Back
AI Search & Visibility
,

AI Search Statistics (October 2026): 125 AI Search, AEO and GEO Stats From AirOps Research

Saven Jude Sajen
,
•
October 29, 2026
Citation source statistics (October 2026)
Retrieval and fan-out statistics (October 2026)
Content structure statistics (October 2026)
Content freshness statistics (October 2026)
Offsite and community statistics (October 2026)
Brand visibility and consistency statistics (October 2026)
ChatGPT ads statistics (October 2026)
Content team and budget statistics (October 2026)
How we keep these statistics current
Changelog

This page collects 125 AI SEO and AI search statistics from 15 AirOps studies and 3 research posts by our Head of Research & Insights, as of October 2026. The studies measure how ChatGPT, Gemini, Perplexity, Google AI Mode and Google AI Overviews retrieve, cite and mention pages and brands, and include two surveys of marketing and content leaders. Each statistic gives its figure, sample, AI engines and date, and links to the study it comes from.

Methodology

125 statistics from 15 AirOps studies and 3 research posts by our Head of Research & Insights, as of October 2026.

  • Every figure comes from our own research, published between April 2025 and October 2026, and is shown as of October 2026.
  • Figures are copied exactly as each study or post publishes them, with no rounding and no figures combined across sources.
  • Samples, engines and dates are the ones each source reports. Where a source gives no field dates, we give its publication month.
  • Survey figures are self-reported by the respondents.
  • Figures from third parties that appear in our studies are left out.

Citation source statistics (October 2026)

Where AI answers take their citations from, by domain, platform and source type. Figures as of October 2026.

  • 84% of AI citations come from domains outside the top 100 cited domains, which together number 148K+ domains, in The Long Tail (March 2026).
  • Wikipedia pages earned a 59.2% citation rate when ChatGPT retrieved them, the highest of any site type across 353,799 pages in The Fan-Out Effect (April 2026).
  • Health publishers earned a 46.4% ChatGPT citation rate across 3,484 retrieved pages in The Fan-Out Effect (April 2026).
  • Reddit pages earned a 29.9% ChatGPT citation rate, and only 0.59% of Reddit pages were cited in all three runs, in The Fan-Out Effect (April 2026).
  • YouTube pages earned a 2.4% ChatGPT citation rate when retrieved, the lowest of any site type in The Fan-Out Effect (April 2026).
  • Amazon pages earned a 12.2% ChatGPT citation rate across 752 retrieved pages in The Fan-Out Effect (April 2026).
  • At high query match, government pages earned a 49.1% ChatGPT citation rate against 35.2% for non-government pages in The Fan-Out Effect (April 2026).
  • ChatGPT cited 6,371 page and query combinations without the page appearing in any search result for that query, and Reddit made up 11.1% of the pages cited from memory, in The Fan-Out Effect (April 2026).
  • Reddit ranks as the #1 most cited domain in Perplexity and Google AI Mode and #2 in ChatGPT across 5.5M+ answers in The Community Flywheel (November 2025).
  • LinkedIn ranks #2 in Google AI Mode, #3 in Perplexity and #5 in ChatGPT by appearances across 5.5M+ answers in The Community Flywheel (November 2025).
  • YouTube ranks as the #2 most cited source in Gemini and Perplexity and #3 in Google AI Mode across 5.5M+ answers in The Community Flywheel (November 2025).
  • Owned brand.com and product page citation share fell 10% from August 2025 to June 2026 across about 3.5 billion citations in our creator citation study.
  • Creator and social citation share grew 140% from August 2025 to June 2026 across about 3.5 billion AI citations in our creator citation study.
  • YouTube held 4.6% of all AI citations by June 2026 in our creator citation study, tracked monthly from August 2025.
  • Reddit's share of all AI citations peaked near 5.3% around the holidays and settled at 3.3% by June 2026 in our creator citation study.
  • Google AI Overviews pull about one in every 9.5 citations from creator surfaces and about one in 12 from YouTube alone, in a 93-million-citation snapshot across six engines in our creator citation study (August 2026).
  • 89% of LinkedIn citations go to published written content and 7% to personal profiles and company pages combined, in our creator citation study (August 2026).
  • About 99% of YouTube citations point to one specific video in our creator citation study (August 2026).
  • Creator sources make up about 7.73% of AI citations at the top of the funnel and 4.7% at the bottom, in our creator citation study (August 2026).
  • AI search cites LinkedIn about four times as often in B2B answers as in B2C answers, peaking at 7.1% of answers for marketing and agencies, according to our Head of Research & Insights (October 2026).
  • Articles and posts make up two-thirds or more of the LinkedIn pages cited in B2B AI answers, according to our Head of Research & Insights (October 2026).
  • In consumer finance, 39% of LinkedIn citations in AI answers are personal profiles, according to our Head of Research & Insights (October 2026).
  • The share of ChatGPT answers citing LinkedIn fell from about 1 in 22 to 1 in 175, according to our Head of Research & Insights (October 2026).
  • Fewer than 1% of the LinkedIn posts AI search cites are from the past week, and most are over six months old, according to our Head of Research & Insights (October 2026).

Retrieval and fan-out statistics (October 2026)

How ChatGPT searches, retrieves and selects pages before it cites them. Figures as of October 2026.

Content structure statistics (October 2026)

Which on-page signals separate cited pages from the rest. Figures as of October 2026.

  • Pages of 500 to 2,000 words earn the highest ChatGPT citation rates, and pages over 5,000 words were cited 28.6% of the time against 30.5% for pages under 500 words, in The Fan-Out Effect (April 2026).
  • Pages with JSON-LD schema were cited by ChatGPT 38.5% of the time against 32.0% without it, a 6.5 point advantage, in The Fan-Out Effect (April 2026).
  • The schema types with the highest ChatGPT citation rates are MedicalWebPage at 47.0%, BreadcrumbList at 46.2%, FAQPage at 45.6%, Organization at 44.3% and WebSite at 40.6%, in The Fan-Out Effect (April 2026).
  • Pages with both lists and tables were cited by ChatGPT 33.9% of the time against 31.0% for pages with neither, in The Fan-Out Effect (April 2026).
  • Product pages with both lists and tables were cited by ChatGPT 37.8% of the time against 24.4% with neither, a 13 point lift, in The Fan-Out Effect (April 2026).
  • Comparison pages with 3 tables earn 25.7% more ChatGPT citations across 7,500 commercial prompts in From Retrieved to Cited (April 2026).
  • Shortlist pages averaging 10 or fewer words per sentence earn 18.8% more ChatGPT citations in From Retrieved to Cited (April 2026).
  • Commercial pages with 7 to 26 list sections were 6% to 15.2% more likely to earn a ChatGPT citation across 217,508 retrieved pages in From Retrieved to Cited (April 2026).
  • Commercial pages averaging 11 to 14 words per sentence had a roughly 7% higher likelihood of a ChatGPT citation in From Retrieved to Cited (April 2026).
  • Early-discovery pages of 1,301 to 1,500 words have a 10.8% higher ChatGPT citation likelihood in From Retrieved to Cited (April 2026).
  • Shortlist pages with 10 images earn 16.4% more ChatGPT citations, and shortlist pages of 1,501 to 1,800 words earn 8.4% more, in From Retrieved to Cited (April 2026).
  • Comparison pages with 7 distinct price points earn 15.7% more ChatGPT citations in From Retrieved to Cited (April 2026).
  • Validation pages with 15+ price points have an 11.0% higher ChatGPT citation likelihood in From Retrieved to Cited (April 2026).
  • 61% of ChatGPT-cited pages use three or more schema types, against 25% of Google page one results, across 12,000+ URLs in Why Ranking on Page One Isn't Enough (July 2025).
  • FAQ schema appears on 10.5% of ChatGPT-cited pages and 5.4% of Google page one results in Why Ranking on Page One Isn't Enough (July 2025).
  • 68.7% of ChatGPT-cited pages follow a sequential heading structure, against 23.9% of Google's top results, in Why Ranking on Page One Isn't Enough (July 2025).
  • 87% of ChatGPT-cited pages use a single H1, against 64% of Google page one results, in Why Ranking on Page One Isn't Enough (July 2025).
  • ChatGPT-cited pages contain 13.75 distinct list sections on average, more than 17 times the Google page one average, in Why Ranking on Page One Isn't Enough (July 2025).
  • 72.9% of ChatGPT-cited pages include bullet points, against 74.5% of Google page one results, in Why Ranking on Page One Isn't Enough (July 2025).
  • Average sentence length is 17.32 words on ChatGPT-cited pages and 16.54 on Google page one results in Why Ranking on Page One Isn't Enough (July 2025).
  • Informational pages need over 60% cosine similarity between the query and on-page signals such as titles and slugs to be cited, across 300+ cited pages in From Query to Citation (August 2025).
  • 65.5% of cited informational pages had a URL slug that met the 60% similarity threshold in From Query to Citation (August 2025).
  • Titles and slugs of cited informational pages share 27% to 30% of their words with the query in From Query to Citation (August 2025).
  • Pages with at least 30% of the query's terms in the title or slug were twice as likely to earn citations in From Query to Citation (August 2025).
  • Roughly 55% of cited commercial titles and 62% of cited commercial slugs met the 60% similarity threshold in From Query to Citation (August 2025).
  • Titles and slugs of cited commercial pages share only 16% of their words with the query, and nearly 90% have under 30% word overlap, in From Query to Citation (August 2025).
  • Top-performing brands earned up to 4.8x more AI citations across 6,700+ pages from 50 SaaS brands in LLMs Are Picking Favorites (June 2025).
  • Pages with Article, FAQ or HowTo schema were 78% more likely to be cited, and clear heading hierarchies lifted citations 63%, in LLMs Are Picking Favorites (June 2025).
  • Long paragraphs of 3 or more sentences cut featured snippet capture odds by 59%, and concise answers with structured lists raised capture by 74%, in LLMs Are Picking Favorites (June 2025).
  • Pages with 2+ outbound citations per 500 words earned 2.4x more AI citations in LLMs Are Picking Favorites (June 2025).
  • Top-quartile brands for authority earned 40% more AI citations than bottom-quartile brands in LLMs Are Picking Favorites (June 2025).
  • Content with consistent brand voice, current product names and aligned messaging earned up to 41% more LLM citations in LLMs Are Picking Favorites (June 2025).

Content freshness statistics (October 2026)

How recently cited pages were updated, and how long content stays current. Figures as of October 2026.

  • More than 70% of pages cited by ChatGPT were updated within the past 12 months, across 4,000+ cited pages and 900 high-intent queries in The Silent Pipeline Killer (August 2025).
  • 35.2% of ChatGPT-cited pages were updated in the last three months in The Silent Pipeline Killer (August 2025).
  • 53.4% of ChatGPT-cited pages were refreshed within six months in The Silent Pipeline Killer (August 2025).
  • 1 in 5 ChatGPT citations for commercial queries went to content older than a year in The Silent Pipeline Killer (August 2025).
  • Nearly a third of content ChatGPT cited for informational queries was over a year old in The Silent Pipeline Killer (August 2025).
  • 50% of mixed-intent pages, those ChatGPT cited for both commercial and informational queries, were updated in the last three months in The Silent Pipeline Killer (August 2025).
  • 26% of cited pages published more than a year ago were updated in the last three months in The Silent Pipeline Killer (August 2025).
  • For commercial queries, pages not refreshed annually lose over 50% of their chance to be cited by ChatGPT in The Silent Pipeline Killer (August 2025).
  • In SaaS, finance and news, the freshness window narrows to as little as 3 months, against 12 months or more for education and research, in The Silent Pipeline Killer (August 2025).
  • Over 30% of cited healthcare and wellness content was refreshed in the past six months in The Silent Pipeline Killer (August 2025).
  • Pages 30 to 89 days old earned the highest ChatGPT citation rate at 32.8%, against 25.3% for pages under 30 days old, in The Fan-Out Effect (April 2026).
  • Finance pages 30 to 89 days old were cited by ChatGPT 50.2% of the time against 35.1% at 5+ years, in The Fan-Out Effect (April 2026).
  • Travel pages 30 to 89 days old were cited by ChatGPT 44.8% of the time against 26.2% at 5+ years, the largest age gap in The Fan-Out Effect (April 2026).
  • Among pages with a strong query match, pages under a year old hold a 4.2 point ChatGPT citation advantage over pages 1 to 5 years old, 35.4% against 31.2%, in The Fan-Out Effect (April 2026).
  • Pages with a visible last updated schema field earned 1.8x more AI citations across 6,700+ SaaS pages in LLMs Are Picking Favorites (June 2025).
  • The median freshness score across 50 SaaS brands was 56 in LLMs Are Picking Favorites (June 2025).

Offsite and community statistics (October 2026)

How third-party pages, communities and platforms shape which brands appear. Figures as of October 2026.

  • Perplexity references user-generated and community platforms in more than 90% of its answers, against as few as 7% for Gemini, across 5.5M+ answers in The Community Flywheel (November 2025).
  • 88% of Reddit citations in AI answers come from category queries and 12% from brand queries in The Community Flywheel (November 2025).
  • Among branded Reddit citations, 31% of queries check features, 29% ask how to do something and 25% look for factual details, in The Community Flywheel (November 2025).
  • 75% of YouTube citations in AI answers come from non-branded queries in The Community Flywheel (November 2025).
  • For non-branded queries that cite YouTube, 59% seek factual information and 34% focus on how to do a task, in The Community Flywheel (November 2025).
  • For branded queries that cite YouTube, 37% ask how to use the brand, 27% verify features and 25% seek factual details, in The Community Flywheel (November 2025).
  • Nearly 90% of third-party brand mentions in AI answers came from listicles, comparisons and reviews, across 21,311 mentions in GPT-5, Claude Sonnet 4.5 and Perplexity Sonar in our offsite signals study (October 2025).
  • 80% of brands mentioned through third-party content often appeared among the first three companies on the cited page in our offsite signals study (October 2025).
  • Product pages account for 19.3% of first-party brand mentions in AI answers and homepages 7.1%, in our offsite signals study (October 2025).
  • Within each AI model, 37% to 52% of the brands it mentioned appeared in no other model, in our offsite signals study (October 2025).
  • First-party pages made up 13% to 21% of brand mentions in Claude and Perplexity answers, against 4% to 11% in GPT-5, in our offsite signals study (October 2025).
  • About 2% of brand mentions in AI answers came with no citation, in our offsite signals study (October 2025).
  • In answers we tracked in October 2026, forums and threads were the largest community citation source on Google AI Mode at 3.7% and AI Overviews at 3.3%, against near zero on ChatGPT, according to our Head of Research & Insights.

Brand visibility and consistency statistics (October 2026)

How often brands stay visible from one AI answer to the next. Figures as of October 2026.

  • About 30% of brands stay visible from one AI answer to the next, across 800 queries and 45,000+ citations in Staying Seen in AI Search (September 2025).
  • Brands that earn both a mention and a citation are 40% more likely to resurface across runs than brands that are only cited, in Staying Seen in AI Search (September 2025).
  • 28% of AI responses include brands that are both mentioned and cited in Staying Seen in AI Search (September 2025).
  • Brands earning visibility were 3x more likely to be cited than to be both cited and mentioned in Staying Seen in AI Search (September 2025).

ChatGPT ads statistics (October 2026)

How ads appear inside ChatGPT answers for fintech queries. Figures as of October 2026.

  • We studied 2,826,899 ChatGPT answers across 131 fintech brands over sixteen weeks in 2026, and 207,785 of them carried at least one ad, in How ChatGPT Ads Are Reshaping Fintech Answers (September 2026).
  • Weekly ad load in fintech ChatGPT answers rose from near zero in April to about 53% in the week of 13 July 2026, against a 7.35% window average, in How ChatGPT Ads Are Reshaping Fintech Answers.
  • For the same questions tracked in late August and late October 2026, the share of ChatGPT answers with an ad rose from 36.1% to 39.9%, according to our Head of Research & Insights.
  • Distinct ChatGPT advertisers per question rose from 1.82 to 2.14 between late August and late October 2026, 18% more brands across the same prompts, according to our Head of Research & Insights.
  • About half of ad-carrying ChatGPT answers now show 2 or more advertisers, according to our Head of Research & Insights (October 2026).
  • Over the three months to late October 2026, the gap in ChatGPT ad share between commercial and informational prompts roughly doubled, according to our Head of Research & Insights.

Content team and budget statistics (October 2026)

How marketing and content teams are funding, staffing and using AI. Figures as of October 2026.

  • 88.4% of marketing leaders describe their AI adoption as emerging or operational, and 7.4% as advanced or leading, in a survey of 307 leaders for the 2026 Marketing Leaders Reality Index (fielded July to August 2026).
  • Teams that received their full budget request were 2.3 times more likely to grow headcount, in a survey of 307 marketing leaders for the 2026 Marketing Leaders Reality Index (fielded July to August 2026).
  • 72% of 144 content and SEO leaders expected to increase investment in AI tools over the next year in The State of Content Teams in 2025 (fielded January to March 2025).
  • 80% of content teams use AI for drafts or ideation, 73% for SEO, 55% for analytics and 40% for workflow automation, in The State of Content Teams in 2025 (fielded January to March 2025).
  • 21% of content teams have fully integrated AI across the workflow, 63% use it in specific parts and 16% are still testing, in The State of Content Teams in 2025 (fielded January to March 2025).
  • 65% of content leaders say research and ideation take the most time, and 40% name first drafts as a major bottleneck, in The State of Content Teams in 2025 (fielded January to March 2025).
  • 37% of content leaders find editing and approvals resource-heavy, 33% struggle with promotion and tracking and 26% spend significant time on SEO, in The State of Content Teams in 2025 (fielded January to March 2025).
  • Most content teams report productivity gains of 25% to 35% in The State of Content Teams in 2025 (fielded January to March 2025).
  • 27% of content leaders cite budget constraints and 24% report internal pushback against AI-generated content in The State of Content Teams in 2025 (fielded January to March 2025).

How we keep these statistics current

We update a statistic when its study reruns, and log every change below. Figures as of October 2026.

Changelog

  • 29 October 2026: Page created with 115 statistics from 15 studies.
  • 29 October 2026: Added 10 statistics from three October 2026 research posts by our Head of Research & Insights on LinkedIn, forum and ChatGPT ad citations.

Turn the AI-first buyer shift into a category advantage

Whether you’re starting from square one with AI discovery or ramping a successful program, our team will meet you where you are to drive results.
© 2026 AirOps. All rights reserved.