Get a free AI visibility report
.png)
Your page holds position one for "best project management software for agencies." You have held it for a year. Then a prospect mentions they found your competitor through ChatGPT, so you type the same question into ChatGPT yourself. Three tools come up. Yours is not one of them. You check Google's AI Overview for the same search, and your page is not cited there either, even though it sits right underneath as the top blue link.
That gap between ranking and being cited is exactly what Cognizo measures for brands every day, and it is what this study puts numbers on. We pulled together eight studies published between February 2025 and September 2026, covering more than 168 million AI Overview citations, tens of thousands of AI Mode and assistant queries, and per-position citation rates. The short answer: a #1 ranking helps, but it gets you cited far less often than most teams assume, and how much it helps depends heavily on the engine. That is why AI citations need to be tracked as their own metric, prompt by prompt and engine by engine, instead of being treated as a side effect of rankings.
The practical takeaway: a rank report cannot tell you whether buyers see you in AI answers. You need to measure citations directly, which is what Cognizo's Answer Engine Insights does for every tracked prompt, engine by engine.
There is no single dataset that answers "does ranking #1 get me cited?" for every engine. Some studies measure the share of citations that come from top-ranking pages. Others measure the reverse: how often a page at a given position gets cited. Both matter, and they produce very different-looking numbers, so we keep them apart.
| Study | What it measures | Sample and date |
|---|---|---|
| Conductor, AIO vs organic rankings | Citation rate by organic position, share of citations outside the top 10 | 167.9M AI Overview citations, US, June to August 2026 |
| Ahrefs, AI Overview citations update | Ranking position of AI Overview-cited URLs | 863K SERPs, about 4M cited URLs, March 2026 |
| Moz, AI Mode citation study (coverage) | Share of AI Mode citations matching top-10 URLs | Nearly 40,000 US and UK queries, February 2026 |
| SE Ranking, AI Mode research | AI Mode overlap with organic results, AI Overviews and itself | 10,000 US keywords, June 2025 |
| seoClarity, AI Mode rankings overlap | AI Mode citation rate by position | 1,000 transactional queries, 12,011 citations, September 2025 |
| Ahrefs, AI search overlap | Overlap of ChatGPT, Perplexity, Gemini and Copilot citations with Google and Bing | About 15,000 long-tail prompts, July 2025 |
| Seer Interactive, SearchGPT and Bing | Match between ChatGPT search citations and Bing results | 500+ citations, February 2025 |
| Ahrefs, AI Overview brand visibility factors | Correlation of brand signals with AI Overview mentions | 75,000 brands |
Most of these studies come from companies that sell SEO or AI visibility software, Cognizo included. We link every figure to its source and flag where methods differ.
AI Overviews are where rankings matter most, which makes sense, since Google builds them from its own index. But "matter most" still leaves a lot of room.
Conductor matched every AI Overview citation in nine weeks of US searches against the organic results for the same query. The citation rate falls at every position:
| Organic position | Chance the page is cited in the AI Overview |
|---|---|
| 1 | 24.9% |
| 2 | 21.1% |
| 3 | 17.9% |
| 10 | 10.2% |
Source: Conductor, June 15 to August 15, 2026.
Two things are true at once here. A #1 ranking is the single best organic position to hold: position one supplied 8.5% of all AI Overview citations in Conductor's data, more than any other spot, followed by position two at 6.6% and position three at 5.3%. And yet three out of four times, the #1 page is left out of the AI Overview sitting directly above it.
Ahrefs looked at the same question from the other direction. Of roughly 4 million URLs cited in AI Overviews, 37.9% ranked in the top 10 for the query, 31.2% ranked between 11 and 100, and 31.0% did not rank in the top 100 at all. Counting organic blue links only, the shares were 37.1%, 26.2% and 36.7%. Conductor's figure for citations outside the top 10, 57%, lands in the same range.
Ahrefs' July 2025 version of this study found that 76% of cited URLs ranked in the top 10. The authors caution that the two studies are not directly comparable, because their parsing now detects more citations and AI Overviews moved to Gemini 3 in January 2026. Either way, the current data says most AI Overview citations do not come from the first page of results. In practice, that means you can hold position one and still be missing from the answer most buyers read first. Cognizo shows that gap prompt by prompt: whether your page is cited in the AI Overview, and which pages are cited instead.
For the broader picture on how often AI Overviews appear and how much traffic they absorb, see our roundup of Google AI Overviews statistics and our study of what sources AI Overviews cite most.
AI Mode is Google's conversational search experience, and it draws even less from the top results than AI Overviews do.
AI Mode and AI Overviews also disagree with each other. SE Ranking found only 10.7% URL overlap between AI Mode and AI Overview citations for the same keywords. A page that wins one Google AI surface has no guarantee on the other. That is why Cognizo tracks AI Overviews and AI Mode as separate engines rather than one "Google AI" number. Our guide to tracking your brand in Google AI Mode covers how to monitor it separately.
Outside Google, the link between Google rankings and citations is weaker still. Ahrefs ran about 15,000 long-tail prompts through Google, Bing and four AI assistants, then checked where each cited URL ranked:
| Assistant | Citations ranking in Google's top 10 | Citations ranking in Bing's top 10 |
|---|---|---|
| Perplexity | 28.6% | Not reported separately |
| ChatGPT, in-text citations | 8.0% | 8.1% |
| ChatGPT, reference list | 6.1% | 8.1% |
| Average across the five assistants tested | About 11.9% | About 10.0% |
Source: Ahrefs, July 2025. Per-assistant figures for Gemini and Copilot are shown in the study's charts.
Perplexity behaves most like a search engine and shows the highest overlap. ChatGPT shows almost none: fewer than one in ten of its citations rank on Google's first page for the prompt that produced them. Ahrefs notes that Copilot had the highest overlap with Bing's top 10, which fits Microsoft's own documentation that Copilot grounds its answers with Bing searches.
Bing matters for ChatGPT too, but in a narrower way than often claimed. Seer Interactive found that more than 87% of SearchGPT citations matched Bing's top organic results when the same question was run in Bing. That study used about 500 citations, mostly from "best" and "top" questions, and predates ChatGPT search's general release, so treat it as a signal rather than a rule. It does suggest that for ChatGPT, Bing visibility is worth checking alongside Google.
Our explainers on how ChatGPT rank tracking works and Perplexity rank tracking go deeper on each engine, and our Copilot citation sources study shows what Copilot cites instead. Because each engine draws on different sources, Cognizo reports visibility and citations per engine, so a strong Perplexity result does not hide a gap in ChatGPT.
Four mechanisms explain most of the gap, and each one changes what a team should do about it.
Google has confirmed that its AI features use "query fan-out": the system splits a question into many related sub-queries, searches each one, and builds the answer from pages that show up across them. Ahrefs believes this is the main reason AI Overview citations drifted away from the original top 10. A page can rank first for the head term and never appear for the sub-questions the AI actually used. Moz attributes AI Mode's low overlap to the same behavior.
ChatGPT and Copilot lean on Bing, Perplexity runs its own retrieval, and Google's AI features read Google's index through fan-out queries. A #1 Google ranking says nothing about where the same page sits in Bing, or whether another engine's retriever even reached it. Our guide to how AI search engines find, rank and cite pages walks through each pipeline.
An AI answer needs a clear, quotable passage that answers one part of the question. A page can rank first because of its links, authority and overall relevance, and still lack the self-contained sentence or table the model wants to quote. Pages structured around direct answers tend to be easier to cite, which we cover in our guide to optimizing content for RAG and AI retrieval.
Rankings move over days and weeks. AI citations move between runs. SE Ranking ran the same 9,451 keywords through AI Mode three times on one day and found that only 9.2% of cited URLs appeared in all three answers; any two runs shared about 19% of URLs. A single manual check, whether it shows you in or out, tells you very little. This is the main reason Cognizo re-runs every tracked prompt daily: only a daily series shows whether a citation is a stable win or a one-off.
None of this means SEO stopped mattering. The data points to three real links between rankings and AI visibility.
Higher positions are cited more often. Every per-position dataset we found shows the same slope. In Conductor's AI Overview data, position one is cited about two and a half times as often as position ten. In seoClarity's AI Mode data, position one beats position three by nine points. Ranking well raises your odds, it just does not decide the outcome.
Ranking for the sub-questions matters as much as the head term. If fan-out is how AI engines gather sources, the pages that win are the ones that rank across the cluster of related questions, not only the main keyword. That favors sites with deep coverage of a topic over a single strong page.
Off-site signals predict AI visibility better than links. In Ahrefs' 75,000-brand study, branded web mentions correlated with AI Overview brand visibility at 0.664, branded anchors at 0.527, branded search volume at 0.392 and backlinks at 0.218. Ahrefs stresses that correlation is not causation, but the pattern is consistent with the rest of the data: AI engines reward brands that many sources talk about. Our research on third-party AI citations covers how to earn those mentions, and Cognizo's mention share metric tracks how often AI answers name your brand, separately from whether they cite your pages.
YouTube is a good example of where non-ranking citations come from. In Ahrefs' AI Overview data, YouTube made up 5.6% of all cited URLs and 18.2% of cited URLs that did not rank in the top 100 for the query.
For a side-by-side view of how the two disciplines differ, see our comparison of GEO and traditional SEO.
The data supports a clear set of moves. None of them involves abandoning SEO.
1. Stop using rankings as a proxy for AI visibility. If a #1 page is cited one time in four in AI Overviews and far less often elsewhere, a rank report cannot tell you whether buyers see you in AI answers. Track citations directly, per engine. Cognizo's Answer Engine Insights records, for every tracked prompt, whether your brand is mentioned, which of your pages are cited and which competitor pages appear instead, across the major AI engines.
2. Check your top-ranking pages against the AI answer for the same query. Start with the ten keywords where you rank first and that carry the most buyer intent. For each one, look at the AI Overview, AI Mode and ChatGPT answer. Where you rank but are not cited, you have the clearest gap to fix: the page already has authority, it is missing what the AI wants to quote.
3. Build for the fan-out, not just the keyword. List the sub-questions buyers ask around each head term (pricing, comparisons, use cases, "is it worth it", alternatives) and make sure you have a strong, direct answer to each. Cognizo's Prompt Volumes module shows the related prompts people ask AI, which is a practical starting point for that list.
4. Rewrite top pages so they are easy to quote. Put a direct answer in the first lines, use tables where readers compare options, add specific numbers with sources, and keep each section focused on one question. Cognizo's Content Optimization module drafts AI-optimized content from the gaps it finds, so a ranked-but-not-cited page can be refreshed quickly.
5. Do not ignore Bing. For ChatGPT and Copilot, Bing visibility affects whether a page is even in the candidate pool. Confirm your key pages are indexed in Bing Webmaster Tools.
6. Earn mentions on the sources AI engines already cite. Since off-site mentions correlate with AI visibility more strongly than links, find the domains that AI answers cite for your prompts, note which ones mention competitors but not you, and target them with PR and outreach. Cognizo's citation data and source classification show those domains per prompt and per engine.
7. Measure daily, not monthly. With fewer than one in ten AI Mode citations surviving three same-day runs, monthly spot checks mostly capture noise. Use a system like Cognizo that re-runs your prompts every day, so you can see whether a change in citations is a trend or a fluctuation.
8. Connect citations to traffic. AI answers send fewer clicks than blue links, so measure AI referral traffic and conversions separately. Cognizo's AI Traffic Analytics tracks AI referrals and crawler activity next to the visibility data, and our guide to tracking AI Overview click loss in Search Console helps quantify the traffic side.
Rank trackers answer "where does my page sit on Google?" That question still matters, but it no longer tells you whether buyers see your brand when they ask an AI. Cognizo is built to answer the second question.
| Question | What a rank tracker shows | What Cognizo shows |
|---|---|---|
| Is my page visible? | Position on Google for a keyword | Whether the brand is mentioned and the page is cited in each AI engine's answer |
| Who is beating me? | Pages ranking above mine | Competitor brands mentioned and competitor pages cited for the same prompt |
| Which sources matter? | Backlinks to my page | The domains each engine cites, classified by source and page type |
| Is it changing? | Rank history | Daily visibility, share of voice, citation share and mention share per engine |
| What should I fix? | Keyword gaps | Prompts where the brand is missing, with the pages and sources that won instead |
For each tracked prompt, Cognizo runs the query through the major AI engines every day and records which brands are mentioned, which domains and pages are cited, and how your visibility, share of voice, citation share and mention share change over time. Source classification and page type classification show what kind of page each engine prefers for the prompt, whether that is a review site, a listicle, a forum thread, a video or a vendor page. Regions, languages and competitors are unlimited on every plan, which matters because citations vary by market.
That makes the ranked-but-not-cited problem concrete. If your page ranks first on Google for a prompt but Cognizo shows a competitor's comparison page and a YouTube review being cited instead, you know exactly what kind of page to build and which sources to approach. When a gap appears, the content agent can draft the fix, technical site audits flag crawl issues that keep AI engines away from a page, and the results flow back into the same daily tracking. Teams can also pull the data into their own tools through Cognizo's MCP server and API. Our explainer on the AI visibility score covers how the headline metric is built. For a real example of what closing AI visibility gaps can lead to, read how Hat Club turned AI visibility into 20x revenue with Cognizo.
The fastest wins usually sit on pages you already rank with. They have authority and links; they are missing what AI engines quote. Here is a repeatable way to find and fix them, using Cognizo for the AI side.
Step 1: Pick the prompts that matter. Take the 20 to 50 questions closest to a purchase in your category, the ones where your top-ranking pages compete. Phrase them the way a buyer would ask an AI, not as short keywords. "What is the best project management tool for a 20-person agency?" is closer to real AI usage than "project management software." Add them to Cognizo as tracked prompts, with your main competitors.
Step 2: Line up rankings against citations. For each prompt, note where your best page ranks on Google, then check in Cognizo whether that page is cited in AI Overviews, AI Mode, ChatGPT and Perplexity. Sort the list into three groups: ranked and cited, ranked but not cited, and neither. The middle group is your priority list.
Step 3: Read what won instead. For each ranked-but-not-cited prompt, look at the pages the engines cited. Cognizo's page type classification shows whether they are listicles, review sites, forum threads, videos, documentation or vendor pages. If every engine cites comparison listicles and your ranking page is a product page, the fix is a different kind of page, not a better version of the same one.
Step 4: Check the sub-questions. Look at the related prompts in Cognizo's Prompt Volumes for the same topic. If competitors are cited across pricing, alternatives and use-case questions while you only cover the head term, add those answers to the page or build linked pages for them. This is how you get into the fan-out results that AI Overviews and AI Mode draw from.
Step 5: Make the page quotable. Add a two-sentence answer at the top, a comparison table where readers compare, current numbers with sources, and clear headings that each answer one question. Cognizo's content agent can draft the refreshed version from the gap, and its technical site audit confirms the page is reachable and readable for AI crawlers.
Step 6: Go after the sources. List the third-party domains cited for the prompt that mention competitors but not you. Those are your outreach, review and PR targets, and Cognizo's citation data and outreach targets turn them into a working list.
Step 7: Watch the result daily. Re-check citations every day for the following weeks. Because AI answers change between runs, look for a sustained rise in citation share rather than a single appearance. Cognizo's daily tracking shows that trend per engine, so you can tell a real gain from a lucky run.
Starting with ranked-but-not-cited pages is usually faster than building new rankings from scratch, because those pages already have the authority. What they lack is the part an AI engine can quote, and that is a content fix, not a link-building project.
Sometimes. Conductor's analysis of 167.9 million AI Overview citations found that the page in position one was cited 24.9% of the time. That makes position one the single most-cited organic position, but it also means the top result is left out of about three AI Overviews in four.
About 38% in Ahrefs' March 2026 study of 863,000 SERPs, and 43% in Conductor's 2026 data. The rest come from pages ranking lower or not ranking for the query at all. Ahrefs found that 31% of cited URLs did not rank in the top 100.
Only a little. In Ahrefs' study of about 15,000 prompts, 8% of ChatGPT's in-text citations ranked in Google's top 10 for the same prompt. Seer Interactive found more than 87% of early ChatGPT search citations matched Bing's top results, though Ahrefs' larger study saw similar overlap with Bing and Google (about 8% each). Rankings on either engine get a page into consideration; they rarely decide the citation.
AI Mode splits your question into many sub-queries and pulls sources from all of them, so it often cites pages that rank for related questions rather than your exact keyword. Moz found only about 12% of AI Mode citations matched a top-10 URL for the same query. AI Mode answers also change between runs, so you may appear on one check and not the next.
Yes. Higher-ranking pages are cited more often in every dataset we found, and AI engines still need to find and crawl your pages. But SEO alone is not enough. Covering the related sub-questions, structuring pages so they are easy to quote, and earning mentions on third-party sites all affect AI citations in ways rankings do not capture. The practical setup is a rank tracker for SEO alongside an AI visibility platform like Cognizo for citations.
Google's AI Overviews, followed by Perplexity. About 38% of AI Overview citations rank in Google's top 10, and 28.6% of Perplexity's citations did in Ahrefs' assistant study. AI Mode and ChatGPT overlap the least, at around 12% and 8%.
Start with the pages where you rank highly but are not cited. Add direct answers in the first lines, tables and specific numbers, cover the related sub-questions on the page or in linked pages, make sure the page is indexed in Bing, and earn mentions on the sites the AI engines cite for that topic. Then track the result daily per engine, because AI answers change often. Cognizo combines those steps: it finds the prompts where you are missing, shows what was cited instead, drafts the content fix and tracks whether the citation follows.
A rank tracker will not show it. You need prompt-level AI visibility tracking that records the cited pages for each engine. Cognizo runs your prompts daily across the major AI engines and shows which of your pages, and which competitor pages, each engine cites, so you can compare your rankings with your actual AI presence.