Generative Engine Optimization Statistics for 2026

These generative engine optimization statistics bring together 44 verified data points on AI search for 2026: how many people use AI engines, how AI Overviews change click-through rates, how much referral traffic AI sends, which sources engines cite, how much answers vary, and whether llms.txt and schema make a difference. Every figure links to its original study, announcement or a report on it and carries its publication date, so you can check the method before you quote it.

A note on reading them. Many generative engine optimization statistics come from studies run by companies that sell AI visibility or SEO tools, samples differ widely, and the market moves monthly. Treat vendor figures as directional, compare like with like (same engine, same metric, same period), and always cite the date. For the concepts behind these numbers, start with our guide to what generative engine optimization is and how it works.

Key AI search statistics at a glance

AI search usage statistics

1. ChatGPT reached 800 million weekly active users, Sam Altman said at OpenAI DevDay. TechCrunch, October 2025.
2. ChatGPT then reached 900 million weekly active users and 50 million paid subscribers. TechCrunch, February 2026.
3. The Gemini app passed 1 billion monthly active users in August 2026, up from 950 million reported in July 2026, according to Google. Implicator, August 2026.
4. Google AI Overviews had more than 2 billion monthly users across 200+ countries and 40 languages. Alphabet Q2 2025 earnings call, July 2025.
5. Google AI Mode passed 1 billion monthly users, a figure Google announced at I/O and restated on its Q2 2026 call. Search Engine Journal, May 2026.
6. Perplexity handled 780 million queries in May 2025, said CEO Aravind Srinivas at Bloomberg Tech, the latest official figure we found as of October 2026. TechCrunch, June 2025.
7. Monthly visits to generative AI platforms averaged 9.5 billion between June 2025 and May 2026, up 70% year on year. Similarweb, July 2026.
8. ChatGPT’s share of generative AI web traffic fell from about 76% to about 53% between June 2025 and May 2026, as rivals grew faster. Similarweb, July 2026.
9. AI Overviews appeared on 88% of 400 tracked US commercial keywords in August 2026, up from 45.75% in September 2025 (97.08% weighted by search volume). Semrush, October 2026.
10. 18% of Google searches produced an AI summary in a panel of 900 US adults, rising to 53% for searches of 10+ words and 60% for questions. Pew Research Center, July 2025.

AI Overviews CTR statistics

11. Users clicked a traditional result on 8% of visits when an AI summary appeared, against 15% when none did. Pew Research Center, July 2025.
12. Only 1% of visits included a click on a link inside the AI summary itself, in the same Pew panel. Pew Research Center, July 2025.
13. Users ended their browsing session on 26% of visits to result pages with an AI summary, against 16% without one. Pew Research Center, July 2025.
14. An AI Overview correlates with 58% lower CTR for the position-1 result. Ahrefs measured a position-1 CTR of 0.016 in December 2025 against a forecast of about 0.037 without AI Overviews, across 300,000 keywords. Ahrefs, February 2026.
15. The CTR impact shrinks lower on the page: −50.8% at position 2, −32.6% at position 5 and −19.4% at position 10. Ahrefs, February 2026.
16. Organic CTR fell from 1.76% to 0.61% on informational queries with an AI Overview (Seer reports this as a 61% drop; the endpoints imply about 65%), and paid CTR fell 68%, across 3,119 queries from 42 clients. Seer Interactive, November 2025.
17. Being cited in an AI Overview more than doubles organic clicks compared with not being cited: about 20,700 vs 9,400 clicks per million impressions on informational queries, against about 33,500 with no AI Overview (a correlation, not proof of cause). Seer Interactive, April 2026.
18. Organic CTR on queries with an AI Overview was 2.36% in February 2026, against 3.82% on queries without one (53 brands, 5.47 million queries). Seer Interactive, April 2026.
19. 77.6% of Google AI Mode sessions had zero external visits in Kevin Indig’s usability study of 37 US participants. Search Engine Journal, October 2025.
20. 58.5% of US and 59.7% of EU Google searches ended without a click, before AI Overviews were widespread, based on Datos clickstream data. Search Engine Land on SparkToro’s study, 2024.

AI referral traffic statistics

21. AI platforms sent 0.32% of total website traffic in 2026, up from 0.24% in 2025 and 0.02% in 2024, across 101,574 sites. SE Ranking, June 2026.
22. ChatGPT sent 74.78% of AI referrals, followed by Gemini (11.56%), Perplexity (7.23%), Copilot (3.51%) and Claude (2.62%). SE Ranking, June 2026.
23. Google sent about 345 times more traffic than ChatGPT, Perplexity and Gemini combined across about 35,000 websites, where AI was 0.1% of traffic and search 43.8%. Ahrefs, March 2025.
24. AI platforms sent 1.13 billion referrals to the top 1,000 websites in June 2025, up 357% year on year, against 191 billion from Google Search. TechCrunch on Similarweb data, June 2025 figures.
25. About 62–63% of ChatGPT referral visits now land on a homepage, after a May 2026 product change. Similarweb, July 2026.
26. Generative AI traffic to US retail sites rose 693% year on year over the 2025 holiday season. Adobe, January 2026.
27. AI-referred visitors converted 31% better than other traffic during the 2025 holiday season, spent 45% more time on site and were 33% less likely to bounce. Adobe, January 2026.
28. AI drove about 0.19% of visits against about 20.45% for organic search, roughly 108 times apart, across 40 US sites in April 2026. Aleyda Solis, May 2026.

For setting up your own measurement, see our guide to tracking AI referral traffic in GA4.

AI citation source statistics

29. Reddit, YouTube, LinkedIn, Wikipedia and Forbes were the five most-cited domains across ChatGPT, Google AI Mode, Gemini, Perplexity and AI Overviews, in 30 million US citations. Peec AI, March 2026.
30. Wikipedia made up 47.9% of ChatGPT’s top-10 cited sources and 7.8% of all its citations, across 680 million citations. Profound, June 2025.
31. YouTube appeared in 46.62% of Google AI Overview responses and 31.82% of responses across all tracked platforms, in 7.9 million responses. Ahrefs UGC benchmark, October 2026.
32. Only 37.9% of AI Overview citations came from pages ranking in the top 10, down from about 76% in July 2025, and 31% came from beyond the top 100. Ahrefs, March 2026.
33. 87% of ChatGPT citations matched Bing’s top-20 organic results, against 56% for Google, in a sample of 500+ citations. Seer Interactive, February 2025.
34. Branded web mentions correlated with AI Overview brand visibility at 0.664, against 0.218 for backlinks, across about 75,000 brands. Ahrefs, May 2025.
35. 44.3% of ChatGPT citations came from the first 30% of a page’s text, based on 18,012 citations analyzed by Kevin Indig. Practical Ecommerce, March 2026.
36. Google’s grounding used a median of about 1,929 words of source text per query, with the top source contributing 28% of it, across 7,060 queries. PPC Land on DEJAN’s study, December 2025.
37. Adding quotations raised a source’s visibility in a simulated generative engine by about 41%, and adding statistics by about 31%, while keyword stuffing scored below baseline. GEO paper, Aggarwal et al., KDD 2024.

Our guide to writing content that AI engines cite turns these findings into practical steps.

AI answer volatility statistics

38. The same list of brands came back less than 1% of the time, and in the same order less than 0.1%, across 2,961 runs of 12 prompts on ChatGPT, Claude and Google AI. Search Engine Journal on SparkToro and Gumshoe, January 2026.
39. 1,000 runs of one prompt at temperature zero produced 80 unique outputs, because of server batching rather than sampling. Thinking Machines Lab, September 2025.
40. Google AI Mode and AI Overviews shared only 13.7% of cited URLs for the same query (540,000 query pairs), and only 0.51% of answers were identical (730,000 query pairs). Ahrefs, December 2025.
41. Grok ran 6.8 fan-out searches per prompt, against 2.1 for ChatGPT and 1.4 for Perplexity, in about 5 million fan-outs. Peec AI, May 2026 (April 2026 data).

Why this happens, and how many runs you need before trusting a number, is covered in our explainer on AI answer volatility.

llms.txt and schema statistics

42. 10.13% of about 300,000 domains had an llms.txt file, and having one showed no relationship with how often a domain was cited by LLMs. Search Engine Journal on SE Ranking’s study, November 2025.
43. Google Search does not use llms.txt, according to Google’s guide to optimizing for generative AI features, which also lists special markup and chunking as unnecessary. Google Search Central, May 2026.
44. Adding JSON-LD schema produced no meaningful AI citation lift across 1,885 pages: −4.6% in AI Overviews, +2.4% in AI Mode and +2.2% in ChatGPT, the latter two not statistically significant. Ahrefs, May 2026.

What these generative engine optimization statistics mean

Read together, the numbers tell a consistent story. AI engines now reach billions of people, and Google’s AI features cut clicks to classic results sharply, yet AI referrals are still well under 1% of most sites’ traffic. Visibility inside the answer therefore matters more than the click: cited brands get roughly twice the clicks of uncited ones on queries with an AI Overview (a correlation, not proof of cause), and AI visitors who do arrive convert well.

Citations come from a broad set of sources, often beyond the top 10, and favor brands that are widely mentioned across the web. Answers vary so much from run to run that single checks are meaningless, which is why visibility has to be measured as a rate over many runs. Shortcuts such as llms.txt and schema show no measurable citation effect so far. Asheard was built around that measurement problem, tracking mentions, citations, position and sentiment across ten AI engines; the AI visibility tracking page explains how. Whichever tool you use, check the date and the method behind any statistic before you put it in a report.

Frequently asked questions

How much traffic does AI search send to websites?

Still very little compared with Google. SE Ranking’s study of 101,574 sites found AI platforms sent 0.32% of total traffic in 2026, up from 0.02% in 2024. ChatGPT accounted for about three quarters of those AI referrals, with Gemini, Perplexity, Copilot and Claude sharing the rest.

How much do AI Overviews reduce clicks?

Ahrefs found that an AI Overview correlates with 58% lower click-through for the top organic result, based on 300,000 keywords compared between December 2023 and December 2025. Pew Research found users clicked a traditional result on 8% of visits with an AI summary, against 15% without one.

Which websites do AI engines cite most?

Reddit, YouTube, LinkedIn, Wikipedia and Forbes topped Peec AI’s March 2026 analysis of 30 million US citations across five AI engines. Shares differ a lot by engine and month: ChatGPT leans toward Wikipedia and editorial sites, while Google’s AI features cite YouTube and social platforms more often.

Does llms.txt improve AI citations?

There is no evidence that it does. SE Ranking found no relationship between having an llms.txt file and how often about 300,000 domains were cited by LLMs. Google says its Search does not use the file, though creating one for other services does no harm.