Why Does a Page Rank in Google but Still Get Ignored by AI?

The short diagnostic: check these causes in order

The order below runs from cheapest to check to most expensive to fix. It is based on dependency and effort, not on proven impact.

Ranking helps, but it isn't the same as being cited

Rank is a strong input to AI citation, not a guarantee of it. Both parts of that sentence are supported by data, and marketers usually hear only one of them.

The case for rank comes from seoClarity. In a snapshot of 362,000 US desktop queries that triggered an AI Overview, 94% of those Overviews cited at least one top-20 URL. Citation odds fell steadily with position: 43% at position 1, 37% at position 2, 31% at position 3, and 7% at position 20 (seoClarity).

The case against treating rank as enough comes from the same dataset and from others. In seoClarity's data, 44% of AI Overview citations came from beyond the top 20. Ahrefs looked at 863,000 keywords and found that only 38% of cited pages ranked in the top 10 for the same query. The rest were split almost evenly between positions 11–100 (31.2%) and pages beyond position 100 (31.0%) (Ahrefs, as reported by Search Engine Journal).

The published overlap numbers disagree because each study measures something different:

Sources: Search Engine Journal, seoClarity.

Don't average these numbers. Ahrefs notes that its parsing improved between versions, so part of the drop from 76% to 38% may come from better detection rather than a real change. The safe reading is that the link between ranking and citation is weaker than earlier measurements suggested. Even the most favorable numbers leave a #1 page uncited more often than not.

Why ranking doesn't carry over: AI answers search more than your query

Google's AI features look for sources across many related searches, not only the query you track. That is the main reason a ranking page can be skipped.

Google describes two mechanisms (Google Search Central):

• Grounding (retrieval-augmented generation): the system pulls pages from the index, then reads specific information inside them to build the answer.
• Query fan-out: the model generates several related queries at the same time to fetch more results across subtopics.

Google's own example is "how to fix a lawn that's full of weeds." That query fans out into searches like "best herbicides for lawns," "remove weeds without chemicals," and "how to prevent weeds in lawn." A page ranking #1 for the original query competes with every page that answers those sub-questions well. Google says this approach is meant to show "a wider and more diverse set of helpful links" (Google Search Central).

Google doesn't publish how many fan-out queries it runs or how it weights them. Research on generative engines describes a similar pipeline. The engine rewrites the query into simpler sub-queries, retrieves sources, summarizes each one, then writes an answer with inline citations (Aggarwal et al.). A ranking page can drop out at any of those stages. It may not match a sub-query, or it may be retrieved and then contribute nothing the summary needs.

Cause 1: No AI answer appears for the query

Some ranking pages get no AI citation simply because no AI answer is shown. Google says AI Overviews appear only when its systems judge them additive to classic Search, "and as such, often don't trigger." It also says AI Mode and AI Overviews may use different models and techniques, so the links they show will vary (Google Search Central).

Confirm it: search the exact query and a few close variants. Note whether an AI Overview appears and check AI Mode separately. Don't treat a result on one surface as a prediction for the other.

Fix it: if the head term rarely triggers an AI answer, stop tracking it as an AI visibility target. Put your effort into the related questions where AI answers do appear.

Cause 2: The page has been switched off for AI features

A page can rank and still be excluded from AI answers because of a setting that someone else configured. This is the check most teams skip.

For Google, the eligibility bar is short. A page must be indexed and eligible to appear in Search with a snippet, and Google says "there are no additional technical requirements" (Google Search Central). That makes snippet controls the first place to look.

Audit these, in this order:

1. Snippet controls. nosnippet, data-nosnippet, and max-snippet limit what Google can show from a page. A max-snippet value set years ago for a different reason can still restrict what AI features show today.
2. noindex. Check both the meta robots tag and the X-Robots-Tag HTTP header.
3. Search Console's Search generative AI control. This site-level setting, rolled out worldwide as of August 31, 2026, excludes a site's links and content from AI Overviews, AI Mode, and Discover AI features, usually within one to two days. Child properties inherit the setting from their parent property. If an agency or another team excluded the domain property, your subfolder property is excluded too (Search Console Help).

One common mistake points teams at the wrong control. Blocking Google-Extended does not remove you from AI Overviews. Google says Google-Extended governs AI training and grounding in other Google systems, not inclusion in Search AI features (Search Console Help). Google also says the generative AI control is not a ranking signal for the rest of Search, so excluding a site won't change its organic rankings.

Cause 3: Non-Google AI crawlers can't reach the page

Google's documentation says nothing about how ChatGPT, Perplexity, Claude, or Copilot choose sources. Treat access for those engines as something to verify, not assume.

Confirm it: look at server-side logs, not browser analytics. Crawlers generally don't run the JavaScript that analytics tools rely on, so a page can be crawled regularly without any trace in your dashboard. Search the logs for user agents such as OAI-SearchBot, GPTBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, and PerplexityBot. These crawlers serve different purposes, so check each vendor's documentation before you allow some and block others in robots.txt.

Keep two limits in mind. A matching user agent is an observation, not an authenticated identity. A bot request also doesn't show that the page was indexed or cited (Tideflow AI). Seeing no requests at all is the useful signal. It tells you the engine may never have fetched the page.

Fix it: remove unintended blocks in robots.txt, your CDN, or firewall bot rules. Then confirm in the logs that requests start arriving.

Cause 4: You rank for the head term, not the sub-questions

If AI answers are assembled from fan-out searches, a page that answers only the headline question gives the system one reason to cite it. Competitors that answer the follow-up questions get several.

Confirm it: write down the five to ten questions a buyer would ask next about the topic. Then check which pages rank for those questions and which pages the AI answer cites.

Fix it: cover those sub-questions inside your strongest existing pages, or in a small number of genuinely distinct pages. Don't create a separate page for every variation. Google says creating separate content for each fan-out query to manipulate responses violates its scaled content abuse policy (Google Search Central).

Illustrative example: a page ranking for "project management software for agencies" might add sections on client permissions, pricing per seat compared with flat pricing, and migrating from spreadsheets. Each section is a real question buyers ask. You don't need three thin new pages.

Cause 5: The page restates what everyone else already says

AI answers draw on many sources at once, so they have little reason to cite a page that adds nothing new. This matters most for pages that rank on authority but read like every other result.

Google says its AI systems "take a look at a variety of sources, so it can be helpful to have a unique viewpoint that stands out." It contrasts a commodity page, "7 Tips for First-Time Homebuyers," with a non-commodity one, "Why We Waived the Inspection & Saved Money" (Google Search Central). This is Google's guidance, not a measured result.

Measured evidence points the same way. In the GEO benchmark of 10,000 queries, adding citations, quotations from relevant sources, and statistics raised a source's visibility in generated answers by up to 40%, and by up to 37% on Perplexity (Aggarwal et al.). Those are best-case figures from tests against 2023–24 engines, and results varied by domain. They still show that what a page contains, and not only where it ranks, affects whether it gets cited once retrieved.

Fix it: for each key section, ask what a summary could only get from your page. Good answers are a number from your own data, a named expert's quote, a specific outcome you observed, or a clear recommendation with the reasoning behind it. Put that material under a clear heading close to the question it answers.

Cause 6: Other formats and sources own the answer

A ranking text page can lose the citation to a video or a community thread that covers the same sub-question. YouTube is the most-cited domain in AI Overviews and accounts for 5.6% of all AI Overview citations. Among cited URLs that didn't rank in Google's top 100 for the keyword, 18.2% were YouTube videos. In a separate comparison, AI Mode leaned more heavily on Wikipedia and Quora (Search Engine Journal). The mix varies by category.

Confirm it: list the domains cited in the AI answers for your priority questions. If videos, forums, or review roundups dominate, another optimization of your own page won't change much.

Fix it: publish video versions of pages where the answer benefits from being shown. Earn genuine inclusion in the third-party sources that already get cited. Google warns that inauthentic mentions aren't as helpful as they seem and that its spam systems filter them (Google Search Central). For the tradeoffs of pursuing roundup placements, see whether third-party listicles are good GEO or just AI-era link building.

Skip these popular fixes for Google

Google's optimization guide rules out several quick fixes for its own AI features (Google Search Central):

• llms.txt or other AI-specific files. Google says Search doesn't use them.
• Special schema or markup for AI. It isn't required.
• Chunking or rewriting content "for AI." Google says its systems can surface the relevant part of a page that covers several topics.
• One page per fan-out query. This counts as scaled content abuse.

This guidance applies to Google Search only. It doesn't tell you what other engines use, but it's a good reason to be skeptical of any tactic sold as a shortcut.

Measure four signals separately

A single "cited or not" check will mislead you. Track these four signals as separate outcomes:

A brand can be mentioned without being cited, and cited without being mentioned. Citations also vary in how much of the answer they support and where they appear in it (Aggarwal et al.). Because answers are generated, check the same prompts repeatedly and across engines instead of trusting a single screenshot.

On the Google side, traffic from AI features is included in Search Console's Performance report under the "Web" search type (Search Console Help). Be patient after changes. Google says recrawling can take anywhere from several days to several months (Google Search Central). Compare results against a baseline taken before the change, not against last week's one-off check.

Where a connected workflow helps

Running this diagnostic by hand gets hard at step four. You need to see, across several engines and repeated checks, which competitors and which sources each answer relies on. Then you need to connect any content change back to crawler visits and conversions.

Tideflow AI is built for that loop. It monitors tracked prompts across ChatGPT, Gemini, Perplexity, Google AI Mode, and Google AI Overviews. For each answer, it reports mentions, citations, the competing brands, and the sources the answer draws on. It turns gaps into plans: improve an existing page, create new content, or pursue inclusion in sources AI already cites. It logs AI crawler visits server-side through a Cloudflare Worker or Express middleware (Tideflow AI) and attributes AI referrals and conversions.

Tideflow is in early access, and it is explicit that neither monitoring nor publishing guarantees a citation (Tideflow AI). If you're comparing monitoring tools, see Tideflow AI vs Profound and Tideflow AI vs Peec AI.

Frequently Asked Questions

Is optimizing for AI answers a separate discipline from SEO?

Mostly no, at least for Google. Google says its generative AI features are rooted in its core Search ranking and quality systems, so indexing, snippet eligibility, and content quality still come first (Google Search Central). What changes is what you measure and which questions you cover. The argument is laid out in Is AEO Really a New Discipline?

Does being cited in AI Overviews mean I'll be cited in AI Mode or ChatGPT?

No. Google says AI Overviews and AI Mode may use different models and techniques and show different links. Google's documentation doesn't describe how ChatGPT or Perplexity select sources at all. Check each surface separately.

What to do first

This week, pick three ranking pages you expect AI to cite. For each one, check whether an AI answer appears for its queries, then audit snippet controls, noindex, and the Search Console generative AI setting, including any parent property. Next, check your server logs for AI crawler requests. Only after those checks pass should you invest in sub-question coverage and original evidence. For a step-by-step approach to ChatGPT specifically, continue with How to Get Mentioned by ChatGPT.

Sources

• seoClarity, The Overlap Between AI Overviews and Organic Rankings (July 2025)
• Matt G. Southern, Google AI Overview Citations From Top-Ranking Pages Drop Sharply (March 2026)
• Google Search Central, Google's Guide to Optimizing for Generative AI Features on Google Search; AI Features and Your Website
• Aggarwal et al., GEO: Generative Engine Optimization (2024)
• Google Search Console Help, Search generative AI control
• Tideflow AI, Detect AI Bots; AI Visibility & Content Platform; Tideflow AI vs Profound: Which AI Visibility Platform Should You Choose?