# Getting Cited by ChatGPT, Claude and Google AI Overviews | Nadya > What gets a site cited in AI answers, per the evidence: Google's AI features guidance, crawler access, llms.txt data, the GEO paper's limits and Ahrefs studies. Canonical: https://nadiaconstructionstudio.com/blog/ai-search-geo [Home](https://nadiaconstructionstudio.com/) / [Blog](https://nadiaconstructionstudio.com/blog) # Getting Cited by ChatGPT, Claude and Google AI Overviews By Nam Bok Rodriguez · October 16, 2026 · 5 min read What gets a page cited in AI answers is mostly what gets it found in search: it must be crawlable and indexable, it must answer a real question clearly, and other sites must talk about you. Google says AI Overviews need no special files or markup, llms.txt shows little measurable effect so far, and the best research on "generative engine optimization" points to sourced facts, statistics and quotations rather than tricks. Here is what the evidence actually supports. ## What does Google say about AI Overviews and AI Mode? Google's [AI features documentation](https://developers.google.com/search/docs/appearance/ai-features), last updated December 2025, is direct: - There are no additional requirements to appear in AI Overviews or AI Mode, and no special optimizations are necessary. - You do not need to create new machine readable files, AI text files or markup to appear in these features. - A page must be indexed and eligible to be shown in Google Search with a snippet. - Existing SEO fundamentals still apply: allow crawling, use internal links, provide a good page experience, keep structured data consistent with visible text, and write helpful, reliable, people first content. The same page lists the controls if you want less exposure: nosnippet, data-nosnippet, max-snippet and noindex. It points to Google-Extended for AI training and grounding in other Google systems. Blocking Google-Extended is not how you leave AI Overviews, which follow normal Search controls. ## Can ChatGPT and Claude reach your site? Each assistant uses its own crawlers, and blocking the wrong one removes you from its answers. - **ChatGPT.** OpenAI's [crawler documentation](https://developers.openai.com/api/docs/bots) says OAI-SearchBot is used to surface websites in ChatGPT's search features, and sites that opt out will not appear in ChatGPT search answers. GPTBot is a separate crawler for training, and ChatGPT-User fetches pages when a user asks. To be visible in ChatGPT search, allow OAI-SearchBot in robots.txt. - **Claude.** Anthropic's [help center](https://support.claude.com/en/articles/8896518-does-anthropic-crawl-the-web-what-is-claude-bot) describes ClaudeBot for training, Claude-User for fetching pages when a person asks Claude a question, and Claude-SearchBot for improving search results. Blocking Claude-User or Claude-SearchBot can reduce your visibility in Claude's answers. Firewall and bot protection rules can block these crawlers even when robots.txt allows them, so check both. ## Does llms.txt help you get cited? llms.txt is a [proposed format](https://llmstxt.org/) from September 2024: a Markdown file at your site root that summarizes your content for language models. The idea is reasonable. The evidence of effect is thin so far. - An [Ahrefs study of 137,210 domains](https://ahrefs.com/blog/llmstxt-study) found that 97% of llms.txt files received zero requests in May 2026. GPTBot was the most frequent named AI bot fetching them. - SE Ranking's analysis of about 300,000 domains, reported by [Search Engine Journal](https://www.searchenginejournal.com/llms-txt-shows-no-clear-effect-on-ai-citations-based-on-300k-domains/561542/), found no clear link between having the file and being cited in AI answers. - Google says it does not need AI text files for Search, as quoted above. A [roundup by Optimizely](https://www.optimizely.com/field-notes/articles/does-llms-txt-work/) also reports Google's Gary Illyes saying in July 2025 that Google does not support llms.txt. We publish an llms.txt on our own site because it takes minutes and is useful for developer tools and agents that do read it. We do not expect it to change citations, and we would not sell it as a strategy. ## What does the GEO research actually show? The most cited paper is [GEO: Generative Engine Optimization](https://arxiv.org/abs/2311.09735) by Pranjal Aggarwal and colleagues, including researchers at Princeton, accepted to KDD 2024. The authors report that their methods can boost visibility by up to 40% in generative engine responses. In [the full paper](https://arxiv.org/html/2311.09735), the best methods were adding quotations, adding statistics and citing sources. Keyword stuffing performed worse than doing nothing. The limits matter as much as the headline: - The main experiments used a simulated engine built on GPT 3.5 Turbo over the top five Google results, plus a 200 query subset on Perplexity. - The authors note that only two engines were tested and that methods may need revision as these systems evolve. - They did not measure effects on traditional search rankings. - Results varied by domain, and citing sources helped lower ranked pages far more than top ranked ones. Our reading: the paper supports writing that is specific, sourced and quotable. It does not support a formula, and today's assistants differ from what was tested. ## Does ranking in Google still matter for AI citations? Yes, but less directly than it did. Ahrefs found in July 2025 that about 76% of URLs cited in AI Overviews came from pages in the top 10. Its [March 2026 update](https://ahrefs.com/blog/ai-overview-citations-top-10/), covering 863,000 keyword results and 4 million AI Overview URLs, put that share at about 38%. The author suggests AI Overviews now draw more from pages that rank for related "fan out" queries. The practical point: cover a topic and its follow on questions well, not just one head keyword. ## Do brand mentions matter more than links? Ahrefs' [study of 75,000 brands](https://ahrefs.com/blog/ai-overview-brand-correlation/) found the strongest correlation with AI Overview visibility was branded web mentions, at 0.664, ahead of branded anchors and branded search volume. These are correlations, not proof of cause. Still, they fit common sense: models describe the brands that the web describes. Being named in reviews, comparisons, forums, podcasts and press is part of the work. ## What actually works, in practice? - **Be reachable.** Indexed in Google, eligible for snippets, and open to OAI-SearchBot, Claude-SearchBot and Claude-User. - **Answer first.** Put a direct two or three sentence answer at the top of the page, then the detail. - **Use question style headings** that match how people ask. - **Cite and quantify.** Specific numbers with sources and clear quotations are what the GEO paper found most useful, and they make your page worth citing. - **Publish something original.** Your own data, prices, methods or examples give an assistant a reason to cite you instead of a summary. - **Keep structured data consistent** with what is visible on the page. - **Earn mentions elsewhere.** Get named on the sites people and models trust in your field. - **Measure.** Track referrals from AI assistants in analytics and test your key questions in each assistant monthly. ## An AI search readiness checklist - Key pages are indexed and not blocked by noindex or nosnippet. - robots.txt and firewall rules allow Googlebot, OAI-SearchBot, Claude-SearchBot and Claude-User. - Each important page opens with a direct answer. - Headings are phrased as the questions buyers ask. - Claims carry numbers and links to sources. - At least one page offers original data or a method others can cite. - Structured data matches visible content. - Related follow on questions are covered on the same page or linked pages. - You have a plan for earning mentions on third party sites. - AI referrals are tracked, and key questions are tested monthly. - llms.txt, if you publish one, is treated as optional. If you want your site to be easy for both search engines and AI assistants to read and cite, without the snake oil, we can help. [Start a project](https://nadiaconstructionstudio.com/#start) and tell us which questions you want to be the answer to. ## Sources - [Google Search Central: AI features and your website](https://developers.google.com/search/docs/appearance/ai-features) - [OpenAI: Overview of OpenAI crawlers](https://developers.openai.com/api/docs/bots) - [Anthropic Help Center: Does Anthropic crawl the web?](https://support.claude.com/en/articles/8896518-does-anthropic-crawl-the-web-what-is-claude-bot) - [llms.txt proposal](https://llmstxt.org/) - [Ahrefs: llms.txt study](https://ahrefs.com/blog/llmstxt-study) - [Search Engine Journal: llms.txt shows no clear effect on AI citations](https://www.searchenginejournal.com/llms-txt-shows-no-clear-effect-on-ai-citations-based-on-300k-domains/561542/) - [Optimizely: Does llms.txt work?](https://www.optimizely.com/field-notes/articles/does-llms-txt-work/) - [Aggarwal et al.: GEO: Generative Engine Optimization (arXiv)](https://arxiv.org/abs/2311.09735) - [Ahrefs: AI Overview citations and top 10 rankings](https://ahrefs.com/blog/ai-overview-citations-top-10/) - [Ahrefs: AI Overview brand visibility factors (75K brands)](https://ahrefs.com/blog/ai-overview-brand-correlation/) Start a project ## Tell us what you want *built.* A few lines is enough. The person who reads it is the person who builds it. - A reply within one business day - A written scope with price and date before any work - Code, infrastructure and accounts in your name Prefer email? [nam@nadiaconstructionstudio.com](mailto:nam@nadiaconstructionstudio.com)