AI Search Market Share Statistics 2026

Statistics 11–20

Google Search vs AI Search: How Big Is the Shift?

The rapid growth of ChatGPT and other generative AI platforms has created a widespread perception that conventional search is being replaced. The available evidence presents a more complex picture.

AI-assisted information seeking is expanding rapidly, but Google continues to operate at enormous scale. At the same time, Google is incorporating generative AI directly into Search, making the boundary between “traditional search” and “AI search” increasingly difficult to define.

11. Google held approximately 91.1% of worldwide search-engine market share in August 2026

12. Bing held approximately 4.5% of worldwide conventional search-engine market share

13. Google was approximately 20 times larger than Bing by measured worldwide search-engine share

14. Google accounted for nearly 40% of measured website traffic in an Ahrefs study

15. ChatGPT accounted for approximately 0.21% of website traffic in the same Ahrefs dataset

16. Google sent approximately 190 times more traffic to websites than ChatGPT

17. Approximately 65% of ChatGPT usage was classified as search or search-like activity by Ahrefs

18. ChatGPT’s estimated search-like activity equated to approximately 11.86% of Google’s search volume under Ahrefs’ methodology

19. AI Overviews exceeded 2.5 billion monthly active users globally

20. Google AI Mode exceeded one billion monthly active users

AI Search Is Growing Alongside Google, Not Simply Replacing It

Statistics 11–20 challenge the idea that the search market can be divided neatly into “Google” on one side and “AI search” on the other.

Standalone AI assistants are clearly becoming significant information-discovery platforms. However, conventional Google Search remains vastly larger as a source of website traffic, while Google itself is integrating generative AI into the search experience at billion-user scale.

This creates an important measurement problem. A search completed through a Google AI Overview or AI Mode is simultaneously part of Google’s ecosystem and part of the broader transition towards generative search.

What This Means for Search Visibility

For organisations, the evidence suggests that abandoning conventional SEO in favour of AI optimisation would misunderstand the current market.

Google continues to generate substantial discovery and website traffic. At the same time, visibility increasingly depends on whether information can also be understood, selected, cited and recommended within generative systems.

SEO and Generative Engine Optimisation therefore increasingly operate as connected disciplines. Traditional rankings remain commercially important, while entity authority, source eligibility, citation visibility and recommendation confidence become additional dimensions of search performance.

Evidence Sources for Statistics 11–20

This section combines three different forms of evidence: Statcounter conventional search-engine share, Ahrefs website-traffic and search-volume analysis, and Google’s first-party reporting on AI Overviews and AI Mode.

View Statcounter Worldwide Search Engine Market Share

View Ahrefs ChatGPT vs Google Analysis

View Google Search Ecosystem Update

View Google I/O 2026 AI Announcements

Statistics 21–30

UK AI Search Adoption & Changing Search Behaviour

Worldwide platform statistics establish the scale of generative AI, but they do not tell us how extensively people in the United Kingdom are actually using AI tools or how those tools compare with established search behaviour.

UK evidence from Ofcom and Ipsos iris shows substantial adoption of AI assistants alongside continued heavy use of Google and conventional online search. The evidence therefore points towards expansion of the discovery ecosystem rather than a simple transfer of users from one search channel to another.

21. 54% of UK adults report using AI tools such as ChatGPT, Copilot or Gemini

22. AI-tool usage reaches 79% among UK adults aged 16–24

23. 74% of UK adults aged 25–34 use AI tools

24. 15.8 million UK online adults visited at least one major AI chatbot in June 2025

25. Major AI chatbots reached 32% of UK online adults

26. ChatGPT reached 13 million UK online adults in June 2025

27. ChatGPT reached 26% of UK online adults

28. UK users generated approximately 1.8 billion ChatGPT visits during the first eight months of 2025

29. The equivalent UK ChatGPT visit figure was approximately 368 million in 2024

30. UK ChatGPT visits increased by approximately 389% between the comparable 2024 and 2025 periods

What the UK Data Shows

The UK evidence demonstrates that generative AI adoption is no longer confined to a small group of early adopters. More than half of adults report using AI tools, while adoption among younger adults is considerably higher.

ChatGPT has also established a particularly strong position. Its measured UK audience reached 13 million online adults by June 2025, while total UK visits expanded dramatically during the first eight months of that year.

However, the statistics describe different behaviours. Ofcom’s 54% figure measures reported use of AI tools generally. The 32% figure measures visits to major AI chatbots among UK online adults. The 26% figure measures ChatGPT’s audience reach specifically.

They should therefore be understood as complementary indicators of adoption rather than competing estimates of a single UK AI search market share.

AI Growth Does Not Mean Google Has Disappeared

The growth rates above are substantial, but they need to be viewed alongside the continuing scale of conventional search.

Ofcom’s Online Nation 2025 reported Google Search being used by 82% of UK adults and estimated approximately three billion UK Google searches per month.

It also reported that 95% of UK adult internet users had searched online for information during the preceding three months.

The evidence therefore supports a picture of coexistence: AI assistants are gaining substantial audiences and usage while conventional search remains deeply embedded in UK digital behaviour.

What UK Adoption Means for Businesses

The UK data suggests that organisations increasingly need to think about search visibility across more than one discovery environment.

A potential customer may begin with Google, encounter an AI Overview, continue research through ChatGPT or Gemini, compare alternatives, return to conventional search and finally visit a business directly.

This makes the relationship between traditional SEO, entity authority, content quality, brand signals, citations and Generative Engine Optimisation increasingly important. Visibility needs to persist across the discovery journey rather than depend on a single ranking or platform.

Evidence Sources for Statistics 21–30

The principal evidence for this section comes from Ofcom’s Adults’ Media Use and Attitudes 2026, Ofcom’s research into the emergence of answer engines, Ipsos iris audience measurement cited by Ofcom, and Online Nation 2025.

View Ofcom Adults’ Media Use and Attitudes 2026

View Ofcom — The Era of Answer Engines

View Ofcom Online Nation 2025

Statistics 31–40

AI Search Clicks, Referrals & the Changing Economics of Website Traffic

Market share measures how audiences are distributed between platforms. For businesses and publishers, however, another question is equally important: what happens after a user receives an AI-generated answer?

Research into Google AI Overviews, ChatGPT referrals and AI citation behaviour indicates that generative search can change the relationship between visibility and website traffic. A brand or source may influence an answer without receiving the type of click traditionally associated with organic search.

31. AI summaries appeared on 18% of Google searches measured by Pew Research Center

32. 58% of participants encountered at least one Google AI summary during the study month

33. Traditional search-result links received clicks on 8% of visits when an AI summary appeared

34. Traditional search-result links received clicks on 15% of visits without an AI summary

35. Traditional-result clicking was approximately 47% lower when an AI summary appeared

36. Only 1% of visits containing an AI summary produced a click on a cited source

37. 26% of browsing sessions ended after a search-results page containing an AI summary

38. 16% of sessions ended after search-results pages without an AI summary

39. Ahrefs estimated a 58% reduction in position-one CTR when AI Overviews were present

40. AI Overview CTR effects extended throughout Google’s organic top ten

Visibility and Traffic Are Becoming Different Measurements

Traditional SEO developed around a relatively clear sequence: a search was performed, results were displayed, a ranking attracted a click and the user arrived on a website.

Generative search introduces another possibility. A search platform can retrieve information from multiple sources, synthesise an answer and satisfy much of the user’s information requirement before an external website is visited.

This does not make source visibility irrelevant. It changes what visibility can mean. A source can contribute to an AI answer, appear as a citation, influence a recommendation or reinforce a brand without necessarily producing an immediate click.

The Commercial Implication

If AI-generated answers reduce the need to click through for some informational searches, organisations may need to evaluate search performance using a broader set of indicators.

Organic sessions and click-through rates remain important, particularly for transactional and commercially valuable journeys. But they may no longer describe the complete influence of search visibility.

Additional indicators can include AI citations, brand mentions, recommendation visibility, source selection, branded-search growth, direct traffic, assisted conversions and whether an organisation appears during AI-mediated comparison and research.

The strategic objective therefore expands from generating rankings and clicks to maintaining visibility throughout a search ecosystem in which some discovery takes place before the user reaches a website.

Evidence Sources for Statistics 31–40

The primary evidence comes from Pew Research Center’s analysis of Google search behaviour and Ahrefs’ large-scale comparison of informational keywords with and without AI Overviews.

View Pew Research Center — Google AI Summaries & Click Behaviour

View Ahrefs — AI Overviews Reduce Clicks

Statistics 41–50

AI Citations, Brand Mentions & Source Selection

As AI search grows, market share alone cannot explain which organisations actually gain visibility inside generated answers. Citation selection, brand mentions, source eligibility and retrieval behaviour create another competitive layer within the AI search ecosystem.

Research from Semrush and Ahrefs demonstrates that being cited by an AI system is not the same as being mentioned as a brand, while conventional Google rankings do not fully determine which pages or domains AI systems select as sources.

41. 61.7% of measured AI citations were “ghost citations”

42. 74.9% of measured AI appearances contained a citation

43. Only 38.3% of measured appearances contained a brand mention

44. Only 13.2% of measured appearances contained both a citation and a brand mention

45. ChatGPT cited sources in 87% of measured appearances but mentioned brands in only 20.7%

46. Gemini mentioned brands in 83.7% of measured appearances but cited sources in 21.4%

47. Comparative content generated 2.4 times more brand mentions than informational content

48. Only approximately 12% of AI-cited URLs ranked in Google’s top ten for the original prompt

49. AI-cited URLs showed approximately 10% overlap with Bing’s top ten

50. 37.1% of Google AI Overview citations ranked in the conventional organic top ten

AI Visibility Is More Than Being Cited

Statistics 41–50 reveal an important distinction between source authority and brand visibility.

A website can provide information that an AI system considers useful enough to cite without the organisation being named prominently in the answer. Conversely, a brand can be mentioned during a recommendation or comparison without its own website necessarily being presented as the supporting citation.

This creates at least two distinct forms of generative visibility: citation visibility, where a source is selected as evidence, and brand visibility, where an organisation becomes part of the generated answer itself.

For commercially important prompts, the strongest outcome may be achieving both: being recognised as a relevant entity while also possessing authoritative content that can support the recommendation.

AI Citation Selection Is Not the Same as Organic Ranking

The Ahrefs evidence also challenges the assumption that AI systems simply select the pages already ranking highest in conventional search.

Only around 12% average exact-URL overlap was observed between AI citations and Google’s top ten in the 15,000-prompt study. Google AI Overviews showed a stronger relationship with Google’s own rankings, but even there more than one-third of cited URLs did not rank within the organic top 100 for the same query.

This does not mean that conventional SEO is unimportant. Rankings can reflect many of the same underlying signals that make a source useful: topical relevance, authority, crawlability, content quality, entity clarity and external validation.

The evidence instead suggests that AI retrieval and citation introduce an additional selection layer. Organisations may therefore need to optimise not only for ranking eligibility but also for source eligibility.

Research Limitation

These studies analyse defined prompt sets, platforms and search-result datasets. They should not be interpreted as proving that the same citation or brand-mention rates apply across every industry, country or query.

AI systems also change rapidly. Citation behaviour observed during one research period may shift as models, retrieval systems, interfaces and search integrations are updated.

Evidence Sources for Statistics 41–50

The primary evidence in this section comes from Semrush research into the relationship between citations and brand mentions, together with Ahrefs studies comparing AI citations with conventional Google and Bing rankings.

View Semrush — Ghost Citations Study

View Ahrefs — AI Citations vs Google Rankings

View Ahrefs — Google AI Overview Citation Selection

Statistics 51–60

ChatGPT Retrieval, Reasoning, Query Fan-Out & Source Diversity

AI search does not necessarily process a user’s prompt as a single conventional search query. Generative systems can interpret the request, break it into additional searches, retrieve information from multiple sources and construct an answer from the resulting evidence.

Research into ChatGPT reasoning behaviour demonstrates how significantly this retrieval process can change depending on the reasoning configuration. Greater reasoning depth can produce more searches, more citations and a wider source pool — while also changing which domains ultimately appear in the answer.

51. ChatGPT’s citation rate increased from 50% to 68% under higher reasoning

52. The measured citation rate increased by 18 percentage points

53. Average citations per response increased from 2.6 to 4.5

54. Average citations per response increased by approximately 73%

55. ChatGPT query fan-out increased 4.6 times under higher reasoning

56. Minimal reasoning generated 245 web searches across the tested prompts

57. High reasoning generated 1,130 web searches across the same test

58. Only 25.6% of cited domains overlapped between minimal and high reasoning

59. High reasoning drew citations from 173 unique domains

60. 99 domains appeared under high reasoning that did not appear under minimal reasoning

One Prompt Can Create an Entire Search Journey

Conventional keyword analysis often begins with the assumption that one user query corresponds broadly to one search-results page.

Generative AI can operate differently. A complex prompt can trigger multiple retrieval actions as the system investigates products, organisations, attributes, evidence, comparisons and supporting information before constructing its response.

This means an organisation may become visible through a secondary or tertiary retrieval path even when it is not the obvious source for the user’s original wording.

Conversely, ranking strongly for the obvious keyword does not guarantee that the organisation will survive every stage of a deeper AI research process.

Reasoning Depth Changes the Competitive Source Set

The 25.6% domain overlap between reasoning modes is one of the most significant findings in this section.

It suggests that AI citation visibility can be more dynamic than a conventional fixed ranking position. Changes in reasoning depth can alter the searches performed, the sources encountered and the evidence ultimately selected.

For organisations measuring GEO performance, this creates a methodological challenge. Testing one prompt once cannot establish whether a brand has durable generative visibility.

More robust measurement may require repeated testing across prompt variants, user scenarios, reasoning depths, platforms and stages of the buyer journey.

Research Limitation

The study tested 100 prompts across 20 buyer journeys. Its results demonstrate substantial differences between the reasoning configurations examined but should not be interpreted as universal citation rates or retrieval patterns for every ChatGPT prompt.

Reasoning systems, retrieval methods and product interfaces can also change. These statistics should therefore be treated as evidence of how reasoning depth can affect retrieval and source selection rather than as permanent platform constants.

Evidence Source for Statistics 51–60

The principal evidence comes from Semrush and Kevin Indig’s 2026 research comparing ChatGPT’s minimal- and high-reasoning configurations across 100 prompts and 20 buyer journeys.

View Semrush — ChatGPT Reasoning & AI Visibility Study

Statistics 61–70

AI Citation Freshness, Content Recency & Source Eligibility

The rapid development of AI systems, products, research and markets creates another factor that can influence source selection: freshness. Information that was authoritative several years ago may remain historically valuable while becoming less useful for answering a current question.

Large-scale Ahrefs research across approximately 17 million citations from seven AI platforms found that AI systems tended to cite fresher content than conventional Google organic results. The evidence does not establish freshness as a universal ranking factor, but it indicates that recency can form part of the competitive environment for AI citation visibility.

61. AI-cited content was approximately 25.7% fresher than Google organic content

62. The freshness study analysed approximately 17 million AI citations

63. The citation-freshness research covered seven AI platforms

64. ChatGPT’s in-text reference URLs were approximately 393 days newer

65. ChatGPT citation URLs were approximately 458 days newer

66. ChatGPT citation URLs showed a 65-day larger freshness gap than its in-text references

67. More than one-third of Google AI Overview citations can come from pages outside Google’s organic top 100

68. 26.2% of Google AI Overview citations came from pages ranking between positions 11 and 100

69. 62.9% of AI Overview citations came from URLs outside Google’s organic top ten

70. An AI system was approximately three times more likely to cite a ranking domain than the exact ranking page in one ChatGPT study

Freshness Can Influence the AI Source Environment

The evidence in this section should not be reduced to the claim that newer content automatically outranks or replaces older content in AI systems.

Authoritative evergreen research, legislation, historical evidence, technical standards and primary documentation may remain highly valuable regardless of publication date.

What the data does indicate is that AI-cited content can display a meaningful freshness advantage over conventional organic results. For subjects where facts, products, prices, technologies, regulations or market conditions change quickly, maintaining current information may therefore strengthen source eligibility.

Freshness should consequently be understood alongside relevance, evidence, authority, entity clarity and corroboration rather than as a standalone optimisation tactic.

Content Maintenance Becomes a Visibility Discipline

For organisations with large content libraries, the research creates a practical question: how much existing material remains sufficiently current to support present-day AI answers?

Updating a page should not mean changing a date while leaving the underlying information untouched. Meaningful maintenance involves reviewing evidence, statistics, products, regulations, terminology, examples, references and conclusions to determine whether they still accurately represent the subject.

This is particularly relevant for research-led organisations because original studies, observations and statistics can become useful source material for AI systems when their methodology, publication date, evidence and scope are clearly communicated.

Research Limitation

A correlation between citation selection and fresher content does not prove that publication age itself caused the citation. Newer pages may also differ from older pages in relevance, structure, topical coverage, evidence or other characteristics.

The findings should therefore be interpreted as evidence of a freshness pattern within the measured citation environment rather than proof of a universal AI freshness ranking factor.

Evidence Sources for Statistics 61–70

The evidence combines Ahrefs’ large-scale analysis of content freshness across approximately 17 million AI citations with its research into AI Overview citation rankings and ChatGPT citation overlap with conventional Google results.

View Ahrefs — AI Citation Content Freshness Research

View Ahrefs — AI Overview Citation Rankings

View Ahrefs — ChatGPT Citations vs Google Rankings

Statistics 71–80

AI Referral Traffic, ChatGPT Growth & Where AI Sends Users

AI platforms increasingly function as discovery environments, but their impact extends beyond the answers displayed inside the interface. When users follow links from ChatGPT and other generative systems, AI becomes a measurable referral channel for external websites.

The available evidence shows rapid growth in ChatGPT referral traffic while also demonstrating that AI referrals remain highly concentrated and substantially smaller than Google as an overall source of website visits.

71. ChatGPT outbound referral traffic increased by approximately 206% during 2025

72. A 206% increase means ChatGPT referral traffic grew to approximately 3.06 times its starting level

73. More than 30% of measured ChatGPT referral traffic went to just ten domains

74. More than 20% of measured ChatGPT referral traffic went to Google

75. Google alone received at least two-thirds of the referral traffic captured by the top ten destinations

76. ChatGPT used its web-search feature on approximately 34.5% of measured queries in February 2026

77. Approximately 65.5% of measured ChatGPT queries did not trigger the web-search feature

78. Computer and technology destinations accounted for 36% of UK ChatGPT outgoing referral activity in Ofcom’s analysis

79. News and media accounted for 9.0% of ChatGPT outgoing traffic compared with 3.9% from Google

80. Science and education represented 6.9% of ChatGPT outgoing traffic compared with 2.1% from Google

AI Referral Traffic Is Growing — But Distribution Matters

A 206% increase in ChatGPT outbound referral traffic demonstrates that generative AI is developing into a measurable acquisition channel.

However, rapid percentage growth should be interpreted alongside the channel’s starting scale and the concentration of traffic among leading destinations. Earlier evidence in this report showed that Google remained approximately 190 times larger than ChatGPT as a source of website traffic in the Ahrefs dataset.

AI referral traffic is therefore important because of its growth and its role in emerging discovery journeys — not because it has already replaced conventional search traffic.

The concentration of referrals also suggests that source selection and destination authority can have a disproportionate effect on which websites benefit from AI-mediated discovery.

The Search Journey Is Becoming Cross-Platform

The finding that more than one-fifth of measured ChatGPT outbound referral traffic went to Google is particularly revealing.

It indicates that some users do not choose between AI and conventional search once and remain within that environment. Instead, they can move between platforms as the information need develops.

A user might ask ChatGPT for an initial explanation, move to Google to investigate a company or product, return to an AI assistant for comparison and finally reach a website through branded search or direct navigation.

This makes attribution more difficult. The platform responsible for the final website session may not be the platform that originally created awareness or shaped the user’s shortlist.

AI and Google Can Send Different Types of Referral Traffic

The Ofcom/Similarweb category evidence suggests that outgoing traffic from ChatGPT is distributed differently from outgoing traffic from Google.

These percentages compare the composition of outgoing traffic, not absolute visitor numbers. Google may still send substantially more total traffic to a category even where that category represents a smaller percentage of Google’s much larger referral base.

Research Limitation

The Semrush referral analysis uses US clickstream data and should not be presented as a direct measurement of UK referral behaviour. The Ofcom/Similarweb evidence provides UK context but measures outgoing traffic categories rather than total search demand.

Referral traffic also captures only interactions that result in an external visit. It cannot measure every occasion on which an AI answer influences awareness, trust, consideration or a later purchase.

Evidence Sources for Statistics 71–80

This section combines Semrush and Kevin Indig’s large-scale analysis of ChatGPT referral and search behaviour with Ofcom’s UK reporting based on Similarweb data.

View Semrush — ChatGPT Search & Referral Traffic Research

View Ofcom — The Era of Answer Engines

Statistics 81–90

Commercial Search, Ecommerce, Research Behaviour & AI Recommendation Visibility

The commercial importance of AI search depends on more than the number of people using ChatGPT, Gemini or other assistants. Search has traditionally influenced purchasing because consumers use it to research products, services, organisations and alternatives before making decisions.

UK evidence shows that online search remains deeply embedded in product research, while referral data indicates that AI assistants are beginning to participate in ecommerce, health, education, news and other discovery journeys. The result is an increasingly fragmented path between initial research and final conversion.

81. 95% of UK adult internet users searched online for information during the previous three months

82. 71% used online search when researching products or purchases

83. 70% searched online for UK news

84. 58% searched online for hobbies and interests

85. 53% of UK adults reported often seeing AI-generated summaries in search

86. Ecommerce and shopping accounted for 4.5% of ChatGPT outgoing traffic in Ofcom’s comparison

87. Ecommerce and shopping represented 2.2% of Google’s outgoing traffic in the same comparison

88. Ecommerce represented approximately twice the share of outgoing traffic from ChatGPT as from Google

89. Health accounted for 4.2% of ChatGPT outgoing traffic compared with 3.0% from Google

90. Comparative content generated 2.4 times more brand mentions in measured AI answers

AI Search Is Moving Into Commercial Discovery

Commercial search has traditionally been described through keywords such as “best”, “reviews”, “compare”, “near me”, product names and service categories. Generative interfaces can combine many of these research steps into a conversational journey.

A user can describe a problem, add requirements, reject unsuitable options, request comparisons and ask for recommendations without repeatedly returning to a conventional search-results page.

This creates a potentially important change in commercial visibility. The competitive question is no longer limited to which company ranks for a purchase-intent keyword. It can also include which organisations enter the AI system’s candidate set and remain present as the user’s requirements become more specific.

For retailers, service businesses and other commercial organisations, recommendation visibility may therefore become an increasingly important complement to conventional ranking visibility.

From Search Ranking to Recommendation Eligibility

Traditional SEO remains important because users continue to conduct enormous volumes of conventional searches and AI systems themselves can retrieve information from the web.

However, commercial AI visibility introduces an additional requirement: the system must understand enough about an organisation, product or service to determine when it is relevant to a particular user scenario.

That can require clear entity information, product and service attributes, pricing or commercial information where appropriate, geographical relevance, independent validation, evidence of expertise, customer-use context and consistent information across external sources.

The objective becomes not merely ranking for a phrase, but building enough machine-readable and externally corroborated evidence for the organisation to remain eligible when AI systems construct a shortlist.

What Businesses Need to Measure

As commercial discovery becomes more fragmented, organisations may need to measure performance across both conventional and generative search environments.

• Organic search rankings and traffic.
• Commercial keyword visibility.
• AI brand mentions.
• AI citation visibility.
• Product and service recommendation visibility.
• Comparison-query visibility.
• Share of AI answers mentioning the organisation.
• AI referral traffic.
• Branded search demand.
• Direct traffic following research activity.
• Qualified leads and conversions influenced by AI discovery.

Evidence Sources for Statistics 81–90

The evidence in this section combines Ofcom’s UK search-behaviour reporting, Ofcom’s analysis of outgoing ChatGPT and Google traffic using Similarweb data, and Semrush research examining citation and brand-mention behaviour across generative search platforms.

View Ofcom — Online Nation 2025

View Ofcom — The Era of Answer Engines

View Semrush — AI Citations & Brand Mentions Research

Statistics 91–100

AI Search Market Outlook 2026, Key Findings & Final Conclusions

The evidence across the first 90 statistics shows a search market undergoing structural change rather than a simple replacement of Google by standalone AI assistants.

Google remains enormous, ChatGPT has established substantial direct usage, Gemini is expanding, generative answers are becoming part of mainstream search, and AI systems increasingly influence which sources, brands and organisations users encounter during research. The final ten statistics bring these developments together and establish the scale of the emerging hybrid search market.

91. Google AI Overviews have more than 2.5 billion monthly active users globally

92. Google AI Mode has more than 1 billion monthly active users globally

93. AI Overviews have at least 2.5 times the reported monthly active audience of AI Mode

94. Google says AI Mode query volume has more than doubled every quarter since launch

95. Google Search still accounted for approximately 40% of measured website traffic in Ahrefs’ study

96. ChatGPT accounted for approximately 0.21% of measured website traffic in the same dataset

97. Google generated approximately 190 times more measured website traffic than ChatGPT

98. Approximately 65% of ChatGPT activity was classified as search or search-like under Ahrefs’ methodology

99. ChatGPT search-like activity was estimated at approximately 11.86% of Google search volume under the Ahrefs methodology

100. AI search can already influence a much larger share of discovery than AI referral traffic alone suggests

Key Findings from the 100 Statistics

AI Search Market Outlook for 2026

The available evidence does not support declaring a single percentage for AI’s share of the entire search market. The platforms, studies and methodologies measure different behaviours.

What can be established is that generative AI has become a significant part of information discovery. Standalone AI platforms have attracted large audiences, Google has integrated generative answers directly into Search, and AI systems increasingly retrieve and cite information from across the web.

This creates a hybrid search environment rather than two completely separate markets. Users can move between Google, ChatGPT, Gemini, AI Overviews, websites, social platforms, publishers and direct brand destinations during the same research journey.

The organisations best positioned for this environment will therefore need visibility across multiple discovery surfaces rather than depending on one ranking system or one traffic source.

What This Means for SEO and GEO

SEO remains important because conventional search remains enormous and many of the technical, content and authority signals developed through strong SEO also support machine discovery.

Generative Engine Optimisation adds another layer. Organisations increasingly need to consider whether AI systems can identify them, retrieve their information, understand their expertise, corroborate their claims and consider them relevant during comparison or recommendation.

The objective is therefore not to replace SEO with GEO. It is to build search authority capable of operating across conventional search results and generative discovery systems.

That requires strong technical foundations, useful content, entity clarity, original evidence, external authority, accurate information, citation eligibility and sufficient brand signals for an organisation to be recognised as a credible candidate within AI-mediated discovery.

How CGO Media Interprets AI Search Market Statistics

AI search research currently uses multiple methodologies. Studies can measure platform visits, monthly audiences, prompts, search-like activity, web retrieval, citations, brand mentions, referral traffic, click behaviour or conventional search-engine share.

CGO Media does not treat these measurements as interchangeable. Statistics are interpreted according to the population, geography, time period and methodology used by the original source.

Where CGO Media calculates a percentage, ratio or comparison from published source figures, it is identified as a CGO Media calculation rather than presented as a statistic published directly by the underlying source.

This approach allows evidence from adoption, behaviour, clicks, citations, retrieval and traffic studies to be considered together without pretending that each study measures the same phenomenon.

Research Review & Update Policy

AI search is changing rapidly. Platform audiences, interfaces, citation behaviour, retrieval systems and search integrations can change materially within months.

CGO Media reviews its research pages as new primary evidence, platform disclosures and substantial independent studies become available. Earlier figures may be retained where they provide useful historical context, but their dates and scope should remain visible.

For further information about how CGO Media evaluates evidence, visit the

Research Methodology

page.

Evidence Sources for Statistics 91–100

The final statistics use Google’s first-party reporting on AI Overviews and AI Mode together with Ahrefs research comparing Google, ChatGPT, website referral traffic and search-like activity. Global figures are identified as global and should not be interpreted as UK-specific measurements.

View Google Search Updates

View Google 2026 AI Updates

View Ahrefs — ChatGPT vs Google Search & Traffic

CGO Media Research

References & Evidence Sources

The 100 statistics in this report draw on primary platform disclosures, regulatory research, large-scale clickstream datasets, behavioural studies and independent analyses of AI search, citations, rankings and referral traffic.

CGO Media distinguishes between statistics reported directly by the original source and calculations produced from published figures. Where a percentage, ratio or comparison has been calculated by CGO Media, it is identified within the relevant statistic.

1. Ofcom — Adults’ Media Use and Attitudes 2026

UK regulatory research covering media behaviour and adoption of AI tools among adults.

Used for: UK AI-tool adoption, including overall adult usage and adoption among younger age groups.

Ofcom — Adults’ Media Use and Attitudes

2. Ofcom — The Era of Answer Engines

Ofcom analysis examining generative AI’s impact on search experiences and online behaviour, incorporating evidence from Ipsos iris and Similarweb.

Used for: UK chatbot reach, ChatGPT reach and outgoing traffic-category comparisons between ChatGPT and Google.

Ofcom — The Era of Answer Engines

3. Ofcom — Online Nation 2025

Ofcom’s annual analysis of UK online behaviour, platforms and internet use.

Used for: UK ChatGPT visit growth, Google Search usage, AI-summary exposure and information-search behaviour including product research, news, hobbies and interests.

Ofcom — Online Nation 2025

4. Pew Research Center — Google AI Summaries and Search Behaviour

Pew Research Center behavioural analysis published in July 2025, based on 900 US adults and 68,879 Google searches.

Used for: AI-summary prevalence, traditional-result click rates, citation clicks, search-session endings and source behaviour within Google AI summaries.

Pew Research Center — Google Users Are Less Likely to Click When an AI Summary Appears

5. Statcounter GlobalStats — Search Engine Market Share

Statcounter browser and device measurement data tracking conventional search-engine market share.

Used for: Google and Bing conventional search-engine market-share comparisons.

Statcounter — Search Engine Market Share

6. Statcounter GlobalStats — AI Chatbot Market Share

Statcounter’s measurement of AI chatbot market-share activity.

Used for: ChatGPT and Gemini AI chatbot share comparisons.

Statcounter — AI Chatbot Market Share

7. Similarweb — Generative AI Platform Traffic

Similarweb digital traffic intelligence covering visits to major generative AI platforms.

Used for: generative AI website-visit distribution, including ChatGPT, Gemini and Claude traffic comparisons.

Similarweb — Generative AI Statistics

8. Ahrefs — AI Overviews and Organic Click-Through Rates

Ahrefs analysis of approximately 300,000 informational keywords examining the relationship between AI Overviews and organic click-through rates.

Used for: estimated CTR effects for positions one, two and ten when AI Overviews appear.

Ahrefs — AI Overviews and Organic CTR Research

9. Ahrefs — ChatGPT vs Google Search & Website Traffic

Ahrefs analysis covering approximately 76,000 websites and comparing Google traffic with ChatGPT traffic and estimated search-like activity.

Used for: Google’s share of measured website traffic, ChatGPT referral traffic, the approximately 190× traffic difference, search-like ChatGPT activity and comparison with Google search volume.

Ahrefs — ChatGPT vs Google Search & Traffic

10. Ahrefs — AI Citation and Search Ranking Overlap

Ahrefs research comparing AI citations with conventional Google and Bing search rankings across approximately 15,000 prompts.

Used for: approximate overlap between AI-cited URLs and Google or Bing top-ten organic results.

11. Ahrefs — ChatGPT Citations vs Google Rankings

Ahrefs analysis of 3,311 short-tail terms comparing exact-URL and domain-level overlap between ChatGPT citations and Google’s organic top ten.

Used for: approximately 10% exact-URL overlap, 31.8% domain overlap and the distinction between page-level and domain-level citation visibility.

12. Ahrefs — Google AI Overview Citation Rankings

Large-scale Ahrefs research examining approximately 863,000 keyword SERPs and four million URLs cited by Google AI Overviews.

Used for: the 37.1% of citation URLs ranking in the organic top ten, 26.2% ranking between positions 11 and 100, and 36.7% outside the organic top 100.

13. Ahrefs — AI Citation Content Freshness

Ahrefs analysis of approximately 17 million citations across seven AI platforms examining the relationship between citation selection and content freshness.

Used for: the estimated 25.7% freshness difference between AI-cited and Google organic content and ChatGPT freshness comparisons of approximately 393 and 458 days.

14. Semrush / Kevin Indig — AI Citations & Brand Mentions

Research covering 3,981 domain appearances across 115 prompts and 14 countries, examining ChatGPT, Google AI Overviews, AI Mode and Gemini.

Used for: ghost citations, citation rates, brand-mention rates, ChatGPT and Gemini differences and the relationship between comparative content and brand mentions.

Semrush — Ghost Citations Study

15. Semrush / Kevin Indig — ChatGPT Reasoning & Source Selection

Research comparing minimal- and high-reasoning configurations across 100 prompts and 20 buyer journeys.

Used for: citation-rate changes, citations per response, 4.6× query fan-out, 245 versus 1,130 web searches, 25.6% domain overlap, unique-domain counts and high-reasoning-only sources.

16. Semrush / Kevin Indig — ChatGPT Search & Referral Behaviour

Analysis of more than one billion lines of US clickstream data examining ChatGPT referral growth, destination concentration and use of web search.

Used for: 206% outbound referral growth during 2025, concentration among leading domains, referrals to Google and the approximately 34.5% web-search-feature rate measured in February 2026.

17. Google — Search & AI Product Updates

First-party Google reporting on the expansion of generative AI within Google Search.

Used for: global AI Overview and AI Mode audience figures, reported AI Mode query growth and contextual information about Google’s AI search ecosystem.

Google — Search Product Updates

Google — AI Updates

Research Methodology

CGO Media reviews evidence according to what each dataset actually measures. Studies based on UK adults are used to describe UK behaviour; US behavioural studies are identified as US evidence; global platform disclosures are described as global figures.

The report does not convert website traffic share into search-query share, chatbot traffic into conventional search-engine market share, referral-category percentages into prompt percentages, or citation rates into market-share estimates.

Where multiple studies address related behaviour, they are used to build a broader picture of the search ecosystem without assuming that the underlying measurements are directly equivalent.

Where calculations are derived from published figures — including percentage changes, ratios or combined shares — they are labelled as CGO Media calculations.

A fuller explanation of the research process is available on the

CGO Media Research Methodology

page.

Research Review & Update Policy

The AI search market changes rapidly. CGO Media reviews this report as new regulatory evidence, platform disclosures, behavioural studies and independent research become available.

Earlier statistics may remain in the report where they provide useful evidence of market development over time. Publication dates, study populations and methodologies should therefore be considered when comparing figures.

Material changes to major statistics, calculations or conclusions should be incorporated into future revisions of the page.