AI Crawlers Ignore Your llms.txt — What Gets a WooCommerce Store Cited

The 300,000-Domain Verdict

Large-scale research settles what speculation couldn’t — llms.txt doesn’t move AI citation rates.

SE Ranking analysed 300,000 domains using both traditional statistical methods and machine-learning models to measure whether llms.txt affected how often AI systems cited a site. The result was unambiguous: no statistically significant correlation between llms.txt adoption and AI citation frequency. In fact, their XGBoost model performed better when the llms.txt variable was removed entirely — the file introduced noise, not signal.

Adoption itself tells its own story. After roughly 18 months of conference panels, agency blog posts, and WordPress plugin integrations, only about 10.13% of domains in the dataset had implemented llms.txt. That’s one in ten sites, a long way from the near-universal adoption of standards like robots.txt or sitemaps that AI crawlers actually rely on.

An SE Ranking study of 300,000 domains found no statistically significant correlation between llms.txt adoption and AI citation frequency — removing the variable actually improved model accuracy.

A separate controlled study by Semrush reached the same conclusion: no statistical correlation between implementing llms.txt and improved AI-result performance. Two independent large-scale studies, same verdict. The question for WooCommerce operators isn’t whether to implement llms.txt. It’s why they’d prioritise it over changes that move measurable outcomes.

You may be interested in: Shopify Got UCP for Free — WooCommerce Stores Must Build It

AI Crawlers Skip the File

500 million bot events tell a definitive story about what AI systems actually fetch.

Limy monitored over 500 million AI bot traffic events across the brands it tracks during a 90-day window in early 2026. The finding cuts through every theoretical argument for llms.txt: only 408 of those requests targeted /llms.txt directly. GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, and Google-Extended overwhelmingly skip the file and crawl HTML pages instead.

That’s not a sampling artifact. That’s 500 million events showing that the major AI crawlers don’t treat llms.txt as a meaningful navigation document for their answer pipelines. They read your HTML, parse your structured data, and follow your internal links — the same content architecture that traditional search engines have rewarded for decades.

A separate analysis of 137,000 websites found that 28% had published an llms.txt file, and 97% of those files were never requested by any bot at all. The infrastructure exists to serve the file. The crawlers aren’t asking for it.

Across 500 million monitored AI bot events in 90 days, only 408 requests targeted /llms.txt directly — GPTBot, ClaudeBot, and PerplexityBot overwhelmingly crawl HTML instead.

For a WooCommerce store owner evaluating where to spend their next hour of technical effort, the data is clear. The AI systems that decide whether to cite your content don’t read your llms.txt. They read your product pages, your FAQ sections, your schema markup, and your answer-first content blocks.

Google Said No, on the Record

Google’s senior search team hasn’t just ignored llms.txt — they’ve explicitly compared it to a discredited standard.

Google’s Gary Illyes confirmed at Search Central Live in July 2025 that Google does not support llms.txt and is not planning to. That’s not ambiguity. That’s a senior member of Google’s search relations team giving a direct, public answer.

John Mueller went further on Bluesky, drawing a comparison that should settle the discussion for any SEO-literate operator: “To me, it’s comparable to the keywords meta tag.” The keywords meta tag is the canonical example of a self-declared signal that search engines abandoned precisely because anyone could claim anything in it. The parallel is deliberate and pointed.

Google’s own AI-features documentation, updated in May 2026, lists llms.txt among the tactics you do not need for AI Overviews or AI Mode. No major AI provider — OpenAI, Google, Anthropic, Meta, or Mistral — has publicly committed to using llms.txt in their production search or answer surfaces as of Q1 2026.

When Mueller was asked whether Google’s own behaviour counted as an endorsement of llms.txt, his response was characteristically direct: no. The file exists. The proposal is real. But the systems that decide which content surfaces in AI-generated answers don’t use it.

What High-Authority Sites Already Know

The sites winning AI citations share a common profile — and llms.txt isn’t part of it.

Here’s the thing: the sites that appear most frequently in ChatGPT, Perplexity, and Claude responses aren’t winning because of a text file at their domain root. They’re winning because of the same foundations that built authority in traditional search.

Genuine authority on a specific topic. Not broad, thin content across dozens of verticals, but deep, consistent coverage that AI models can recognise as a reliable source on a defined subject. For a WooCommerce store, that means owning your product category’s information architecture.

Consistent mentions across high-quality external sources also matter — editorial citations from industry publications, product reviews from authoritative sites, and brand signals that appear across multiple trusted domains. AI models weight external validation heavily when selecting which sources to cite.

Factor	Impact on AI Citations	Evidence Level
llms.txt file present	No measurable correlation	300,000-domain study + 500M bot events
Structured schema markup	Strong positive correlation	Multiple cross-platform citation analyses
Answer-first content	Strong positive correlation	AEO/GEO research consensus
External authority signals	Primary ranking factor	Consistent across all AI citation studies
Site speed and technical SEO	Indirect positive effect	Traditional SEO foundations carry into AI

SE Ranking’s own data revealed another telling pattern. High-traffic sites (100,001+ visits) showed lower llms.txt adoption at 8.27% compared to mid-traffic sites at 10.54%. The sites with the most resources and the most to gain from AI visibility chose not to prioritise the file. That’s not an oversight — it’s a signal about where experienced operators see real value.

The WooCommerce Citation Stack That Works

The specific technical and content infrastructure that drives AI visibility for WordPress stores.

If llms.txt doesn’t move AI citations, what does? The answer isn’t a single tactic — it’s a stack of interconnected systems that make your store’s content machine-readable, authoritative, and directly answerable.

Clean Product and FAQ schema markup sits at the foundation. AI systems parsing your WooCommerce product pages need structured data they can extract without ambiguity. That means complete schema.org Product JSON-LD with price, availability, reviews, and brand attributes — not the minimal markup most themes generate by default.

FAQ schema (FAQPage JSON-LD) gives AI models pre-structured question-answer pairs they can cite directly. When Perplexity or ChatGPT encounters a well-formed FAQ block on a page that answers the user’s query, the structured format reduces the extraction cost to near zero. Your content becomes citable by default, not by luck.

Answer-first content architecture completes the stack. Every piece of content that targets an AI-answerable question should deliver the direct answer in the first 100 words, supported by a verifiable statistic. AI citation systems consistently favour content that front-loads the answer over content that builds up to a conclusion. Translation: the opening paragraph is your citation pitch.

You may be interested in: Meta’s One-Click CAPI Sets a Low EMQ Baseline for WooCommerce

Server-side data infrastructure strengthens this stack indirectly but meaningfully. Accurate conversion tracking feeds better signals to ad platforms, which in turn strengthens your domain’s commercial authority signals — the kind of signals AI models weight when deciding which e-commerce source to cite for product-related queries.

Where llms.txt Has a Narrow Use

The file isn’t useless everywhere — but its valid use case is far from AI search citations.

Limy’s own analysis frames it well: llms.txt is not the SEO play it’s being sold as. It’s a Business-to-Agent (B2A) play — a standardised way for a brand to publish a machine-readable surface that AI agents (not search crawlers) can route on.

The demonstrable use case is developer documentation. IDE agents like Cursor, GitHub Copilot, and Claude Code actively read llms.txt when a developer points them at a domain’s documentation. In that context, the file serves as a curated table of contents that reduces token waste and improves answer accuracy for code-related queries.

For a WooCommerce store, this use case barely applies. Your customers aren’t pointing IDE agents at your product catalogue. They’re asking ChatGPT “which WooCommerce hosting plan should I choose” or telling Perplexity “compare these two WordPress themes.” Those queries are answered by AI systems crawling your HTML, not reading a markdown index file.

If your WordPress SEO plugin generates llms.txt automatically at zero effort — leave it on. There’s no downside to having the file. But don’t confuse zero-cost maintenance with zero-cost opportunity. Every hour spent optimising an llms.txt file is an hour not spent on the structured content and schema work that the data shows actually moves AI citations.

Key Takeaways

• No citation correlation: A 300,000-domain study and a 500-million-event crawler analysis both confirm llms.txt has no measurable impact on AI citation frequency for websites.
• Google explicitly rejected it: Google’s senior search team compared llms.txt to the discredited keywords meta tag and confirmed it is not a supported signal for AI Overviews or AI Mode.
• AI crawlers read HTML, not llms.txt: Of 500 million AI bot events monitored over 90 days, only 408 targeted /llms.txt — the rest crawled standard HTML pages.
• Structured content drives citations: Clean Product and FAQ schema, answer-first content, and external authority signals are the consistent factors in AI citation frequency.
• Narrow valid use exists: llms.txt serves IDE agents reading developer documentation — a real but limited use case that doesn’t apply to most WooCommerce stores.

References

• SE Ranking. “LLMs.txt: Why Brands Rely On It and Why It Doesn’t Work.” 2025. seranking.com
• Limy. “LLMs.txt in 2026: The Full Guide.” May 2026. limy.ai
• Search Engine Journal. “Google’s Mueller Says llms.txt Can’t Help LLMs Differentiate Sites.” June 2026. searchenginejournal.com
• DerivateX Agency. “LLMs.txt: The Complete Guide for SEO and AI Search (2026).” June 2026. derivatex.agency
• WolfPack Advising. “What Is llms.txt? Does It Help SEO in 2026?” June 2026. wolfpackadvising.com
• LinkBuildingHQ. “Should Websites Implement llms.txt in 2026?” February 2026. linkbuildinghq.com

If your WooCommerce store’s AI visibility strategy starts with a text file and ends with hope, the data says you’re optimising the wrong layer. Talk to Seresa about building the content and data infrastructure that AI systems actually read.