本文へスキップ

AI Search Statistics for 2027: 100 Numbers Traced to Their Source

100 AI search statistics, each read at its original source on 11 November 2026: who uses AI chatbots and AI Overviews, engine shares, click effects, crawler ratios, business adoption, plus the popular figures we could not trace.

Featured image for AI Search Statistics for 2027: 100 Numbers Traced to Their Source

About half of US adults (49%) now use AI chatbots, 24% use one every day, and 60% read the AI summaries at the top of search results, according to Pew Research Center's survey of 5,119 adults in February 2026. Google says AI Overviews has more than 2.5 billion monthly users and AI Mode more than 1 billion. Google also still receives 91.05% of search engine referrals worldwide (StatCounter, October 2026), so much of the AI answering is happening inside Google, not somewhere that replaced it.

This page collects 100 AI search statistics, and every one links to the page where we read it. Government and survey-research sources lead each section. Vendor studies are labeled with their sample. Every figure was checked against its source on 11 November 2026, and the data runs to November 2026. No 2027 data exists yet: the year in the title means these are the numbers to plan 2027 with. Near the end you will find a table of popular figures we could not trace, with what to quote in their place.

Key statistics

  • 49% of US adults use AI chatbots, up from 33% in 2024, and 24% use them daily (Pew, Feb 2026).
  • 62% of US adults aged 18-64 used generative AI in May 2026, up from 45% in August 2024 (St. Louis Fed).
  • One in three EU residents aged 16-74 used generative AI tools in 2025 (Eurostat).
  • 60% of US adults read AI summaries in search results: 72% of 18-29s, 38% of those 65 and over (Pew, Feb 2026).
  • 2.5 billion monthly users for AI Overviews and 1 billion for AI Mode, by Google's own count (I/O keynote, May 2026).
  • 8% vs 15%: the share of Google visits that ended in a click on a result with and without an AI summary (Pew, March 2025 browsing data).
  • 51.58% of AI chatbot referrals worldwide came from ChatGPT in October 2026; Gemini sent 28.86% (StatCounter). StatCounter's series shifted sharply that month (ChatGPT read 80.47% in September 2026), so treat the October split as provisional.
  • 38,065.7 to 1: pages Anthropic crawled per visitor it referred in July 2025. Google's ratio was 5.4 (Cloudflare).
  • 19.8% of US businesses used AI as of 3 May 2026 (US Census Bureau); 19.95% of EU enterprises did in 2025 (Eurostat).
  • 44.4% of the prompt-engine pairs we track flipped between mentioning and not mentioning the brand at least once (AI Rank Lab, 45 pairs, 997 runs, Feb-Nov 2026).

How many people use AI search and AI chatbots?

In the United States, 49% of adults use AI chatbots and 62% of working-age adults use generative AI of some kind. In the EU, one in three people aged 16-74 used a generative AI tool in 2025. Worldwide, ChatGPT alone had reached around 10% of the adult population by July 2025. The table puts each figure next to who was asked and when, because that is where most of the disagreement between sources comes from.

SourceWho was measuredPeriodWhat was measuredFigure
Pew Research Center5,119 US adults (18+)17-23 Feb 2026Ever use AI chatbots49% (33% in 2024)
Pew Research CenterSame survey17-23 Feb 2026Use chatbots daily24%
Pew Research CenterSame survey17-23 Feb 2026Use chatbots to search for information42%
Pew Research CenterSame survey17-23 Feb 2026Ever read AI summaries at the top of a search result60%
St. Louis Fed, Real-Time Population SurveyUS adults 18-64Aug 2024 to May 2026Use generative AI45% to 62%
FRED series RPSGENAIUSAGESHARENONWORKUS adults 18-64Q3 2026Use generative AI outside work49.49% (41.58% in Q3 2025)
EurostatEU individuals 16-742025Used generative AI toolsOne in three
Reuters Institute Digital News Report 202697,520 people in 48 markets2026 surveyUse AI chatbots for news weekly10% (7% a year earlier)
NBER Working Paper 34255World adult populationThrough July 2025Had adopted ChatGPTAround 10%

Pew's 49% and the Fed's 62% do not contradict each other. Pew asked all adults, including those over 65, about chatbots. The Real-Time Population Survey covers ages 18 to 64 and asks about generative AI more broadly. If you need one US number for "how many people use AI chatbots", we would quote Pew, with the caveat Pew itself gives: the 2024 question went only to people who had heard of chatbots, so the jump from 33% is not perfectly like for like.

How often do Americans use AI chatbots?

A quarter of US adults (24%) use chatbots daily: 12% several times a day, 8% about once a day and 4% almost constantly (Pew, 17-23 Feb 2026). The most common use Pew recorded was searching for information, named by 42% of all adults. That is the closest thing to an official "AI search usage" figure for the US, and it is a survey share, not a count of queries.

Who reads AI summaries in search results?

Six in ten US adults (60%) say they read AI summaries at the top of a search result; 30% say they do not and 10% are not sure. Age splits it sharply: 72% of adults 18-29, 68% of those 30-49, 59% of those 50-64 and 38% of those 65 and over (Pew chart, same survey).

Is AI use at work growing as fast as personal use?

Both are growing, and personal use is ahead. The share of US workers using generative AI for their jobs rose from 33% to 45% between August 2024 and May 2026 (St. Louis Fed). Use outside work reached 49.49% of adults 18-64 in Q3 2026, against 41.58% a year earlier (FRED). An earlier Federal Reserve note put work use at about 41% of the workforce in November 2025.

How many Europeans use generative AI?

One in three EU individuals aged 16-74 used generative AI tools in 2025, most often for private reasons (25%), then for work (15%) and formal education (9%). Denmark (48%) and Estonia (47%) lead the EU; Norway, outside the EU, is at 56% (Eurostat, data extracted December 2025). Eurostat plans to update the page on 27 January 2027, so this is the figure you will be quoting for most of the winter.

What do people use chatbots for?

Mostly not for work. In NBER's study of ChatGPT usage, non-work messages rose from 53% to more than 70% of usage, and three topics (Practical Guidance, Seeking Information and Writing) made up nearly 80% of conversations (NBER w34255, data through July 2025). News is a small slice: 10% of people across 48 markets use AI chatbots for news weekly, 16% of under-35s, and only 20% trust chatbot answers (Reuters Institute, 2026). For scale, the ITU estimates 6 billion people were online in 2025.

Which AI chatbot has the largest market share? It depends on what you count

ChatGPT leads on every measure we could find, but the gap changes with the yardstick. It is used by 44% of US adults, sends 51.58% of chatbot referrals to websites worldwide, and accounts for 52.7% of AI-assistant referrals on sites that run our tag. The rest of the ranking reshuffles depending on whether you count people or clicks.

One warning about the StatCounter column. Its October 2026 figures follow a sudden shift in the series: ChatGPT's worldwide share read 80.47% in September and 51.58% in October, with no note from StatCounter explaining the change. A drop that size in one month looks like a change in what was counted, not in what people use. Quote the October split as provisional and name the month.

AssistantUS adults who use it (Pew, Feb 2026)Share of chatbot referrals worldwide (StatCounter, Oct 2026)Share of AI referral page views on sites with our tag (AI Rank Lab, Jul 2025-Nov 2026)
ChatGPT44%51.58%52.7%
Gemini24%28.86%20.0%
Copilot17%2.18%1.5%
Meta AI14%Not listedNot in our data
Grok8%Not listedNot in our data
Claude6%2.3%13.2%
PerplexityNot in the figures we read15.04%12.6%

Sources: Pew Research Center; StatCounter AI chatbot market share; AI Rank Lab site-tag data.

Read the columns as three different questions:

  • Pew counts people. It is the share of US adults who say they ever use each product. Copilot's 17% is everyone who says they use it, however rarely.
  • StatCounter counts clicks. Its FAQ says the figures are page views referred to the more than 1 million sites that carry its code, "not the number of searches performed". Copilot's 2.18% and Perplexity's 15.04% say more about how each product links out than about how many people use it.
  • Our column counts AI-assistant referrals on a small set of sites. It covers 4,050 page-view events where the referrer was an AI assistant, on 32 sites using the AI Rank Lab tag, from 27 July 2025 to 11 November 2026. These are sites that chose to install our tag, not a random sample, and the counts are page views, not people. We would not generalize Claude's 13.2% from it.

Which to quote: Pew for "how many people use X", StatCounter for "which assistant sends the most traffic". Neither can be added to search engine share. On the same StatCounter method, Google held 91.05% of search engine referrals in October 2026 and Bing 4.8% (StatCounter), but a chatbot referral and a search referral come from different pools and StatCounter does not publish one combined denominator.

How much website traffic comes from AI referrals?

Very little, on the sites we can see. Across sites using our tag, the median per-site share of tracked page-view events that came from an AI assistant was 0.18% (AI Rank Lab tag data, 27 July 2025 to 11 November 2026; 32 sites recorded at least one AI referral). Referrer-based counts probably miss visits too: over the same period we recorded 8,508 page views carrying utm_source=chatgpt.com on 36 sites, about four times the 2,133 ChatGPT page views identified by referrer. We did not find a government or survey-research figure for AI referral share of web traffic, so treat any single percentage you see quoted as one vendor's panel. Our guide to tracking AI referral traffic in GA4 shows how to get your own number.

How many users do Google AI Overviews and AI Mode have?

By Google's own statements, AI Overviews passed 2.5 billion monthly active users in May 2026 and AI Mode passed 1 billion. Both figures come from Sundar Pichai's public remarks, and no outside body audits them. Here is each disclosure in date order, with where it was said.

DateWhereWhat Google said
23 Jul 2025Q2 2025 earnings remarksAI Overviews: over 2 billion monthly users in more than 200 countries and territories. AI Mode: over 100 million monthly active users in the US and India.
19 May 2026I/O 2026 keynoteAI Overviews: over 2.5 billion monthly active users. AI Mode: surpassed 1 billion monthly active users about a year after launch.
22 Jul 2026Q2 2026 earnings remarksAI Mode above 1 billion monthly active users, repeated. Search and Other revenue up 17% year on year.
28 Oct 2026Q3 2026 earnings remarksAI Mode queries are on average three times longer than traditional searches. One in five AI Mode searches use voice, images or documents. Over 50 billion visual searches per month.

Three cautions before you quote these. First, they are company disclosures, and the posts we read do not define what counts as a monthly user of AI Overviews. Seeing one is not the same as reading or trusting it. Second, AI Overviews, AI Mode and the Gemini app are three separate products with three separate user counts, and roundups regularly blur them. Third, the October 2026 remarks gave no new user count for AI Mode or AI Overviews, so the May and July figures are the latest. Our explainer on how AI Mode differs from AI Overviews covers what each one is.

Site owners now get a first-party view as well. Search Console's generative AI performance reports, announced on 3 June 2026 and available to all websites since 31 August 2026, show how often your URLs appeared in AI Overviews, AI Mode and Discover's AI features. The announcement describes impressions; it does not mention clicks.

Three measuring tools pointed at the same group of AI assistant icons, one counting people, one counting clicks and one counting crawler visits, each showing a different result

Do AI Overviews reduce clicks to websites?

Yes, on the searches where they appear. In Pew's tracking of 900 US adults, people clicked a search result on 8% of Google visits that showed an AI summary and on 15% of visits that did not. They clicked a link inside the summary itself on 1% of visits, and ended the browsing session after 26% of pages with a summary against 16% without (Pew, 68,879 searches from March 2025). It observed real browsing instead of a keyword list, and the data is now 20 months old.

How often do AI Overviews appear in Google search results?

Somewhere between 6% and 25% of searches during 2025, depending on the method and the month.

  • User panel: 18% of the Google searches Pew observed in March 2025 produced an AI summary, and 58% of panelists ran at least one search that did (Pew).
  • Keyword panel: Semrush found AI Overviews on 6.49% of queries in January 2025, 24.61% at the July 2025 peak and 15.69% in November 2025, across 10M+ keywords in its database (Semrush, updated 15 December 2025). This is a vendor study.

The two are not the same measurement. Pew counts what people typed. Semrush counts a keyword list, where every keyword weighs the same however often it is searched. Query length matters more than either headline: in Pew's data only 8% of one- or two-word searches produced a summary, against 53% of searches with 10 or more words. We found no primary measurement of the trigger rate for 2026, and we would not quote a 2025 rate as if it were current.

How big is the click loss for a page that ranks first?

Ahrefs estimates a 58% lower click-through rate for the top-ranking page when an AI Overview is present. It compared 300,000 keywords in December 2023 and December 2025 using aggregated Search Console data: position-1 CTR on AI Overview keywords fell from 0.073 to 0.016, against an expected 0.037 without them. Its April 2025 version of the same study found 34.5% (Ahrefs, 4 February 2026). Ahrefs calls this a correlation, and it is desktop CTR only.

One vendor finding points the other way. On more than 200,000 keywords that gained an AI Overview, Semrush measured the zero-click rate falling from 33.75% to 31.53% (Semrush). The two can both be true: Semrush asks whether any click happened after the search, and Ahrefs asks whether the first organic result got it. If your question is "what happens to my page at position 1", use Ahrefs with its caveats. If it is "do people stop clicking altogether", Pew's 8% versus 15% is the better-grounded answer.

What have publishers actually lost?

The Reuters Institute cites Chartbeat data showing Google organic search traffic to more than 2,500 sites fell by 33% globally between November 2024 and November 2025, and by 38% in the US (Digital News Report 2026). That is a measured count on publisher sites, and it does not isolate AI summaries from other causes. Our AI Overviews CTR data post goes deeper on the click studies.

Is AI search replacing Google?

Not in any measured data we could open. Google's share of search engine referrals was 91.05% in October 2026 (StatCounter). Cloudflare, which says more than 20% of the web sits behind its network, put Google at about 88% of referral traffic in a report dated 1 July 2026 (Cloudflare). Alphabet reported 17% revenue growth in Search and Other for Q2 2026 (Q2 2026 remarks) and said in October that its year-over-year query growth rate was the highest in five years (Q3 2026 remarks). The last two are Google describing itself. Still, nothing here supports the forecasts of shrinking search volume that circulate in roundups.

How much do AI crawlers crawl compared with the visitors they send?

Far more than they send back. In July 2025, Anthropic's crawlers fetched 38,065.7 pages for every visitor Anthropic's products referred to a site, and OpenAI's fetched 1,091.4. Google's ratio was 5.4 (Cloudflare, crawl-to-click gap). These are network-level counts from sites behind Cloudflare, and they are 16 months old; we did not find a newer published table.

CompanyPages crawled per referred visit, Jan 2025Pages crawled per referred visit, Jul 2025
Anthropic286,930.138,065.7
OpenAI1,217.41,091.4
Perplexity54.6194.8
Microsoft38.540.7
Google3.85.4

Source: Cloudflare, published 29 August 2025.

  • Volume: AI bots averaged 4.2% of HTML requests on Cloudflare's network between 1 January and 2 December 2025. Googlebot alone made 4.5% (Cloudflare Radar 2025 Year in Review).
  • Which bots: GPTBot's share of AI-only crawling rose from 11.9% to 28.1% between July 2024 and July 2025, and ClaudeBot's from 15% to 23.3% (Cloudflare).
  • Purpose: training accounted for 72% of AI crawling in July 2024 and 79% in July 2025; search fell from 26% to 17% (Cloudflare).
  • 2026 update: Cloudflare says 52% of crawler requests were for AI training as of June 2026, up from 22% in spring 2025, with mixed-use crawlers at over 36% (Cloudflare, 1 July 2026). The denominator here is all crawler requests, search crawlers included, so do not set the 52% beside the 79% above.

One conflict inside a single source: the Radar review's headline says "user action" crawling (a bot fetching a page because a person asked an assistant something) grew by over 15x in 2025, while its body text says over 21x from January to early December. We would quote "over 15x", the more conservative of the two.

How many sites block AI crawlers?

Few overall, many at the top. GPTBot was named in about 4.5% of robots.txt files in the July 2025 HTTP Archive crawl, up from 2.6% in 2024, and in 20.9% of the top 1,000 sites (Web Almanac 2025, about 12.9 million sites). The same crawl found a valid llms.txt on 2.13% of desktop sites. Google's own guidance says Search "itself doesn't use" such files, so that 2.13% is a bet on other assistants.

We publish no crawler counts of our own. The AI Rank Lab tag runs in the browser and most crawlers do not run JavaScript, so our data cannot see them.

How many businesses use AI, and what do they use it for?

About one in five, by official counts on both sides of the Atlantic. The US Census Bureau's Business Trends and Outlook Survey put AI use at 19.8% of businesses as of 3 May 2026, and Eurostat put it at 19.95% of EU enterprises with 10 or more employees in 2025. The two surveys define AI use differently and cover different firm sizes, so the near-match is partly luck.

  • United States: 19.8% overall; 39.7% in the Information sector and 33.9% in Finance and Insurance; 37% of firms with at least 250 employees. Between 20% and 23% expected to be using AI within six months. The Census Bureau widened the question in November 2025 from AI used "in producing goods or services" to AI used "in any business function", so earlier readings are not comparable (US Census Bureau, 26 May 2026).
  • Workers versus firms: although only about 18% of firms had adopted AI at the end of 2025, 78% of the US labor force works at a firm that has, because large employers adopt first (Federal Reserve FEDS Note, 3 April 2026).
  • European Union: 19.95% of enterprises in 2025, up 6.47 percentage points in a year. Large enterprises 55.03%, small ones 17%. Denmark is highest at 42.03% and Romania lowest at 5.21% (Eurostat).
  • What for: marketing or sales is the most common purpose in the EU, named by 34.70% of enterprises that use AI (Eurostat).

What we could not find is an official number for how many companies optimize for AI search specifically. Every "X% of marketers are investing in GEO" figure we saw came from a vendor or agency survey with an unclear sample, so none is published here. Consultancy surveys that report 80%-plus "AI adoption" count any use by anyone in a large organization; the statistical offices above count firms, most of which are small.

How stable are AI answers and citations between runs?

Less stable than most visibility statistics admit. In our own tracking, 44.4% of prompt-engine pairs (20 of 45) flipped between mentioning and not mentioning the tracked brand at least once, and the median overlap in cited domains between one run and the next was 0.41 on a 0-to-1 scale.

Method: we took every prompt-and-engine pair in AI Rank Lab with at least five successful runs between 13 February and 10 November 2026. That left 45 pairs and 997 runs. Overlap is the Jaccard index of cited domains between consecutive runs of the same prompt on the same engine: 1 means the same set of domains, 0 means none in common.

EnginePrompt-engine pairsRunsPairs where the brand mention flippedMedian cited-domain overlap between consecutive runs
Gemini1428328.6%0.26
ChatGPT1224858.3%0.50
Claude1223541.7%0.52
Perplexity622466.7%0.59
All engines4599744.4%0.41

Source: AI Rank Lab prompt runs, 13 February to 10 November 2026. The total includes one Google AI Overview pair with 7 runs, which is too few to report on its own.

Two more numbers from the same set. The median prompt cited 2.76 times as many distinct domains across all its runs as it did in a single run. And a median of only 3.6% of the domains cited for a prompt appeared in every run. So a single check of "who does ChatGPT cite for this question" shows you roughly a third of the sources in play.

The limits are real. This is 45 pairs from a small number of brands that use our product, not a sample of the web, and six Perplexity pairs cannot support a ranking of engines. What it does support is a rule for reading other people's statistics: a citation share measured from one run per prompt carries noise the study probably did not report. The longer write-up is in our AI answer stability study, and what sources LLMs cite covers the domain patterns.

Which AI search statistics have no traceable source?

Several of the most repeated figures are forecasts, vendor models or numbers whose original we could not open. We went through 10 ranking statistics pages on 11 November 2026, and these are the ones we would not quote as fact.

Figure in circulationWhat we foundQuote this instead
"90 million Americans will use AI for search by 2027"It turns up on marketing blogs, sometimes paired with "13 million in 2023". None of the search results we saw names a study, and we could not find an original.42% of US adults use chatbots to search for information (Pew, Feb 2026).
"Search engine volume will drop 25% by 2026"A forecast attributed to a Gartner press release of February 2024; the page would not load for us, so we have not read its wording. It was a prediction, not a measurement, and 2026 is nearly over.Google at 91.05% of search referrals (StatCounter, Oct 2026); Alphabet reporting its highest query growth rate in five years (Oct 2026).
"AI search visitors are worth 4.4x more"On the vendor roundup where we found it, the figure sits in a heading with no number in the text under it. We did not find the dataset behind it.Nothing yet. We know of no primary-source conversion figure for AI referrals across sites.
"AI search traffic will overtake traditional search by 2028"A vendor's projection, drawn as a chart running to 2029. It is a model, not a measurement.Label it as one vendor's forecast, or leave it out.
ChatGPT weekly users (900 million, 1 billion and 1.2 billion all appear)Each is attributed to OpenAI, often through a news report. The OpenAI page we tried would not load for us, so we could not confirm any of them at the origin.Around 10% of the world's adult population by July 2025 (NBER w34255), or 44% of US adults (Pew, Feb 2026).
"About 60% of searches end without a click" next to "8% click when there is an AI summary"Roundups attribute the first to a 2025 consultancy survey, which we did not open, so we do not publish it. The two measure different things and are often printed side by side as one trend.Pew's 8% versus 15%, with its March 2025 date.

The pattern across those pages is consistent. Four of the 10 had a methodology section and none had a consolidated source list. Among the headline figures we checked, not one came from a government statistics office. Several sent readers to a news article or another roundup instead of the original. That is how a forecast turns into a "statistic" after two or three hops.

How to measure your own numbers

Population statistics tell you the size of the audience. They cannot tell you whether an assistant mentions your brand, and the stability data above says one manual check will mislead you. Three numbers are worth collecting for your own site:

  1. Impressions in Google's AI features. Open Search Console and use the generative AI performance report. It is free and it is Google's own count.
  2. AI referral visits. Filter GA4 by referrer and by utm_source=chatgpt.com. Expect a small share.
  3. Mentions and citations for the questions your buyers ask. Run the same prompts repeatedly and keep every run. This is the step we built AI Rank Lab for: brand visibility tracking runs the prompts you choose on ChatGPT, Claude, Gemini and Perplexity when you run them and stores each run, and the citation tracker shows which domains and URLs the engines cite for those prompts. It will not give you Google's internal numbers; nothing outside Google can.

Methodology

We set the rules before collecting anything.

  • Origin only. A number is here only if we opened the page that produced it and read the figure there on 11 November 2026. Where a roundup quoted a figure, we went to the original. If the original would not open, the number was dropped or moved to the "could not trace" table.
  • Source order. Government and intergovernmental statistics first (US Census Bureau, Federal Reserve, Eurostat, ITU), then survey research and academic work (Pew, Reuters Institute, NBER, HTTP Archive), then companies reporting their own numbers (Google, Cloudflare, StatCounter). Two vendor studies (Semrush, Ahrefs) are included with their samples because no primary source measures the same thing; neither is the only support for a headline claim.
  • Kind of number. Each figure is described as a survey share, a measured count or a company disclosure, with the period the data covers. We publish no forecasts as statistics.
  • Count. The 100 in the title is the number of entries in our source ledger: one entry per published figure, each tied to one source and one period.
  • Our own data. The AI Rank Lab figures come from anonymized, aggregated product data pulled on 11 November 2026: prompt runs from 13 February to 10 November 2026 and site-tag referral events from 27 July 2025 to 11 November 2026. Samples are small and come from our customers, and we say so next to each number. We left out our crawler counts (the tag cannot see most bots) and our Google AI Overview runs (7 runs is too few).
  • Left out. Market-size estimates, "X% of marketers say" surveys, user counts for OpenAI, Perplexity and Anthropic that we could not confirm at the origin, and any AI Overview trigger rate for 2026, because we found no primary measurement. Market narrative is in our AI search state of the market report from March 2026; this page is the number reference.
  • Cutoff and updates. Data to 11 November 2026. Known releases that will change this page: Eurostat's enterprise AI update on 15 December 2026, Cloudflare's 2026 year in review (likely in December), and Eurostat's individuals update on 27 January 2027. We will revise the figures when those land and change the date only then.
  • How it was written. AI tools helped with drafting. Every number, date and source name was checked against the pages listed below.

Sources

Every source used on this page, with publisher and publication date. All were opened on 11 November 2026.

  1. Americans and AI 2026: Chatbots, Smart Devices and Views on Impact - Pew Research Center, 17 June 2026.
  2. A majority of Americans overall say they read AI search summaries, but adults 65 and up are least likely (chart) - Pew Research Center, 17 June 2026.
  3. Google users are less likely to click on links when an AI summary appears in the results - Pew Research Center, 22 July 2025.
  4. What Work Does Generative AI Do? - Federal Reserve Bank of St. Louis, 1 September 2026.
  5. Generative Artificial Intelligence, Adoption Rate Outside of Work: Working-Age Adults (RPSGENAIUSAGESHARENONWORK) - FRED, Federal Reserve Bank of St. Louis, updated 20 October 2026.
  6. Monitoring AI Adoption in the U.S. Economy (FEDS Notes) - Board of Governors of the Federal Reserve System, 3 April 2026.
  7. AI use by businesses (America Counts, Business Trends and Outlook Survey) - US Census Bureau, 26 May 2026.
  8. Digital economy and society statistics - households and individuals - Eurostat, data extracted December 2025.
  9. Use of artificial intelligence in enterprises - Eurostat, data extracted December 2025.
  10. Facts and Figures 2025 (press release) - International Telecommunication Union, 17 November 2025.
  11. Digital News Report 2026: Executive Summary - Reuters Institute for the Study of Journalism, University of Oxford, 16 June 2026.
  12. How People Use ChatGPT (NBER Working Paper 34255) - National Bureau of Economic Research, September 2025.
  13. Web Almanac 2025: Generative AI chapter - HTTP Archive, 15 January 2026.
  14. AI Chatbot Market Share Worldwide - StatCounter Global Stats, data for October 2026.
  15. Search Engine Market Share Worldwide - StatCounter Global Stats, data for October 2026.
  16. StatCounter Global Stats FAQ - StatCounter, read 11 November 2026.
  17. Google I/O 2026: Sundar Pichai's opening keynote - Google, 19 May 2026.
  18. Alphabet Q2 earnings call: CEO's remarks (Q2 2025) - Google, 23 July 2025.
  19. Alphabet earnings call Q2 2026: Sundar Pichai remarks - Google, 22 July 2026.
  20. Alphabet earnings call Q3 2026: Sundar Pichai remarks - Google, 28 October 2026.
  21. The crawl-to-click gap: Cloudflare data on AI bots, training, and referrals - Cloudflare, 29 August 2025.
  22. Cloudflare Radar 2025 Year in Review - Cloudflare, December 2025 (covers 1 January to 2 December 2025).
  23. Content Independence Day, one year on: building the business model for the agentic Internet - Cloudflare, 1 July 2026.
  24. Search Generative AI performance reports in Search Console - Google Search Central, 3 June 2026.
  25. Guide to optimizing for generative AI features on Google Search - Google Search Central, 15 May 2026.
  26. Semrush AI Overviews Study - Semrush, 15 December 2025.
  27. Update: AI Overviews Reduce Clicks by 58% - Ahrefs, 4 February 2026.
  28. AI Rank Lab first-party data: aggregated, anonymized prompt-run and site-tag data, extracted 11 November 2026. Methods and sample sizes are stated beside each figure.

Want these numbers for your own brand instead of the whole market? Start with 100 free credits, no card needed, which covers about one brand visibility prompt run.

Get a free AI ranking consultation

Want your brand to show up when ChatGPT, Gemini and Perplexity answer questions in your market? Tell us about your site and our team will send a personalized strategy.

This form is protected by reCAPTCHA. Your data is handled securely and we will never spam you.

よくある質問

ほかにご質問は? チームに質問する

How many people use AI search and AI chatbots?

In the US, 49% of adults use AI chatbots and 42% use them to search for information (Pew Research Center, 5,119 adults, February 2026). Among US adults aged 18-64, 62% used generative AI in May 2026 (St. Louis Fed). In the EU, one in three people aged 16-74 used generative AI tools in 2025 (Eurostat).

What share of Americans use AI chatbots daily?

About a quarter. Pew found 24% of US adults use chatbots daily in February 2026: 12% several times a day, 8% about once a day and 4% almost constantly.

Which AI chatbot has the largest market share?

ChatGPT, on every measure. It is used by 44% of US adults (Pew, February 2026) and sent 51.58% of AI chatbot referrals worldwide in October 2026 (StatCounter), ahead of Gemini at 28.86% and Perplexity at 15.04%. StatCounter counts referral clicks, not users, which is why Copilot is 17% in Pew and 2.18% in StatCounter. Treat the October StatCounter split as provisional: ChatGPT's share read 80.47% a month earlier and StatCounter has not explained the shift.

How many users do Google AI Overviews and AI Mode have?

Google said on 19 May 2026 that AI Overviews has over 2.5 billion monthly active users and that AI Mode had passed 1 billion. A year earlier it reported over 2 billion for AI Overviews and over 100 million for AI Mode in the US and India. These are company disclosures and are not independently audited.

Do AI Overviews reduce clicks to websites?

Yes. Pew tracked 900 US adults in March 2025 and found they clicked a result on 8% of Google visits with an AI summary, against 15% without one, and clicked a link inside the summary 1% of the time. Ahrefs estimates a 58% lower click-through rate for the first organic result when an AI Overview is present (300,000 keywords, December 2023 vs December 2025).

How often do AI Overviews appear in Google search results?

Pew observed an AI summary on 18% of Google searches in March 2025. Semrush, using a keyword panel of 10M+ keywords, measured 6.49% in January 2025, 24.61% in July 2025 and 15.69% in November 2025. We found no primary measurement for 2026.

Is AI search replacing Google?

No measured data shows that. Google held 91.05% of search engine referrals worldwide in October 2026 (StatCounter), and Alphabet said in October 2026 that its query growth rate was the highest in five years. Much of the AI answering happens inside Google, through AI Overviews and AI Mode.

How much do AI crawlers crawl compared with the visitors they send?

In July 2025 Cloudflare measured 38,065.7 pages crawled per referred visit for Anthropic, 1,091.4 for OpenAI, 194.8 for Perplexity and 5.4 for Google. AI bots averaged 4.2% of HTML requests on its network in 2025.

How many businesses use AI?

About one in five by official counts: 19.8% of US businesses as of 3 May 2026 (US Census Bureau) and 19.95% of EU enterprises with 10 or more employees in 2025 (Eurostat). In the EU, 34.70% of AI-using enterprises use it for marketing or sales.

How stable are AI answers and citations between runs?

Not very. Across 45 prompt-engine pairs and 997 runs tracked in AI Rank Lab from February to November 2026, 44.4% of pairs flipped between mentioning and not mentioning the brand at least once, and the median overlap of cited domains between consecutive runs was 0.41 on a 0-to-1 scale.

この記事はいかがでしたか?

ニュースレターを購読して、AI検索最適化の最新インサイトをメールで受け取りましょう。

No spam. Unsubscribe at any time.

See how AI engines read your site

Paste your site. The free audit scores it for SEO, answer engines and generative engines, and lists the fixes.