How to Get Cited by Perplexity and Google AI Overviews in 2026

To get cited by Perplexity and Google AI Overviews, make sure PerplexityBot and Googlebot can crawl and index you, answer the exact buyer question in the first paragraph of a page, and earn mentions on the third-party pages these engines already cite: Reddit, YouTube, LinkedIn, review sites and comparison articles. Then re-test the same questions monthly.

Both engines answer by searching the web, reading a handful of pages, and writing a summary with links. The links are the citations. If your page is not in the handful, you are not in the answer, and the work is to change which pages are in the handful.

This guide is built on a compiled dataset of 16 sources: platform documentation from Google and Perplexity, eleven published studies and surveys, one of which covers 680 million citations, and two of my own query studies. I have kept what each source measured, how big the sample was and when it was run, because the single biggest mistake in this field is quoting a percentage without its denominator. After the dataset I reconcile the places where the studies disagree, then give my own analysis and a ten-step method. If you want the ChatGPT side, read how to get your brand mentioned by ChatGPT first; this piece does not repeat it.

#The short version in one table

Perplexity and AI Overviews are fed differently. Perplexity cites Reddit and YouTube for 42.4% of the top 50 domains' mention share, and 28.6% of its citations rank in Google's top 10. AI Overviews draw 37.9% of citations from the top 10 and 31.0% from beyond the top 100.

Dimension	Perplexity	Google AI Overviews
Index it searches	Its own crawler (PerplexityBot) plus real-time fetching	Google Search index, same quality systems
Eligibility rule (official)	Allow PerplexityBot; changes reflect within about 24 hours	Indexed and snippet-eligible; no extra requirement
Top cited domains	Reddit 21.6%, YouTube 20.8%, Wikipedia 6.3% of top-50 mention share (Ahrefs, 2026)	YouTube most cited and growing 34% in six months; Reddit 21.0% and YouTube 18.8% of top-10 share (Profound, 2025)
Overlap with Google top 10	28.6% of citations (Ahrefs, 2025)	37.9% of citations (Ahrefs, 2026); 54.5% by BrightEdge's looser definition (2025)
Age of cited pages	1,166 days since publication on average (Ahrefs, 2025)	1,432 days for the top 3 citations (Ahrefs, 2025)
Structured data needed	Not documented as required	Google: no special schema needed
Best first move	Be discussed on Reddit, YouTube, LinkedIn and listicles	Rank for the question and its sub-questions; add video

Every number in that table is sourced in the dataset below.

#Why a founder should care now

Citations decide who is on the shortlist, and the clicks that remain go to the cited brands.

• 51% of B2B software buyers start research in an AI chatbot more often than in Google, and 69% chose a different vendor than planned because a chatbot recommended it, in G2's March 2026 survey of 1,076 buyers (G2, 15 Apr 2026).
• 8% of Google visits with an AI summary led to a click on a traditional result, against 15% without one, across 68,879 searches by 900 US adults (Pew Research Center, 22 Jul 2025).
• 1% of visits to pages with an AI summary included a click on a link inside the summary (Pew Research Center, 22 Jul 2025).
• 61% drop in organic click-through rate on queries with an AI Overview, from 1.76% in June 2024 to 0.61% in September 2025, and 35% higher organic click-through for brands cited in the Overview than for brands not cited (0.70% against 0.52%), the second reported as a correlation (Seer Interactive, 4 Nov 2025).

If the answer replaces the click, the citation is the only part of your page the buyer sees. More context sits in AI search statistics 2026.

#What the platforms themselves say

I start with documentation because it is the only part of this topic that is not an estimate.

Perplexity

Perplexity's crawler documentation says PerplexityBot is "designed to surface and link websites in search results on Perplexity" and is "not used to crawl content for AI foundation models." It recommends allowing the bot in robots.txt and says changes "may take up to 24 hours" to reflect. A second agent, Perplexity-User, fetches pages when a user asks a question and "generally ignores robots.txt rules" (Perplexity Crawlers documentation, checked 30 Oct 2026).

Perplexity's help hub describes the pipeline in three steps: it interprets the question with AI, searches the web for "authoritative sources like articles, websites, and journals," and condenses findings into an answer with "numbered footnotes linking to the original sources." Quick Search summarises indexed material; Pro Search goes deeper and may ask follow-up questions (Perplexity, "How does Perplexity work?", checked 30 Oct 2026).

Two practical consequences. First, there is a clean on/off switch: if PerplexityBot is blocked you are not in its index. Second, Perplexity does not publish a ranking formula, so everything beyond access is inferred from studies.

Google AI Overviews and AI Mode

Google's Search Central page on AI features says a page needs to be indexed and eligible to show a snippet, and that "there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." It describes query fan-out as "issuing multiple related searches across subtopics and data sources" (Google Search Central, "AI features and your website", updated 10 Dec 2025).

Google's AI optimization guide of 15 May 2026, as quoted by Frase, adds four statements. You do not need "new machine readable files, AI text files, markup, or Markdown." There is "no requirement to break your content into tiny pieces," and "no ideal page length." Schema "isn't required for generative AI search" though it remains useful for rich results. And valuable, non-commodity content with a first-hand expert point of view "will likely influence your website's presence in generative AI search in the long run" (Google AI optimization guide, 15 May 2026, quotes read through Frase, 7 Jun 2026; I could not open Google's page directly in this session, so wording is as quoted there). The same coverage says the AI features are "rooted in our core Search ranking and quality systems" and that the guide does not cover ChatGPT, Perplexity or other engines.

Read together: Google gives you one rule (be indexable and snippet-eligible) and one principle (original, expert content). Everything else you have heard about AI Overview "hacks" is, by Google's own account, not required.

#The compiled dataset

The dataset below compiles 16 sources run between June 2025 and October 2026, covering from 50 queries to 680 million citations. Studies from Ahrefs, Profound, Semrush and BrightEdge use different engines, samples and denominators, so figures can be compared in direction but not added together.

#	Source	Date	Sample	Engine	Headline finding
1	Google Search Central, AI features	updated 10 Dec 2025	Documentation	AIO, AI Mode	Indexed plus snippet-eligible; no extra requirements
2	Google AI optimization guide	15 May 2026	Documentation	AIO, AI Mode	No special files, markup or chunking required
3	Perplexity Crawlers docs	checked 30 Oct 2026	Documentation	Perplexity	PerplexityBot surfaces sites; about 24 hours to reflect changes
4	Perplexity help hub	checked 30 Oct 2026	Documentation	Perplexity	Search, condense, numbered footnotes
5	Ahrefs, 50 most-cited sites in Perplexity	1 Jun 2026, updated 2 Sep 2026	3.1 million US queries	Perplexity	Reddit 21.6%, YouTube 20.8%, Wikipedia 6.3%
6	Ahrefs, AI Overview citations vs top 10	2 Mar 2026	863K keyword SERPs, 4M AIO URLs	AIO	37.9% top 10, 31.2% ranks 11 to 100, 31.0% beyond 100
7	Ahrefs, AI search overlap	11 Aug 2025	15,000 long-tail queries	ChatGPT, Perplexity, AIO and others	Perplexity 28.6% in Google top 10; overall 12%
8	Ahrefs, content freshness	28 Jul 2025	17 million citations, 7 platforms	Assistants and AIO	Assistants 25.7% fresher than organic; AIO top 3 not fresher
9	Ahrefs, brand visibility correlations	12 Dec 2025	75,000 brands	ChatGPT, AI Mode, AIO	Branded mentions 0.664 to 0.709; YouTube mentions about 0.737
10	Profound, AI platform citation patterns	5 Jun 2025	680 million citations, Aug 2024 to Jun 2025	ChatGPT, Perplexity, AIO	Reddit 46.7% of Perplexity's top 10; Reddit 21.0% and YouTube 18.8% of AIO's
11	Semrush, most-cited domains	10 Nov 2025	3 months, 230,000+ prompts tracked weekly for AI Mode	ChatGPT, AI Mode	LinkedIn in nearly 15% of AI Mode answers; Reddit swings 60% to 10% in ChatGPT
12	BrightEdge, rank overlap	18 Sep 2025	16 months of AIO tracking	AIO	54.5% overlap with organic; 71.0% in B2B tech
13	Seer Interactive, AIO CTR	4 Nov 2025	3,119 search terms	AIO	Cited brands 35% higher organic CTR (correlation)
14	Pew Research Center	22 Jul 2025	68,879 searches, 900 US adults	AIO	8% vs 15% click rate; 1% click inside summary
15	G2, Answer Economy survey	15 Apr 2026	1,076 B2B software buyers	All chatbots	51% start in chatbots; 69% switched vendor
16	Shruti Kaushik, 50-query study	17 Oct 2026	50 B2B buyer questions	Google AI Mode	YouTube 48%, LinkedIn 44%, Reddit 30%; Wikipedia 0%

Sources 1 to 4 are first-party documentation, 5 to 15 are third-party measurements, and 16 is mine. Several of the measuring firms sell visibility tracking, which is a reason to read their conclusions with care; I flag that in the methodology.

#Which domains get cited

Perplexity

Ahrefs' monthly ranking, based on over 3.1 million US queries across all topics, is the most current public list for Perplexity. Reddit holds 21.6% of mention share among the top 50 domains and YouTube 20.8%, together 42.4%. Wikipedia is a distant third at 6.3%, followed by Facebook at 4.5% and Amazon at 3.6%. Healthline, Good Housekeeping, the New York Times, TripAdvisor, Consumer Reports, WebMD, Forbes, TechRadar and NerdWallet each hold between 1.1% and 2.6%. LinkedIn is 24th at 0.9% and Trustpilot 26th at 0.8%. Ahrefs describes Reddit as the most concentrated top source of any assistant it examined (Ahrefs, updated 2 Sep 2026).

Two cautions from the same page. The ranking is one US snapshot across all topics, so a B2B software question may look different from a health question. And mention share is measured against the top 50 only, not against all citations.

Profound's earlier analysis of 680 million citations found Reddit supplied 46.7% of Perplexity's top 10 cited sources (Profound, 5 Jun 2025). The gap between 46.7% and 21.6% is mostly a denominator effect, covered in the disagreement section.

Google AI Overviews and AI Mode

Profound found Reddit supplied 21.0% of AI Overviews' top 10 cited sources and YouTube 18.8% (Profound, 5 Jun 2025). Ahrefs' 2026 study adds that YouTube is the most-cited domain in AI Overviews and grew 34% over six months, and that 18.2% of AI Overview-cited pages that do not rank in Google's top 100 are YouTube URLs, equal to 5.6% of all AI Overview-cited URLs in the dataset (Ahrefs, 2 Mar 2026).

For AI Mode, a sibling product, Semrush found LinkedIn cited in nearly 15% of responses across more than 230,000 prompts tracked weekly between July and October 2025 (Semrush, 10 Nov 2025).

My own B2B check

On 17 October 2026 I ran 50 B2B buyer questions through Google AI Mode: agencies and ghostwriters, B2B software, LinkedIn, Reddit and AI search. YouTube was cited in 24 of 50 answers (48%), LinkedIn in 22 (44%) and Reddit in 15 (30%). At least one of the three appeared in 41 answers (82%). Wikipedia appeared in none. Answers cited 7.14 distinct domains on average, 239 distinct domains appeared in total, and 200 of those (84%) appeared only once. Agency and software vendor sites supplied 276 of 357 citations (77.3%) (Who Does AI Cite for B2B Buyer Questions). The browser was in India, so a US run will differ.

The pattern across all of this: platforms where people talk (Reddit, YouTube, LinkedIn) form the top tier, and below them the field is long and flat. For B2B questions, a small agency or vendor page can win an individual citation because no single publisher dominates.

#Do you need to rank on Google first?

For AI Overviews, partly. For Perplexity, much less.

• 37.9% of AI Overview citations came from the top 10 organic blocks, 31.2% from positions 11 to 100 and 31.0% from beyond the top 100, across 863K keyword SERPs and 4 million AI Overview URLs. With standard blue links only, the split was 37.1%, 26.2% and 36.7% (Ahrefs, 2 Mar 2026).
• 54.5% of AI Overview citations came from pages that also rank organically in September 2025, up from 32.3% in May 2024, and 71.0% in B2B tech (BrightEdge, 18 Sep 2025).
• 28.6% of Perplexity's citations ranked in Google's top 10, the highest of the assistants Ahrefs tested in 15,000 long-tail queries, while only 12% of AI-cited URLs overall ranked in Google's top 10 for the original prompt (Ahrefs, 11 Aug 2025).

The mechanism Ahrefs proposes is query fan-out: the engine searches several variations of the question, so the cited page often ranks for a related query, not the one typed. Ahrefs marks this as a hypothesis, not a confirmed finding. Google's own description of fan-out as "multiple related searches across subtopics" fits it.

The practical reading: ranking in the top 10 for your core query gets you roughly a third of the way, ranking for sub-questions gets you more, and ranking nowhere on Google still leaves about a third of citations available, mostly through video and community content.

#Does freshness matter?

Freshness helps assistants more than it helps AI Overviews. In Ahrefs' study of 17 million citations, assistants cited content 25.7% fresher than organic Google results by publication date and 13.1% fresher by last update, while the top 3 AI Overview citations were slightly older than organic results.

Ahrefs measured the average age of cited content in days:

Platform	Days since published	Days since last updated
Google AI Overviews (top 3)	1,432	1,067
Organic Google results	1,416	1,047
Perplexity (in-text)	1,166	993
Gemini	1,118	831
Copilot	1,056	865
ChatGPT (in-text references)	1,023	865
ChatGPT (citations)	958	989

Source: Ahrefs, 28 Jul 2025, 17 million citations across 7 platforms. The article's text and table disagree by a few days on three figures (organic published age 1,432 vs 1,416, organic updated age 1,046 vs 1,047, and AI Overview updated age 1,047 vs 1,067); I used the table.

Look at the scale. Perplexity's average cited page was about 3.2 years old (1,166 days) and 2.7 years since its last update (993 days). That is not a recency-obsessed engine. Fresher than organic Google is a difference of roughly 250 days on publication date, not a requirement to publish this week.

Google's guidance on freshness is narrower. It asks local and ecommerce sites to keep Merchant Center feeds and Google Business Profile current, because operational data should be "kept fresh" (as quoted by Frase). It does not tell publishers to date-bump articles.

#What structure gets cited

This is where I can say the least with confidence, because no primary study isolates page structure.

What the sources do support:

1. Google says chunking is unnecessary. "No requirement to break your content into tiny pieces" and "no ideal page length" (Google guide, as quoted by Frase). Writing short sections is good for readers; it is not a ranking input Google acknowledges.
2. Schema is not required. Google says no special schema.org data is needed for AI features. Ahrefs' 75,000-brand study did not find schema among the leading correlates; branded web mentions did (0.664 for ChatGPT, 0.709 for AI Mode, 0.656 for AI Overviews, against 0.266 for Domain Rating for ChatGPT) (Ahrefs, 12 Dec 2025).
3. Question-led pages match fan-out. If the engine runs several related searches, a page that answers the main question and the two or three obvious follow-ups can be retrieved by more of them. This is my inference from Google's description of fan-out and Ahrefs' overlap result, not a measured effect.
4. Video is a distinct surface. YouTube is the top or second Perplexity source (20.8%), the top AI Overview source, and the strongest correlate of AI visibility in Ahrefs' brand study (about 0.737).

The honest summary: nobody has published a controlled test showing that a particular heading format or answer length raises citation odds. What is measured is where the engines look, not how a page inside that place should be formatted.

#What the data disagrees on

This section is the reason to read more than one study. Four conflicts matter.

1. How much of AI Overviews comes from the top 10?

BrightEdge says 54.5% of citations overlap with organic rankings (September 2025). Ahrefs says 37.9% come from the top 10 (March 2026), and its earlier study put the figure near 76% in July 2025 on 1.9 million citations. Three reasons reconcile most of the gap.

• Different definitions. BrightEdge counts a citation as overlapping if the page ranks organically for the query. Ahrefs counts only positions 1 to 10, and tests it two ways (all SERP blocks, and blue links only).
• Method changes. Ahrefs notes its parsing has improved since July 2025, so the fall from about 76% to 37.9% may not be like for like. It also notes AI Overviews moved to Gemini 3 in January 2026, a possible real change.
• Sector mix. BrightEdge's 71.0% is B2B tech specifically. Sector matters.

My reading: do not plan around a single number. Plan around "somewhere between a third and a half of citations come from the top 10, and the rest come from deeper pages, video and community content."

2. How concentrated is Reddit in Perplexity?

Profound: Reddit is 46.7% of Perplexity's top 10 sources (Aug 2024 to Jun 2025). Ahrefs: Reddit is 21.6% of top 50 mention share (2026). These measure different things. Profound's denominator is the top 10 domains only, so every share is inflated relative to Ahrefs' top 50. Ahrefs adds the data is a 2026 snapshot while Profound's window ended in June 2025. Both agree on the direction: Reddit is the largest single source for Perplexity, by a wide margin.

Semrush adds volatility: close to 60% of ChatGPT responses cited Reddit in early August 2025, falling to around 10% by mid-September (Semrush, 10 Nov 2025). That is ChatGPT, not Perplexity, but it warns that any snapshot can move several times over within weeks. See Why Reddit matters for B2B for the full Reddit series.

3. Does freshness matter?

Ahrefs finds assistants cite fresher content than organic Google, yet its own table shows AI Overview top 3 citations are slightly older than organic. Google's guide talks about "up-to-date web pages" in the grounding sense, meaning the index is current, not that newer pages win. The consistent reading is that freshness is a modest tilt for assistants (Perplexity at 1,166 days still has an average cited page over three years old) and not a lever for AI Overviews. The study also uses first-seen dates by Ahrefs' crawler, which can be later than true publication dates.

4. Is the click gone, or is the citation a way to keep it?

Pew shows clicks fall when a summary appears (8% against 15%). Seer shows brands cited in the Overview see 35% higher organic click-through than uncited brands (0.70% against 0.52%). Both can be true: the pool of clicks shrinks, and cited brands take a larger share of what remains. The Seer figure is correlation, because brands that get cited may already be stronger. Pew measured US adults on all queries, Seer measured 3,119 search terms from its clients. Do not combine them into one percentage.

Smaller conflicts worth knowing

• The overlap between ChatGPT and Google top 10 is about 8% (Ahrefs) while Perplexity is 28.6%. The engines behave differently, so a "rank in AI" strategy that treats them as one engine will underperform.
• Branded mentions correlate with AI visibility (0.664 to 0.709), but the Ahrefs authors note these are correlations, not proof of cause. Companies that are visible everywhere are mentioned and cited more for many reasons.

#Shruti's analysis

This section is my synthesis. It uses only the figures cited above; where I calculate, I show the arithmetic.

1. The citation tier list has three levels, and only the first is hard to enter. Level one is the platforms where people talk: Reddit and YouTube together hold 42.4% of Perplexity's top 50 mention share (21.6 plus 20.8), and in my 50-query AI Mode check YouTube, LinkedIn and Reddit between them appeared in 82% of answers. Level two is high-authority publishers and review sites (Wikipedia 6.3%, Forbes 1.2%, NerdWallet 1.1% in Perplexity). Level three is the long flat tail: in my study 200 of 239 cited domains (84%) appeared in exactly one answer. A small company is not going to out-rank Reddit, but it can be one of the 200. That is where the opening is.

2. For B2B, vendor and agency pages are a larger share of citations than the global studies suggest. In my AI Mode run, agency and software vendor sites supplied 276 of 357 citations (77.3%), while publishers and analyst firms supplied 3.1% and review directories 1.1%. Ahrefs' Perplexity list is dominated by consumer categories (health, autos, retail), and its authors say results vary by topic. So the Reddit-and-YouTube headline is right for the engine as a whole, and a B2B founder should still publish on their own domain, because their category is not the category in the headline studies. This is a conclusion from two samples of different engines, so I would test it on your own questions.

3. Rankings are a gate, not the prize. If between 37.9% (Ahrefs) and 54.5% (BrightEdge) of AI Overview citations come from organically ranking pages, then ranking is useful but cannot be the whole plan. A page ranking in the top 10 has roughly a third to a half of the available citation mass; the remainder is reached through fan-out queries, video, and community content. I would spend about as much effort on being the best answer to the sub-questions as on the head term.

4. The most reliable lever is the one nobody can fake: being named by other people. The strongest correlates in Ahrefs' 75,000-brand study are branded web mentions (0.664 to 0.709) and YouTube mentions (about 0.737), and Domain Rating at 0.266 for ChatGPT is much weaker. Correlation is not cause, but it agrees with the platform tier list above and with Google's own statement that original, first-hand content is what influences presence over the long run.

5. Do not build a strategy on Reddit share alone. Semrush saw ChatGPT's Reddit citations go from close to 60% to about 10% in six weeks. A source mix that moves like that is a reason to be present on several surfaces and to re-measure monthly. Reddit is the single largest source for Perplexity and still not a safe single bet.

6. Freshness is a maintenance task, not a content calendar. Perplexity's average cited page is 993 days since its last update. The gain from updating is real but modest (assistants 13.1% fresher than organic by last update), so I would refresh pages that already earn impressions or citations once or twice a year, with new data and dates, instead of publishing weekly.

7. What I would not do. Chase chunking, special schema for AI, or AI text files for Google. Google says none is needed. If you are curious whether an AI text file helps other systems, see llms.txt adoption in 2026.

I run this method on my own site and report the weekly numbers openly in the AI visibility case study. In my 17 October 2026 baseline, Google AI Mode named me for 1 of 12 queries and cited my site for none, Perplexity completed 7 of 12 queries before hitting its free limit and named me in none, so I am working from a low base, not from success.

#The 10-step method

Run this in order. Steps 1 to 3 take a day; the rest is monthly.

Step 1. Open the gates. Confirm robots.txt allows PerplexityBot and Googlebot, and that key pages are indexed and snippet-eligible. Perplexity says changes take up to 24 hours. The free AI search readiness checker tests this in seconds.

Step 2. Write 20 buyer questions. Use the phrasing buyers type: "best X for Y", "X vs Y", "how to choose X", "how much does X cost". Include 5 questions on each stage: category, comparison, pricing, implementation, alternatives.

Step 3. Run them in both engines and log everything. For each question in Perplexity and Google (AI Overviews or AI Mode), record whether you are named, which URLs are cited, the domain type (platform, publisher, vendor, agency, review site), and the competitors named. The example sheet in how to get your brand mentioned by ChatGPT works unchanged.

Step 4. Count your citation gaps. For each question, list the cited pages that do not mention you. Sort by domain. If Reddit threads, YouTube videos or listicles recur, those are your targets.

Step 5. Make your own page the best answer to the question. Put a direct 40 to 60 word answer under the heading, then the numbers, the method, and a table. This fits Google's stated preference for original, expert content and costs nothing.

Step 6. Cover the sub-questions. For each core question, write down the three follow-ups a buyer would ask and answer them on the page or on linked pages. This serves query fan-out.

Step 7. Show up on the level-one platforms. Answer real questions in the relevant Reddit communities under your own name and disclose affiliation; publish a short YouTube walkthrough on the same question; post the data on LinkedIn. The aim is useful pages that exist, not link drops. Reddit rules are covered in the Reddit B2B guide.

Step 8. Get onto the listicles and review pages already cited. From step 4, pitch the authors of cited comparison pages with a factual correction or a data point. Keep your company description identical across your site, LinkedIn, Crunchbase and review profiles.

Step 9. Refresh, do not republish. Update pages with new data and a visible date once or twice a year, starting with those already getting impressions.

Step 10. Re-test monthly and keep a history. Rerun the same questions on the same date each month. Report mention rate (questions where you are named), citation rate (questions where your URL is cited) and share of answer per engine. Expect swings; only trend lines over several months mean anything.

#Common mistakes

• Optimising only the homepage. In my 50-query study, vendor and agency pages supplied the bulk of citations, and they were blog and category pages, not homepages.
• Quoting a single percentage. Overlap with Google's top 10 ranges from 28.6% to 54.5% depending on engine and definition.
• Treating one tool's tracker as the truth. Ahrefs, Semrush, Profound and BrightEdge all sell visibility tooling. Their numbers are useful, and they are not neutral.
• Testing once. Semrush's ChatGPT Reddit share moved by about 50 points in six weeks.
• Ignoring personalisation. Logged-in Perplexity and ChatGPT sessions can recognise you. In my case study, a logged-in Perplexity session named my product Klype because of personalisation, so I excluded it. Test in a logged-out session where possible.
• Believing a markup or file trick will do it. Google says there is none.

#Limits of this evidence

• Most studies measure citations, not clicks or revenue. Only G2, Pew and Seer touch behaviour.
• Samples skew to the US, English and consumer topics. My 50-query run came from a browser in India, on B2B questions only.
• The Ahrefs and Profound datasets come from tracker panels of prompts, not from real user logs.
• Google's guide is read through a secondary quote, and the freshness study is from July 2025, before AI Overviews moved to Gemini 3 in January 2026.
• Correlation studies (Ahrefs' 75,000 brands, Seer's CTR) do not prove cause.

#Methodology

Source selection. I included a source if it met three tests: it states a sample size or documents an official behaviour, it has a visible publication date between June 2025 and October 2026, and it covers Perplexity, Google AI Overviews or AI Mode. I prefer platform documentation first, then studies from firms with direct access to citation data (Ahrefs Brand Radar, Profound, Semrush, BrightEdge), then independent behavioural research (Pew, Seer, G2). I excluded listicles, vendor guides without data, and any figure I could not trace to a page that states its sample. I did not use unsourced statistics.

Date range. Sources span 28 July 2025 (Ahrefs freshness) to 17 October 2026 (my own study), with documentation checked on 30 October 2026. Percentages from different studies are never added or averaged in this article; where I do arithmetic (for example 21.6 plus 20.8 equals 42.4), the inputs come from one study.

My own data. The 50-query study is documented in a separate article with the question list. It used Google AI Mode with a fresh query per question on 17 October 2026 from India. The 12-query baseline used a logged-in ChatGPT session and a free Perplexity account, and was incomplete for two engines, as reported there.

Conflicts of interest. Ahrefs, Semrush, Profound and BrightEdge sell AI visibility tracking, and this site also offers a visibility check. I describe measurements, flag correlations as correlations, and show denominators.

Limits. I could not open Google's AI optimization guide page in this session, so its wording is taken from a secondary quote and marked as such. No figure in this article is estimated or invented.

#Sources

• Google Search Central, "AI features and your website", updated 10 Dec 2025. https://developers.google.com/search/docs/appearance/ai-features
• Google Search Central, "AI optimization guide", 15 May 2026 (quotes via Frase). https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
• Frase, "AEO and GEO Are Still SEO: What Google's Official AI Guide Actually Says", 7 Jun 2026. https://frase.io/blog/google-ai-optimization-guide
• Perplexity, "Perplexity Crawlers", checked 30 Oct 2026. https://docs.perplexity.ai/docs/resources/perplexity-crawlers.md
• Perplexity, "How does Perplexity work?", checked 30 Oct 2026. https://hub-prod.perplexity.ai/hub/faq/how-does-perplexity-work
• Ahrefs, "The 50 Most-Cited Websites in Perplexity", 1 Jun 2026, updated 2 Sep 2026. https://ahrefs.com/blog/most-cited-domains-perplexity/
• Ahrefs, "Update: 38% of AI Overview Citations Pull From The Top 10", 2 Mar 2026. https://ahrefs.com/blog/ai-overview-citations-top-10
• Ahrefs, "Only 12% of AI Cited URLs Rank in Google's Top 10 for the Original Prompt", 11 Aug 2025. https://ahrefs.com/blog/ai-search-overlap/
• Ahrefs, "New Study: AI Assistants Prefer to Cite Fresher Content (17 Million Citations Analyzed)", 28 Jul 2025. https://ahrefs.com/blog/do-ai-assistants-prefer-to-cite-fresh-content/
• Ahrefs, "Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews (75k Brands Studied)", 12 Dec 2025. https://ahrefs.com/blog/ai-brand-visibility-correlations
• Profound, "AI Platform Citation Patterns", 5 Jun 2025. https://www.tryprofound.com/blog/ai-platform-citation-patterns
• Semrush, "The Most-Cited Domains in AI: A 3-Month Study", 10 Nov 2025. https://www.semrush.com/blog/most-cited-domains-ai/
• BrightEdge, "AI Overview Citations Now 54% from Organic Rankings", 18 Sep 2025. https://www.brightedge.com/resources/weekly-ai-search-insights/rank-overlap-after-16-months-of-aio
• Seer Interactive, "AIO Impact on Google CTR: September 2025 Update", 4 Nov 2025. https://www.seerinteractive.com/blog/aio-impact-on-google-ctr-september-2025-update
• Pew Research Center, "Google users are less likely to click on links when an AI summary appears in the results", 22 Jul 2025. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
• G2, "In the Answer Economy, Don't Win the Click, Win the Answer", 15 Apr 2026. https://company.g2.com/news/g2-research-the-answer-economy
• Shruti Kaushik, "Who Does AI Cite for B2B Buyer Questions? 50-Query Study", 17 Oct 2026. /writing/who-ai-cites-b2b-buyer-questions-2026
• Shruti Kaushik, "Can ChatGPT find a founder? A live AI visibility case study", 17 Oct 2026. /writing/ai-visibility-case-study-shruti-kaushik