the past year, “add an llms.txt file” has been one of the most repeated pieces of advice in the AEO and GEO communities. So when Google updated its AI Search optimization guidance on Monday, June 15, 2026, to say plainly that it does not use llms.txt files at all, it landed like a small earthquake in the SEO world. If you have already published one — or were about to — here is exactly what Google said, what it means for your Google rankings and AI Overviews visibility, and whether the file is still worth keeping for ChatGPT, Perplexity, and Copilot.
What is llms.txt and why did everyone start adding it?
The llms.txt file is a proposed standard: a plain markdown file placed at the root of your domain (yoursite.com/llms.txt) that gives large language models a curated, machine-readable summary of your most important content. The idea borrowed its logic from robots.txt and XML sitemaps — give the machines a clean map instead of making them crawl and guess.
As Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO) exploded as disciplines in 2025 and 2026, llms.txt became a checklist item in nearly every “how to rank in AI” guide. The problem: the technical recommendation was packaged and sold long before any major platform actually confirmed it used the file. That gap is exactly what Google’s June 15 update addressed.
What did Google say about llms.txt on June 15, 2026?
On June 15, 2026, Google updated its AI Search optimization guide and Search Central documentation to clarify its position in unusually direct language. The key points:
- Google Search does not use llms.txt files — not for traditional rankings and not for its generative AI features.
- You do not need the file to appear in Google Search, “including its generative AI capabilities” like AI Overviews and AI Mode.
- llms.txt files neither help nor hurt your rankings. Google simply ignores them.
- It is “completely fine” to create and maintain llms.txt files for other services or systems that do use them — doing so “won’t harm (nor help) your visibility or rankings in Google Search.”
That last point matters. Rather than dismissing the file outright, Google struck a more balanced tone than in its earlier comments, acknowledging that there are AI search surfaces beyond Google’s own where site owners may legitimately want to optimize. The message is not “llms.txt is useless” — it is “llms.txt does nothing for Google, and that’s a different question from whether it helps elsewhere.”
Does llms.txt help your Google rankings or AI Overviews visibility?
No. This is now confirmed directly by Google, not inferred. Whether you are chasing a blue-link ranking, a spot inside an AI Overview, or a citation in AI Mode, the llms.txt file plays zero role. Google grounds its generative answers in the same content it already crawls and indexes — your actual pages, your structured data, and the authority signals pointing to them. A summary file sitting at your root is not part of that pipeline.
If you spent time hand-building an llms.txt file specifically to win Google AI visibility, that time did not move the needle. The good news: it also did no damage. You can leave the file in place or remove it without any Google ranking consequence either way.
Should you still create an llms.txt file for ChatGPT, Perplexity, and Copilot?
This is the more useful question, and Google’s wording deliberately leaves the door open. Google is one AI search surface; ChatGPT, Perplexity, and Microsoft Copilot are others, and they each ingest content differently. Some tools and crawlers in that ecosystem do read machine-readable context files, and pairing schema.org structured data with an llms.txt file can increase an engine’s extraction confidence on those surfaces.
So the honest 2026 answer is: keep an llms.txt file if you want to support non-Google AI tools that honor it, but set your expectations correctly. It is a low-cost, low-risk supplement — not a ranking lever, and not a substitute for the content and authority signals that actually drive citations. Do not let “we added llms.txt” become the box you check instead of doing real GEO work.
What actually improves your GEO and AEO visibility in 2026?
If llms.txt is not the answer, what is? The patterns that consistently earn AI citations have nothing to do with a single root file:
- Self-contained, quotable passages. Content structured in clear 200–400 word blocks that answer a specific question earns far more AI citations than long, meandering prose.
- Named-source statistics and quotes. Including concrete figures and attributed quotations can lift AI visibility by 30–40% versus generic copy, because answer engines favor verifiable, earned authority.
- Clean structured data. Valid schema.org markup helps every engine — Google included — understand what your page actually is.
- Third-party authority. AI engines disproportionately cite earned, off-site mentions, so digital PR and being referenced by trusted sources matters more than ever.
- Crawlability. Make sure the AI crawlers you want (such as Google-Extended, OAI-SearchBot, and PerplexityBot) can actually reach your content.
The bigger lesson from June 15 is about how to read GEO and AEO advice in a fast-moving market. When a tactic gets packaged and sold before any platform confirms it, treat it as a hypothesis, not a requirement. Google just reminded everyone that the fundamentals — quality content, structured data, and genuine authority — are still what AI search rewards. That is the same playbook our team builds around in our GEO work, and it has not changed because of one file.
Want a second opinion on whether your site is set up to be found and cited by AI? Rosh Media helps businesses optimize for Google AI Overviews, ChatGPT, Perplexity, and Copilot — without chasing every shiny new file that hits the SEO forums.
Comments are closed