Category: GEO

AI search runs on websites that brands do not own. Wikipedia and Reddit together took 25.1% of ChatGPT’s US citations in January and February 2026 (Similarweb), and in a University of Toronto study of 143,249 AI answers, third-party websites earned 75.0% of citations while the brand’s own site earned 11.1% (Chen and Koudas, 2026).

For this report, we compiled citation data published between July 2025 and October 2026, covering March 2025 to September 2026. The sources are Similarweb’s US citation panel, Pew Research Center’s analysis of 68,879 Google searches, three peer-reviewed or preprint academic studies with more than 850,000 citations between them, and one agency dataset of 155,312 citations that we use only for engines no other source covers.

Our position is that the most-cited websites in AI search are the places where people other than the brand describe it. Each engine draws from a different pool of those sites, and the Methodology section at the end lists every dataset.

Most-Cited Websites in ChatGPT

In the table below, we show the 15 domains ChatGPT cited most often in the US, as a share of all its citation events. Similarweb recorded these citations from ChatGPT’s web browsing mode across its monitored prompts.

Most-Cited Websites in ChatGPT, US (January to February 2026)

Rank	Website	Citations	Share of ChatGPT citations
1	wikipedia.org	118,285	13.15%
2	reddit.com	107,680	11.97%
3	openai.com	55,876	6.21%
4	walmart.com	26,118	2.90%
5	youtube.com	23,976	2.67%
6	linkedin.com	21,736	2.42%
7	reuters.com	20,451	2.27%
8	nih.gov	19,962	2.22%
9	google.com	19,478	2.17%
10	media-amazon.com	17,477	1.94%
11	wikimedia.org	17,329	1.93%
12	facebook.com	15,813	1.76%
13	ebay.com	15,730	1.75%
14	amazon.com	15,343	1.71%
15	github.com	14,569	1.62%

Source: Similarweb, AI Citation Analysis, US, January to February 2026.

Our team took three points from this data:

• We found that Wikipedia and Reddit took 25.1% of ChatGPT’s citations between them, more than the next eight domains combined.
• ChatGPT’s third most-cited domain was its own publisher, openai.com, ahead of Reuters, NIH, and every retailer.
• Counting wikimedia.org, Wikipedia’s projects reached 15.1% of ChatGPT’s citations.

Most-Cited Websites in Google AI Mode

In the table below, we show the 10 domains Google AI Mode cited most often over the same period and prompt set.

Most-Cited Websites in Google AI Mode, US (January to February 2026)

Rank	Website	Citations	Share of AI Mode citations
1	fandom.com	42,332	7.16%
2	wikipedia.org	30,792	5.21%
3	youtube.com	29,032	4.91%
4	reddit.com	24,764	4.19%
5	google.com	16,878	2.85%
6	facebook.com	14,410	2.44%
7	amazon.com	10,886	1.84%
8	nih.gov	8,821	1.49%
9	github.com	8,728	1.48%
10	apple.com	8,127	1.37%

Source: Similarweb, AI Citation Analysis, US, January to February 2026.

What the AI Mode data shows:

• We found that AI Mode spreads its citations more widely than ChatGPT: its top domain took 7.16%, about half of Wikipedia’s share in ChatGPT.
• Google’s own properties are prominent here too, with google.com fifth and YouTube, which Google owns, third.
• Wikipedia and Reddit are the only two domains in the top four of both engines.

Most-Cited Websites in Google AI Overviews and Other Engines

Pew Research Center tracked the Google searches of 900 US adults in March 2025 and recorded the sources linked in each AI summary. In the table below, we show what Pew found about the sources Google’s AI Overviews cite.

Sources Cited in Google AI Summaries, US (March 2025)

Measure	AI summaries	Standard search results
Wikipedia, YouTube, and Reddit combined, share of linked sources	15%	17%
.gov websites, share of linked sources	6%	2%
Share that linked to a news website	5%	5%
Searches that produced an AI summary	18%	Not applicable

Source: Pew Research Center, 68,879 Google searches by 900 US adults, March 2025.

No first-party or academic source publishes a recent domain ranking across Gemini, Perplexity, and AI Overviews together. For that comparison, we use one agency dataset, Suff Digital’s September 2026 study, and label it as such. In the table below, we show the most-cited website per engine on the same 2,984 questions.

Most-Cited Website by AI Engine (September 2026, agency data)

Engine	Most-cited website	Citations from that engine
Google AI Overviews	youtube.com	6,894
Google AI Mode	youtube.com	5,162
Perplexity	wikipedia.org	979
ChatGPT	nih.gov	954
Gemini	reddit.com	273

Source: Suff Digital, an SEO agency, 155,312 citations from 2,984 questions across five engines, September 2026.

Three points stand out across these engines:

• We found that government sites are three times as common in Google’s AI summaries as in its standard results (6% against 2%).
• In the agency data, ChatGPT cited YouTube 5 times on the same questions where AI Overviews cited it 6,894 times.
• No engine’s top source is a company’s own product or service site; each is a platform, an encyclopedia, or a government source.

Third-Party Websites Earn Most AI Citations

University of Toronto researchers collected 90 days of daily AI answers from ChatGPT, Gemini, and Perplexity on commercial questions in two markets, then sorted every cited website by who owns it. In the table below, we show how citations split between the brand being studied, its competitors, and everyone else.

Who Gets Cited in AI Answers: Own Site, Competitors, and Third Parties (June to September 2026)

Market	Answers	Brand’s own site	Competitors’ sites	Third-party sites
US home internet (consumer)	143,249	11.1%	13.8%	75.0%
Business software for architecture, engineering, and construction firms (niche B2B)	14,850	2.6%	46.7%	50.8%

Source: Chen and Koudas, University of Toronto (2026), ChatGPT, Gemini, and Perplexity via their APIs.

The same research group’s 2025 study compared AI search with Google on US automotive questions. Google returned 39.5% brand content, 15.4% social, and 45.1% earned media, while AI search returned 81.9% earned media and 18.1% brand content (Chen et al., 2025).

What the ownership data shows:

• We found that in the consumer market, a brand’s own site earned about one in nine citations, and third-party sites earned three in four.
• In the niche B2B market, competitors’ sites took 18 times the brand’s share, so a smaller vendor’s visibility depends on being mentioned on pages it does not control.
• The 2026 study found that every page ChatGPT cited had first been retrieved by its own search, and that retrieval explained 32% of the variation in citations against 2.8% for page features.

How Concentrated Are AI Citations?

A few domains take a large share of citations, and the share depends on the topic. In the table below, we compare concentration across four datasets.

Citation Concentration in AI Search (2025 to 2026)

Dataset	What was measured	Concentration
Yang (2025)	366,087 citations from OpenAI, Google, and Perplexity models, news sources only	Top 20 news outlets: 67.3% (OpenAI), 31.9% (Google), 28.5% (Perplexity)
Mental health citation audit (2026)	15,942 citations from ChatGPT, Perplexity, and Google AI Overviews	Top 10 domains: 43.6% of English citations
Similarweb	ChatGPT, US, Jan to Feb 2026	Wikipedia and Reddit: 25.1%
Suff Digital (agency data)	155,312 citations, five engines, Sept 2026	Top 20 websites: 26.2%; 14,616 websites cited at least once

Source: each dataset, linked in the table.

Our team took two points from this data:

• We found that in news, OpenAI’s models leaned hardest on a few outlets, with reuters.com at 22.8% and apnews.com at 12.2% of news citations (Yang, 2025).
• The long tail is still wide: thousands of websites earn at least one citation, so a specific page on a narrow question can be cited without being a household name.

What This Means for Brands

• Earn mentions where AI engines already look, starting with the Reddit threads that shape brand reputation in AI search.
• Plan for third-party coverage as much as owned content, through PR and strategic marketing work that earns placement on cited sites.
• Track each engine separately, because ChatGPT, AI Mode, and AI Overviews draw from different pools.
• Keep your own pages accurate and specific. Google says a page must be indexed and eligible to show with a snippet to appear as a supporting link, and needs no special markup (Google Search Central).
• Size the audience for each engine before splitting budget across them, using current AI search usage figures and Google AI Overviews statistics.

Methodology

This report compiles citation data available as of November 9, 2026. We included a dataset only if it stated its sample, its period, and how it counted a citation. We gave priority to first-party measurement and research firms (Similarweb, Pew Research Center) and to academic studies, and used one agency dataset only for the per-engine comparison that no other source covers.

In the table below, we list every dataset used.

Datasets Used in This Report

Dataset	Type	Sample	Period
Similarweb, AI Citation Analysis	Research firm panel	ChatGPT and Google AI Mode citation events across monitored US prompts	Jan to Feb 2026
Pew Research Center	Nonprofit research, browsing panel	68,879 Google searches by 900 US adults	March 2025
Chen and Koudas, University of Toronto	Academic preprint	158,099 answers, 468,037 citations, ChatGPT, Gemini, Perplexity	June to Sept 2026
Chen, Wang, Chen, and Koudas, University of Toronto	Academic preprint	Controlled queries across several product categories, AI search vs Google	2025
Yang	Academic preprint	366,087 citations from 12 AI search models	2025
Mental health citation audit	Academic preprint	15,942 citations from ChatGPT, Perplexity, Google AI Overviews	2026
Suff Digital	Agency study	155,312 citations from 2,984 questions, five engines	Sept 2026

Each source counts citations differently. Similarweb counts every citation event, Pew counts the sources linked in each AI summary, and the Toronto studies count each URL once per answer, so compare rankings within a dataset rather than percentages across datasets.

Frequently Asked Questions

What is the most-cited website in AI search?

It depends on the engine. Wikipedia was the most-cited domain in ChatGPT in the US at 13.15% of citations, and fandom.com led Google AI Mode at 7.16% (Similarweb, January to February 2026). Wikipedia, YouTube, and Reddit were the top three sources in Google’s AI summaries (Pew Research Center).

What sources does ChatGPT cite most?

Wikipedia (13.15%) and Reddit (11.97%), followed by openai.com (6.21%), walmart.com, and youtube.com, based on Similarweb’s US data for January and February 2026.

Do AI engines cite brand websites?

Less often than third-party sites. In a University of Toronto study of 143,249 answers, the brand’s own site earned 11.1% of citations and third-party sites 75.0%. In a niche B2B market, the brand’s own site earned 2.6%.

Do ChatGPT and Google AI Mode cite the same websites?

Only partly. Wikipedia and Reddit are in the top four of both, but fandom.com led AI Mode while ranking 19th in ChatGPT, and openai.com ranked third in ChatGPT (Similarweb).

How can a brand get cited in AI answers?

Get mentioned on the third-party pages each engine already retrieves for your category, and publish specific pages that answer your buyers’ questions directly. The Toronto study found that ChatGPT cites only pages its own search retrieves, and that pages whose titles and paragraphs match the question are cited more.

Talk to Genevate

To see which websites AI engines cite in your category, run your buyers’ top questions through ChatGPT, Gemini, Perplexity, and Google, and list the sources. Genevate is a New York City GEO agency for mid-market and enterprise brands. Contact Genevate to turn that list into a content and earned-mention plan.

Sources

1. Similarweb, “Most Cited Domains by LLMs” (2026)
2. Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results” (July 22, 2025)
3. Chen and Koudas, “What Drives AI Citations? Content, Retrieval and Model Knowledge in Generative Engine Optimization (GEO),” arXiv (2026)
4. Chen, Wang, Chen, and Koudas, “Generative Engine Optimization: How to Dominate AI Search,” arXiv (2025)
5. Yang, “News Source Citing Patterns in AI Search Systems,” arXiv (2025)
6. “Sources of Truth: A Multi-Platform, Multilingual Audit of Citations in AI Mental Health Information Queries,” arXiv (2026)
7. Suff Digital, “Most Cited Websites in AI 2026” (Sept 2026)
8. Google Search Central, “AI features and your website”

Visitors who click through from AI answers now convert better than the rest of a site’s traffic in most current data. In September 2026, AI traffic to US retail sites converted 90% better than non-AI traffic, the widest gap Adobe Digital Insights has recorded across more than 1 trillion visits. ChatGPT, the largest source of AI referrals, reached 1.2 billion weekly users the same month (The Verge, reporting OpenAI).

The AI referral numbers most teams see in Google Analytics 4 are still far too low. GA4 files every click from Google’s AI Overviews and AI Mode under Organic Search (Google Analytics Help), and Google says those features send billions of clicks to websites every week. Our estimate below puts the AI clicks GA4 can label at less than a tenth of the total.

For this report, we compiled first-party figures from OpenAI, Google, Adobe, Microsoft, Similarweb, and Pew Research Center, plus one peer-reviewed study and one multi-site GA4 benchmark. Each figure is quoted with its sample, period, and comparison channel, and the Methodology section lists every dataset.

AI Referral Conversion Rates: 2026 Benchmarks

In the table below, we compare the conversion rate of AI referral traffic with a stated comparison channel in five datasets. Each source defines a conversion in its own way, so the ratio inside each row is the number to compare.

AI Referral Conversion Rates by Dataset (2024 to 2026)

Dataset	Sample	Period	AI referral conversion	Comparison channel	Ratio
Adobe Digital Insights	1 trillion+ visits to US retail sites	Sept 2026	90% higher	All non-AI traffic	1.9x
Microsoft Clarity	1,277 publisher and news domains	One month, 2025	1.34% (subscriptions)	Search, 0.55%	2.4x
Similarweb	US desktop panel	Sept 2025	7% (ChatGPT)	Google, 5%	1.4x
The SEO Works	63 B2B and B2C sites in GA4	Sept 2025 to Aug 2026	7.18%	Organic search, 6.08%	1.18x
Kaiser and Schulze, Marketing Science	973 ecommerce sites in GA4	Aug 2024 to July 2025	Below organic search	Organic search, about 13% higher	About 0.9x

Source: each dataset, linked in the table. The Clarity and Kaiser and Schulze ratios are our arithmetic from published figures, rounded.

Our team took three points from this data:

• Four of the five datasets found AI referrals converting above the comparison channel, from 1.18x to 2.4x.
• The only dataset below parity covers August 2024 to July 2025, the year AI traffic was still converting worse in Adobe’s own data. Kaiser and Schulze also found the AI conversion rate rising through that year while traditional channels stayed flat.
• The smallest advantage came from the GA4 benchmark that includes B2B sites. In that study, B2B product sites were the one segment below parity, at 0.62x across ten sites (The SEO Works).

How AI Conversion Rates Changed From 2025 to 2026

Adobe has tracked AI traffic to US retail sites since October 2024, which makes it the longest consistent series available. In the table below, we show how the conversion gap between AI and non-AI traffic moved over 18 months.

AI vs. Non-AI Conversion Rate, US Retail Sites (2025 to 2026)

Month	AI traffic conversion compared with non-AI traffic	Source
March 2025	38% worse	Adobe, April 2026
August 2025	9% worse	Adobe, October 2026
March 2026	42% better	Adobe, April 2026
September 2026	90% better	Adobe, October 2026

Source: Adobe Digital Insights, based on more than 1 trillion visits to US retail sites. Non-AI traffic includes channels such as paid search and email.

Three points stand out in the Adobe series:

• The gap swung 128 points in 18 months, from 38% worse to 90% better.
• AI visitors in September 2026 also spent 73% more time on site and generated 83% more revenue per visit than non-AI visitors (Adobe).
• Studies built on 2024 and early 2025 data describe an earlier stage of the channel, so older “AI converts worse” findings no longer reflect current performance.

ChatGPT Referral Traffic Is Growing

ChatGPT referrals did not peak in 2025. First-party data shows ChatGPT’s audience, its search volume, and the clicks it sends to websites all growing through 2026.

In the table below, we show ChatGPT’s weekly active users as OpenAI has reported them.

ChatGPT Weekly Active Users (October 2025 to September 2026)

Date reported	Weekly active users	Source
October 6, 2025	800 million	TechCrunch, reporting Sam Altman
February 27, 2026	900 million	OpenAI
July 2026	1 billion	The Verge, reporting an OpenAI spokesperson
September 29, 2026	1.2 billion	The Verge, reporting OpenAI

Source: OpenAI figures as published or reported. These are company statements and are not independently audited.

OpenAI also says ChatGPT now handles about one-fifth as many search queries as Google, counting any question that triggers a web search (The Verge, Nov 3, 2026). In the US, 44% of adults said in February 2026 that they use ChatGPT, up from 18% in 2023 (Pew Research Center).

In the table below, we show what that audience growth means for referral traffic to websites.

AI and ChatGPT Referral Traffic Growth (2025 to 2026)

Dataset	What was measured	Change	Period
Similarweb	Referral visits from AI platforms, worldwide web	+117.4% year over year	June 2025 to May 2026
Similarweb	ChatGPT referrals after brand names became clickable	+157.7% week over week	May 7 to 13, 2026, vs prior week
Business Insider, reporting Graphite’s analysis of Similarweb data	Referral traffic from AI services worldwide, 60% of it from ChatGPT	+92% year over year	Year to fall 2026
Adobe	AI traffic to US retail sites	+71% year over year	September 2026
Microsoft Clarity	AI traffic to 1,200+ publisher and news sites	+155.6%	Eight months, 2025

Source: each dataset, linked in the table.

What the growth data shows:

• No dataset in this report shows AI or ChatGPT referral traffic shrinking. Every industry in Similarweb’s 2026 data at least doubled its AI referrals year over year.
• ChatGPT’s usage did flatten for about seven months after the first half of 2025, according to Graphite’s analysis. Growth resumed in 2026, and ChatGPT’s monthly visits reached a record 61.35 billion in September (Business Insider).
• ChatGPT’s May 7, 2026 change, which turned brand names in answers into links, moved referrals to a higher baseline that held in the weeks Similarweb tracked.

Google AI Mode and AI Overviews Traffic Is Hidden in Organic Search

The largest source of AI-driven website visits does not appear as AI traffic in any standard report. Google’s own documentation defines GA4’s Organic Search channel as non-ad search clicks “including Google’s AI Overviews and AI Mode,” and says the AI Assistant channel added in May 2026 excludes them (Google Analytics Help).

In the table below, we show where each type of AI-influenced visit lands in a default GA4 property.

Where AI-Influenced Visits Appear in GA4 (November 2026)

How the visitor arrives	GA4 default channel	Reported as AI traffic?
Clicks a link in ChatGPT, Gemini, Copilot, or a similar assistant	AI Assistant	Yes
Clicks a link in a Google AI Overview	Organic Search	No
Clicks a link in Google AI Mode	Organic Search	No
Clicks a link in an AI app that strips the referrer	Direct	No
Searches the brand on Google after an AI recommendation	Organic or Paid Search	No
Types the brand’s URL after an AI recommendation	Direct	No

Source: Google Analytics Help, Default channel group; Similarweb on referrer-less AI visits.

Search Console does not close the gap. AI Overview and AI Mode clicks are counted inside the “Web” search type (Google Search Central), and the Generative AI performance report that reached all sites on Aug. 31, 2026 shows impressions only (Search Console Help). Google’s Gary Illyes said on Nov. 2, 2026 that Google has nothing to announce yet about adding AI click data (Search Engine Journal).

In the table below, we compare the scale of Google’s AI features with the AI referral traffic analytics tools can identify.

Google AI Features Compared With Measurable AI Referrals (2026)

Measure	Figure	Source
AI Overviews monthly active users	More than 2.5 billion	Google, May 19, 2026
AI Mode monthly active users	More than 1 billion	Google, July 22, 2026
Clicks to websites from AI features in Google Search	“Billions” every week	Google, July 22, 2026
Google AI feature clicks per month, conservative estimate	At least 8.7 billion	Genevate estimate (2 billion a week x 52 weeks / 12 months)
Referral visits from all standalone AI platforms per month	770.7 million	Similarweb, June 2025 to May 2026 average

Source: Google statements and Similarweb, linked in the table. The estimate row is our arithmetic.

We drew three conclusions from this comparison:

• On the most conservative reading of “billions,” Google’s AI features send at least 11 times as many clicks to websites as every standalone AI assistant combined, and GA4 labels all of them Organic Search.
• That puts the AI clicks GA4 can identify at less than 10% of the clicks AI answers send to websites (0.77 billion of at least 9.4 billion a month).
• Our estimate covers clicks only. AI answers that lead to a later brand search or a typed URL add more visits that no tool attributes to AI.

Why Measured AI Conversions Understate AI’s Value

AI answers often send buyers to a brand without a click. Similarweb found that people who get a brand recommendation from ChatGPT are 2.5 times more likely to visit that brand’s site within seven days than an unrecommended competitor’s site, and 55.9% of those AI-influenced visits arrived through search, compared with 40.4% of visits with no AI influence (Similarweb, US desktop, July to December 2025).

Those visits convert inside Organic Search, Paid Search, and Direct, where they raise the apparent performance of those channels. Kaiser and Schulze make the same point in Marketing Science: ChatGPT may play a role in discovery that last-click data cannot capture (Kaiser and Schulze). The GA4 benchmark from The SEO Works also says its AI figures should be read as a floor, because AI Overview clicks cannot be separated (The SEO Works).

Our position is that the conversion rates in this report are accurate for the traffic they measure, and that the AI channel’s total contribution to leads and revenue is larger than any current analytics tool reports. Our AI visibility to revenue guide covers how to connect the two.

How to Measure AI Referral Conversions in GA4

You can get closer to the real number with four changes.

1. Report the AI Assistant channel next to Organic Search, with conversion rates and session counts for both. GA4 began assigning the “ai-assistant” medium on May 13, 2026 (Google Analytics release notes), so compare periods after that date.
2. Add any assistant your audience uses that the default list misses to a custom channel group, using the referrer domain.
3. Track AI Overview and AI Mode impressions in the Search Console Generative AI performance report, and set them beside organic clicks and conversions for the same pages.
4. Add a required “How did you hear about us?” field to your contact form with AI assistants as an option, so AI-influenced leads that arrive through search or direct still get counted.

What This Means for Marketing Teams

• Treat AI as a growing acquisition channel. Every first-party dataset in this report shows ChatGPT and AI referral traffic rising through 2026.
• Expect AI visitors to convert above your organic baseline, and test it on your own site with session counts reported next to each rate.
• Read Organic Search as partly AI traffic. Clicks from AI Overviews and AI Mode are counted there, so a flat organic line can hide growth in AI-driven visits.
• Our AI search statistics report and Google AI Overviews statistics cover the adoption data behind these trends.
• B2B teams can review how Genevate works with technology and B2B companies, and ecommerce teams with consumer brands.

Methodology

This report compiles published data available as of November 9, 2026. We included a dataset only if it stated its sample, its time period, and, for conversion figures, its comparison channel. We gave first-party platform data (OpenAI, Google, Adobe, Microsoft) and independent measurement firms (Similarweb, Pew Research Center) priority over agency and vendor studies, and used one agency benchmark for its B2B segment.

In the table below, we list every dataset used.

Datasets Used in This Report

Dataset	Type	Sample	Period
OpenAI (direct and as reported by TechCrunch and The Verge)	Company-reported usage	ChatGPT users worldwide	Oct 2025 to Nov 2026
Google (I/O 2026 keynote, Q2 2026 earnings, product documentation)	Company-reported usage and documentation	Google Search users worldwide	May to Nov 2026
Adobe Digital Insights	First-party analytics	1 trillion+ visits to US retail sites	Oct 2024 to Sept 2026
Microsoft Clarity	First-party analytics	1,277 publisher and news domains	2025
Similarweb	Clickstream panel	Worldwide web and US desktop panels	June 2025 to May 2026
Pew Research Center	National survey	US adults, Feb 17 to 23, 2026	2023 to 2026
Kaiser and Schulze, Marketing Science 45(4)	Peer-reviewed study, GA4 data	973 ecommerce sites	Aug 2024 to July 2025
The SEO Works	Agency benchmark, GA4 data	95 sites (63 with conversion data)	Sept 2025 to Aug 2026

The Google click estimate assumes the lowest plausible value for Google’s statement that AI features send “billions of clicks to websites every week,” which is 2 billion a week, or about 8.7 billion a month. Google’s figure dates from July 2026 and covers all countries, while Similarweb’s 770.7 million figure is a 12-month worldwide average for web traffic only. The two are not measured the same way, so we treat the 11x result as an order-of-magnitude estimate.

Frequently Asked Questions

What is the conversion rate of AI referral traffic?

In current data, AI referrals convert above the rest of a site’s traffic. Adobe found AI traffic to US retail sites converted 90% better than non-AI traffic in September 2026, and Similarweb found ChatGPT visitors converted at 7% versus 5% for Google visitors.

Is ChatGPT referral traffic growing or falling?

It is growing. ChatGPT reached 1.2 billion weekly users in September 2026, up from 800 million in October 2025, and Similarweb measured AI referral visits up 117.4% year over year, with ChatGPT the largest source.

Does Google Analytics show traffic from Google AI Mode and AI Overviews?

No. Google’s default channel group documentation places clicks from AI Overviews and AI Mode in Organic Search, and the AI Assistant channel excludes them. Search Console reports AI feature impressions but not separate AI clicks.

Why does AI referral traffic look small in GA4?

GA4 labels only clicks from standalone assistants such as ChatGPT as AI traffic. Clicks from Google’s AI features, AI apps that strip the referrer, and brand searches that follow an AI recommendation are all credited to other channels, so the AI line shows a fraction of AI-driven visits.

Does AI traffic convert better than Google organic traffic?

In current data, yes. Similarweb measured 7% for ChatGPT visitors against 5% for Google visitors, and The SEO Works found 7.18% against 6.08% for organic search across 63 sites. Because Google’s organic channel now includes AI Overview and AI Mode clicks, part of Google’s own number is AI traffic too.

Talk to Genevate

If your AI referral line looks small, it is probably measuring a fraction of the real channel. Genevate is a New York City GEO agency for mid-market and enterprise brands, and we set up AI search measurement alongside the work that earns AI visibility. You can contact Genevate to see what AI search is already sending your business.

Sources

1. Adobe, “AI traffic now generates 83% higher revenue per visit for U.S. retailers heading into the holiday shopping season” (Oct 29, 2026)
2. Adobe, “U.S. retailers see surge in AI traffic, but many websites are not entirely readable by machines” (Apr 16, 2026)
3. OpenAI, “Scaling AI for everyone” (Feb 27, 2026)
4. The Verge, “ChatGPT now has 1.2 billion weekly users, OpenAI says” (Sept 29, 2026)
5. The Verge, “ChatGPT and Gemini both just passed 1 billion users” (Aug 11, 2026)
6. The Verge, “OpenAI says ChatGPT is catching up to Google in search” (Nov 3, 2026)
7. TechCrunch, “Sam Altman says ChatGPT has hit 800M weekly active users” (Oct 6, 2025)
8. Google, “I/O 2026: Welcome to the agentic Gemini era” (May 19, 2026)
9. Google, “Q2 2026 earnings call: Remarks from our CEO” (July 22, 2026)
10. Google Analytics Help, “Default channel group”
11. Google Analytics Help, “What’s new in Google Analytics”
12. Google Search Central, “AI features and your website”
13. Search Console Help, “Generative AI performance report (Search)”
14. Search Engine Journal, “Google Says No AI Click Reporting News Yet, Ahead Of UK Deadline” (Nov 2, 2026)
15. Similarweb, “AI Referral Traffic by Industry: What 770 Million Monthly Visits Reveal” (Sept 3, 2026)
16. Similarweb, “ChatGPT Referral Traffic Near Triples Overnight” (May 25, 2026)
17. Similarweb, “How to Optimize Your Homepage for AI Traffic” (Aug 9, 2026)
18. Business Insider, “ChatGPT’s growth is bouncing back as rivals struggle to close the gap” (Nov 9, 2026)
19. Microsoft Clarity, “AI Traffic Converts at 3x the Rate of Other Channels (Study)” (Nov 6, 2025)
20. Pew Research Center, “The share of Americans who say they use ChatGPT has more than doubled since 2023” (June 11, 2026)
21. Kaiser and Schulze, “ChatGPT Referrals to E-Commerce Websites,” Marketing Science 45(4)
22. The SEO Works, “How Much Traffic Do AI Assistants Actually Send, And Does It Convert Better?”

To choose a generative engine optimization (GEO) agency, score each candidate from 0 to 3 on eight criteria: AI search depth, measurement, proof of results, content and earned-mention execution, scope and contract clarity, leadership, industry fit, and honest claims. Google’s own guidance for hiring a search consultant adds a rule that applies to GEO too: no one can guarantee a #1 ranking, so treat any guaranteed placement as a reason to walk away (Google Search Central).

This guide gives you the scorecard and the questions to ask on a sales call, plus the red flags to watch for. Disclosure: Genevate is a GEO agency, so you may end up comparing us with others. The scorecard works the same way for every agency, including ours.

The GEO Agency Evaluation Scorecard

Each criterion gets a score from 0 (no evidence) to 3 (clear, checkable evidence). The table lists what to ask, what a strong answer sounds like, and what should worry you.

Criterion	Ask the agency	Strong answer (3)	Red flag (0)
AI search depth	Which AI platforms do you work on, and what do you change on a client’s site and off it?	Lists platforms such as ChatGPT, Perplexity, Google AI Overviews, and Copilot, and describes specific work for each	GEO is one line inside an SEO package
Measurement	How do you track AI visibility, and how do you tie it to leads?	Describes a prompt set, a tracking method, a reporting schedule, and a link to qualified leads	Reports only “mentions” with no method
Proof of results	Can I speak with past clients?	Offers references and describes results with dates and scope	Only logos, no references
Execution	Who writes the content, and how do you earn third-party mentions?	Lists the content and earned-media work included each month	Advice only, with all work left to your team
Scope and contract	What is the minimum term, and who owns the content and data if we leave?	States term, out clause, and ownership in writing	Will not discuss terms before you sign
Leadership	Who does the work, and what have they published?	Named team members with a public track record	No names, or work handed off after the sale
Industry fit	Which clients are like us?	Relevant industry pages or named, comparable clients	No work in your sector
Honest claims	What can you promise?	Promises a method and reporting, not a placement	Guarantees AI placements or a #1 result

Source: criteria are Genevate’s editorial framework, built from Google’s hiring guidance and the public service pages cited below.

Key Findings

• Eight scores of 3 give a maximum of 24, but a 0 on honest claims should end the search.
• Ask every agency the same questions in the same order, so the scores can be compared.

How to Weight the Scorecard

Not every criterion deserves the same weight. The weights below are our editorial suggestion for a mid-market or enterprise buyer, and you should change them to fit your goals.

Criterion	Suggested weight (points out of 100)	Why it carries that weight
AI search depth	20	GEO is the service you are buying
Measurement	15	Without a method, you cannot judge the work
Proof of results	15	Past work with references is the best predictor available
Execution	15	AI engines read content and third-party mentions, so someone has to produce both
Scope and contract	10	Terms decide how much risk you carry
Leadership	10	Senior people should do the work you were sold
Industry fit	10	Familiarity with your buyers shortens the learning curve
Honest claims	5	Low weight, but a 0 disqualifies the agency

Source: Genevate editorial suggestion. Multiply each 0 to 3 score by weight, divide by 3, and add the results for a score out of 100.

Key Findings

• The three heaviest criteria (AI search depth, measurement, and proof) add up to 50 points.
• Honest claims carry the fewest points because it works as a pass or fail test.

Questions to Ask a GEO Agency

Google’s guidance for hiring a search consultant lists questions that carry over to GEO (Google Search Central). Use these on the first call.

1. What results do you expect, in what timeframe, and how do you measure success?
2. What is your experience in my industry?
3. How long have you been in business, and who will do the work?
4. Will you share every change you make, with the reasoning behind it?
5. Can I contact past clients?
6. Which AI platforms do you track, and how many prompts?
7. How do you report AI referral traffic and conversions in GA4?
8. What is the minimum term, and is there an out clause?
9. Who owns the content and the tracking data if we leave?
10. Does your advice match Google’s official guidance on generative AI features?

Genevate publishes its own answers to several of these. Its FAQ describes a three-phase process (strategy in month 1, content execution from month 2, then ongoing measurement) and says engagements are ongoing retainers tailored to each client. Genevate runs those retainers month to month, with no long-term contract.

You can read how that applies to your category on the generative engine optimization services page.

Red Flags When Hiring a GEO Agency

Google lists the warning signs for search consultants. Each one applies to GEO agencies.

Red flag	What Google or the agency itself says
Guaranteed placement	“No one can guarantee a #1 ranking on Google” (Google Search Central)
Claimed special access	Beware of SEOs that allege a “special relationship” with Google or advertise a “priority submit” (Google Search Central)
Secrecy	“Be careful if a company is secretive or won’t clearly explain what they intend to do” (Google Search Central)
Required special files or markup	Google says no new machine-readable files, AI text files, or special schema are needed for AI Overviews and AI Mode (Google Search Central)
Link schemes	Google says to avoid SEOs that talk about link popularity schemes (Google Search Central)
Placement promises from GEO agencies	“No generative engine optimization agency, contractor, or freelancer can guarantee placements” (WebFX)

Source: Google Search Central, “Do You Need an SEO? Tips for Hiring an SEO” (updated June 5, 2026) and “Optimizing your website for generative AI features on Google Search”; WebFX service page.

Key Findings

• Google and at least one large agency agree that placement guarantees are a warning sign.
• An agency that sells AI-specific files or markup as a requirement is contradicting Google’s published guidance for its own AI features.

What the Contract Should Say

Agencies publish very different terms. The table shows three agencies that post them, plus Genevate’s own term.

Agency	Published term	Source
WebFX	6-month commitment	WebFX
Breaking B2B	6 or 12 months, with a lower monthly rate on the 12-month term	Breaking B2B
Siege Media	12 months with a 30-day out clause	Siege Media
Genevate	Month to month, no long-term commitment	Genevate

Source: each agency’s own page, checked Nov 5, 2026. The Genevate row states our own terms.

A long minimum term protects the agency more than the client. Genevate works month to month because we expect the results to be the reason clients stay. If the work stops producing measurable gains in AI visibility and qualified leads, a client can leave without waiting out a contract.

Before you sign, get four items in writing: the minimum term, the exit terms, the content and data ownership, and the reporting you will receive each month. Our GEO pricing guide compares published agency prices and engagement models if you are building a budget alongside this scorecard.

Ask How the Agency Measures Results

AI referral traffic is still small, so a good agency reports more than clicks. The SEO Works found that across 95 websites, organic search delivered 39 times more sessions than AI assistants from September 2025 to August 2026 (The SEO Works). Marie Haynes found that 45 of 100 sites she sampled tracked no key events at all (Marie Haynes).

Those two findings suggest three questions for any agency:

• Which conversions will we track, and are they set up before the work starts?
• How do you report AI-assisted visits that arrive through branded search or direct traffic?
• How do you report Google AI Overviews, which reach your site as ordinary organic sessions (The SEO Works)?

For background on the AI search numbers behind these questions, read our AI search statistics report and the Google AI Overviews statistics.

How Genevate Fits the Scorecard

Genevate is a New York City GEO agency that works with mid-market and enterprise brands. In our ranking of the best-value GEO companies in the US, Genevate ranks first and First Page Sage ranks second. The ranking is our editorial opinion, so score every agency yourself.

To see how Genevate answers the scorecard, review the generative engine optimization services, the about page, or contact Genevate and bring the ten questions above.

Frequently Asked Questions

How do I choose a GEO agency?

Score each candidate from 0 to 3 on AI search depth, measurement, proof of results, execution, scope and contract, leadership, industry fit, and honest claims. Ask all of them the same questions, and drop any that guarantee a placement. Google Search Central says no one can guarantee a #1 ranking.

What questions should I ask a GEO agency?

Ask about expected results and timeframe, experience in your industry, who does the work, which AI platforms they track, how they report conversions, contract terms, and data ownership. Google’s hiring guidance also suggests asking whether you can contact past clients.

Can a GEO agency guarantee that my brand appears in ChatGPT?

No. WebFX says no GEO agency, contractor, or freelancer can guarantee placements, and Genevate’s FAQ says no ethical GEO agency can guarantee specific AI placements. AI platforms decide what to show, and those decisions change as models update.

Does my GEO agency need special files or markup for Google AI features?

Google says no. Its documentation states that you do not need new machine-readable files, AI text files, or special schema.org markup to appear in AI Overviews and AI Mode (Google Search Central). Standard SEO fundamentals still apply.

How long should a GEO contract be?

Published terms range from month to month up to 12 months. WebFX asks for 6 months, Breaking B2B offers 6 or 12, and Siege Media uses 12 months with a 30-day out clause. Genevate works month to month with no long-term commitment. Ask for exit terms in writing.

Talk to Genevate

If you are comparing GEO agencies for next year, bring the scorecard and the ten questions to every call and score the answers the same way. Genevate is happy to be scored. You can contact Genevate to start the conversation, or review our services first.

Sources

1. Google Search Central, “Do You Need an SEO? Tips for Hiring an SEO”
2. Google Search Central, “Optimizing your website for generative AI features on Google Search”
3. WebFX, AI search optimization services
4. Breaking B2B, pricing
5. Siege Media, content marketing services
6. The SEO Works, AI referral traffic benchmark
7. Marie Haynes, “GEO is a distraction” (Oct 21, 2026)
8. First Page Sage, The Top Generative Engine Optimization (GEO) Agencies
9. Genevate, FAQ

An Expert Interview with Emily Connor, Founder and CEO of Driven Metrics

Driven Metrics is One of the Best GEO Agencies for Small to Mid-Sized Businesses

At Genevate, we help companies build authority, improve visibility in AI-generated answers, and generate qualified leads through GEO and content strategy. One of the most common questions we hear from clients is also the most important one: how do we know if this is actually working? Today, we’re sitting down with Emily Connor, Founder and CEO of Driven Metrics, one of the best GEO agencies for small to mid-sized businesses. Emily works with companies at the intersection of SEO, GEO, and performance marketing, and she has built rigorous frameworks for connecting AI visibility to real business outcomes. We asked her how companies should think about measuring GEO’s impact in a landscape where the tracking tools are still catching up to the channel.

Q: Emily, GEO is still an emerging discipline and measurement is genuinely difficult. What is the honest starting point for a company trying to understand whether their GEO investment is working?

A: The honest starting point is acknowledging that no single metric tells the full story, and that anyone claiming otherwise is overstating what current tools can prove. GEO operates across a range of AI platforms, ChatGPT, Perplexity, Google AI Overviews, Claude, and others, and most of those platforms do not expose referral data the way traditional search does. That means measurement requires a composite approach: prompt auditing to track citation frequency across platforms, branded search volume as a proxy for awareness, direct and dark social traffic patterns, AI-specific referral data where it is available, and lead quality analysis that looks at whether inbound inquiries are arriving more informed and further along in their evaluation than they were before GEO investment began. None of these is a clean attribution signal on its own. Together, they build a picture that is meaningful and directionally reliable.

Q: Walk us through prompt auditing. What is it, and why does it matter as a GEO measurement tool?

A: Prompt auditing is the practice of systematically querying AI platforms with the questions your target buyers are likely to ask and recording which brands, sources, and answers appear. It is the GEO equivalent of rank tracking in traditional SEO. If a company is investing in GEO, they should be running regular prompt audits across the AI platforms most relevant to their category and tracking whether their brand is being cited, in what context, and with what frequency relative to competitors. This tells you where you are winning AI visibility and where gaps exist. At Driven Metrics, we build structured prompt audit frameworks for every GEO client because citation frequency is the most direct signal available that GEO content is doing what it is supposed to do. It is not a revenue number, but it is the upstream indicator that everything else depends on.

Q: How does branded search volume connect to GEO performance, and why is it a useful proxy metric?

A: When AI systems cite a brand in response to a high-intent question, a meaningful portion of recipients who are further along in their evaluation will follow up with a branded search to learn more about the company they just heard referenced. That behavior shows up in Google Search Console as an increase in branded query volume, even though the original discovery happened in an AI environment that did not generate a direct referral click. Tracking branded search over time, particularly for non-navigational branded queries that suggest the searcher is evaluating the company rather than returning to it, gives you a measurable upstream signal of AI-driven discovery. It is an imperfect proxy, and it needs to be interpreted alongside other signals, but for companies working with Genevate on authority building and AI visibility, branded search trend is one of the clearest business-facing indicators that the GEO work is expanding top-of-funnel awareness.

Q: Lead quality is something GEO proponents often cite as an outcome. How do you actually measure that, and is it real?

A: It is real, and it is measurable, though it requires intentional tracking infrastructure to surface. Companies that have invested in GEO and thought leadership content consistently report that inbound leads arriving after AI-assisted research arrive more educated, ask better questions, and require less nurturing before a sales conversation becomes productive. The way to measure this is through lead intake data: tracking how prospects describe how they found the company, what they already know about the service, and how far along they are in their decision process at first contact. Over time, comparing the quality characteristics of leads arriving before and after GEO investment tells a genuine story about the channel’s contribution to pipeline quality, even when direct attribution is not available. Marketing-qualified lead rate, sales cycle length, and close rate by lead source are all metrics worth isolating and tracking against GEO activity timelines.

Q: What should a company realistically expect from GEO measurement in the first six to twelve months, and what patience does it require?

A: Expect to be building measurement infrastructure and establishing baselines in the first three to six months, and expect to start seeing directional signal, not conclusive proof, in months six through twelve. GEO compounds over time in the same way that SEO does: the content authority that earns AI citation today continues to earn it as new users query those platforms weeks and months later. Companies that enter GEO expecting the same attribution clarity they get from paid search will be disappointed, because the channel does not work that way. Companies that understand they are building a durable organic visibility asset, one that influences buyers during research phases that happen outside of tracked sessions, and that commit to the composite measurement framework needed to observe its impact, consistently find that the business case becomes clearer as the investment matures. Genevate’s approach to authority building and AI visibility, paired with the kind of performance tracking Driven Metrics brings to GEO strategy, is exactly the combination that makes that case over time.

To learn more about Driven Metrics and explore GEO and SEO strategy for small to mid-sized businesses, visit driven-metrics.com.

Google AI Overviews now reach more than 2.5 billion people a month, and they sit on top of a growing share of searches. This page collects the click-through rate (CTR) and zero-click figures from studies published between July 2025 and October 2026, each linked to its original source.

Key Findings

• Google showed an AI Overview on 39.4% of US desktop searches in June 2026, up from 25.8% in July 2025 (Comscore).
• In a 2025 study of real browsing, people clicked a regular search result on 8% of visits when an AI summary appeared and on 15% when none did (Pew Research Center).
• An AI Overview correlated with a 58% lower CTR for the top-ranking page in December 2025 (Ahrefs).
• 68.01% of US Google searches ended without a click from January to April 2026, up from 60.45% in 2024 (SparkToro).
• Organic CTR on queries with an AI Overview fell to 1.31% in December 2025, then rose to 2.36% by February 2026 (Seer Interactive).

For broader usage data on ChatGPT, Gemini and other assistants, our AI search statistics report covers user counts and market share, while this page covers clicks.

How Often Do AI Overviews Appear in Google Search?

Coverage depends on who measures it and which queries they sample. Panel data from real users gives the highest rates, while a sample of trending news topics gives a much lower one.

Measure	Latest figure	Period	Source
US Google desktop searches with an AI Overview	39.4% (from 25.8%)	Jul 2025 to Jun 2026	Comscore
US AI Overview appearance rate	43%+	Mid-2026	Similarweb
Trending-topic searches with an AI Overview	13.7%	Mar 13 to Apr 21, 2026	Washington University in St. Louis
Searches with an AI summary in a browsing panel	18%	Mar 2025	Pew Research Center
Bing searches with Copilot Search	17.3% (from 11.8%)	Jul 2025 to Jun 2026	Comscore

Source: publishers listed in each row.

Key Findings

• Comscore’s panel puts AI Overviews on about two in five US desktop searches.
• WashU found that 65% of question-form queries produced an overview, compared with 9.5% of other searches.
• The WashU team counted about eight cited sources in an average overview.

Query intent shapes coverage even more than wording. Across 53 brands, Seer Interactive found AI Overviews on 36% of informational queries, 8% of commercial queries, and 5% of transactional queries.

Query type	Share with an AI Overview
Comparison (X vs Y)	95.4%
Question (what, why, how, is, are)	85.9%
Price, cost, or buy	83.4%
Best of	81.3%
Near me	76.9%
Single word	27.3%

Source: Seer Interactive, Jan 2025 to Feb 2026.

Key Findings

• Comparison and question queries show an AI Overview most of the time in Seer’s data.
• One-word searches show one about a quarter of the time.

How Do AI Overviews Change Click-Through Rates?

Each study in the table below finds fewer clicks when an AI Overview appears. The size of the drop depends on method and timing.

Study	Sample	What it measured	With AI Overview	Without AI Overview
Pew Research Center, Mar 2025	900 US adults, 68,879 searches	Visits with a click on a regular result	8%	15%
Ahrefs, Dec 2025	300,000 keywords, Search Console data	Position 1 CTR vs forecast	58% lower	Baseline
Seer Interactive, Feb 2026	53 brands, 5.47M queries	Organic CTR	2.36%	3.82%
Seer Interactive, Feb 2026	Same	Paid CTR	16.21%	21.85%

Source: publishers listed in each row.

Key Findings

• Pew’s panel clicked a regular result on 8% of visits with an AI summary, close to half the 15% rate without one.
• Ahrefs measured top-ranking CTR on AI Overview keywords at 0.073 in December 2023 and 0.016 in December 2025.
• Paid CTR also runs lower on AI Overview pages in Seer’s data.

Pew also tracked what people did inside the summary. They clicked a link in the AI summary on 1% of visits, and they ended their browsing session on 26% of pages with a summary, compared with 16% of pages without one.

Did AI Overview CTR Recover in 2026?

Seer’s monthly data shows a partial rebound. Organic CTR on AI Overview queries fell from 3.19% in January 2025 to 1.31% in December 2025, then rose to 2.36% in February 2026.

Organic CTR with and without an AI Overview, Jan 2025 to Feb 2026. Source: Seer Interactive.

Key Findings

• Queries without an AI Overview gained clicks, from 2.75% CTR in January 2025 to 3.82% in February 2026.
• Two months of recovery is a short window, so treat the 2026 rebound as early data.

Does Being Cited in an AI Overview Bring Clicks?

Seer Interactive found that brands cited in an AI Overview earned 120% more organic clicks per impression than brands left out of it.

Those cited pages still trailed the no-overview baseline by 38% on informational queries.

Washington University in St. Louis also checked accuracy. The team reviewed 98,020 claims and found about 11% were unsupported by, or contradicted, the cited sources.

How Many Google Searches Are Zero-Click in 2026?

SparkToro found that 68.01% of US Google searches ended without a click from January to April 2026, using Similarweb clickstream data. Its 2024 figure was 60.45%, measured with a different panel.

Zero-click searches and AI Overview coverage. Sources: SparkToro (Similarweb panel); Comscore.

Key Findings

• The share of searches that produced any click fell 9.51 points over two years, a 22.9% decline, per SparkToro.
• SparkToro credits most of the shift to AI Overviews, since AI Mode took 0.34% of searches from January to April 2026.
• SparkToro’s data excludes the Google mobile app, where the authors expect zero-click rates to run higher.

What Does Google Say About Clicks?

Google’s statements on clicks differ from the third-party studies. In August 2025, Liz Reid, head of Google Search, wrote that “total organic click volume from Google Search to websites has been relatively stable year-over-year.”

The same post said Google was sending “slightly more quality clicks,” meaning visits where people do not quickly return to the results. In July 2026, Alphabet said it sends “billions of clicks to websites every week through AI features in Search.”

Google has not published a CTR figure for AI Overviews. Its statements cover total click volume across the web, while the studies above measure clicks per search or per page.

How Much Traffic Do AI Assistants Send?

Outside Google, AI assistants still send a small share of site visits. Across 115,254 sites, Ahrefs measured AI assistants at 0.43% of traffic in September 2026, compared with 26.82% from search engines.

Frequently Asked Questions

What percentage of Google searches show an AI Overview?

Comscore measured AI Overviews on 39.4% of US desktop Google searches in June 2026. Similarweb reported a US appearance rate above 43%, and Washington University in St. Louis found 13.7% on a sample of trending searches.

What is the click-through rate when an AI Overview appears?

Seer Interactive measured organic CTR at 2.36% on queries with an AI Overview in February 2026, against 3.82% without one. Pew Research Center found people clicked a regular result on 8% of visits with an AI summary and 15% without.

How much do AI Overviews reduce clicks?

Ahrefs found that an AI Overview correlated with a 58% lower CTR for the top-ranking page in December 2025. Results differ by study because each one uses a different sample and method.

What share of Google searches are zero-click?

SparkToro found that 68.01% of US Google searches ended without a click from January to April 2026. Its 2024 reading was 60.45%.

Do people click links inside AI Overviews?

Rarely. In Pew Research Center‘s 2025 study, people clicked a link inside the AI summary on 1% of visits to pages that showed one.

Does Google say AI Overviews reduce traffic?

Google says total organic click volume has stayed relatively stable and that click quality has gone up. It has not published a click-through rate for AI Overviews.

What to Do With These Numbers

If a large share of your organic traffic comes from informational or comparison queries, check how many of those queries now show an AI Overview and whether your pages are cited in them. Seer Interactive‘s data suggests a citation is worth tracking as its own metric next to rankings.

If you are new to the field, our guide to what generative engine optimization is covers the basics. Genevate’s generative engine optimization services focus on earning those citations in Google and in AI assistants, and our research center holds our other studies. To talk through your own search data, contact our team.

Sources

1. Google, “Sundar Pichai at I/O 2026,” May 19, 2026: https://blog.google/innovation-and-ai/sundar-pichai-io-2026/
2. Comscore, Q2 2026 AI Intelligence Report press release, Sep 22, 2026: https://ir.comscore.com/news-releases/news-release-details/comscores-latest-ai-intelligence-report-shows-competition
3. Similarweb, “AI Search Stats 2026,” Jul 29, 2026: https://aisearch.similarweb.com/blog/gen-ai-stats/
4. Washington University in St. Louis, AI Overview citation study, Sep 29, 2026: https://source.washu.edu/2026/09/washu-research-finds-gaps-in-google-ai-overview-citations-claims/
5. Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results,” Jul 22, 2025: https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
6. Ahrefs, “Update: AI Overviews Reduce Clicks by 58%,” Feb 4, 2026: https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/
7. Seer Interactive, “AIO Impact on Google CTR: 2026 Update”: https://www.seerinteractive.com/insights/aio-impact-on-google-ctr-2026-update
8. SparkToro, “In 2026, Less than One Third of Google Searches Still Send a Click,” Jun 8, 2026: https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/
9. Google, “AI in Search is driving more queries and higher quality clicks,” Aug 6, 2025: https://blog.google/products/search/ai-search-driving-more-queries-higher-quality-clicks/
10. Alphabet, Q2 2026 earnings remarks, Jul 22, 2026: https://blog.google/company-news/inside-google/message-ceo/alphabet-earnings-q2-2026/
11. Ahrefs, AI vs Search traffic benchmark: https://ahrefs.com/benchmarks/ai-vs-search

Among established generative engine optimization (GEO) agencies that publish a price in 2026, managed programs start between $3,000 and $10,000 per month. WebFX‘s 2026 cost guide puts agency GEO services between $1,500 and $50,000+ per month, and AI visibility tracking tools alone start at $50 to $95 per month on public plans.

Our analysts reviewed agency and tool pricing pages in November 2026 and quoted each price as the company states it. Many agencies quote only by proposal, so the prices below come from those that post a rate card.

How Much Does GEO Cost per Month?

Most GEO work is sold as a monthly retainer. WebFX‘s cost guide, published October 15, 2026, sorts agency pricing by company size.

Those bands are WebFX’s market estimate. They are not a survey with a stated sample, so treat them as a rough frame and compare them with the rate cards in the next section.

Buyer or engagement	Cost range	Billing
Small businesses (basic strategy)	$1,500 to $5,000 per month	Retainer
Mid-sized businesses (advanced strategies)	$5,000 to $25,000+ per month	Retainer
Enterprise businesses (complex strategies)	$25,000 to $50,000+ per month	Retainer
One-off projects (audits, migrations, overhauls)	$5,000 to $50,000 per project	Project
Hourly consulting	$50 to $300 per hour	Hourly
AI visibility tracking software	$50 to $1,000+ per month	Subscription

Source: WebFX, “How Much Does Generative Engine Optimization Cost in 2026?”, Oct 15, 2026.

Key Findings

• The widest gap sits in the mid-sized band, where the top of the range is five times the bottom.
• SEO.com, which WebFX runs, gives a narrower $1,500 to $10,000 per month range for answer engine optimization services.

Published GEO Prices From Agencies’ Own Pages

The table lists every agency price we could load on the company’s own site in November 2026. Where a price covers more than GEO, the scope column says so.

Agency	Published price (quoted)	Scope	Term stated
WebFX	“GEO Solutions Starting At $3,000/month”	GEO	6-month commitment
RevenueZen	Foundation “$3,000 / Month” to Total Market “$15,000 / Month”	SEO content program; top tier adds “Full GEO”	Not stated on tiers
Breaking B2B	“Light plan from $4,600/month”; Silver $6,500 (6 months) or $6,000 (12 months); Growth Team $9,500 or $9,000	Google and AI search	6 or 12 months
The Digital Elevator	Managed AEO/SEO “From $5,000/mo”; AI Visibility Mentions “From $3,495/mo”	AEO and SEO	Not stated
Obility	“SEO & GEO Core” $5,450/mo	SEO and GEO, B2B	Not stated
Siege Media	“Our minimums for content marketing are currently $8,000 per month.”	Content marketing	12 months with a 30-day out clause
Grow and Convert	“Starting at $10,000 per month.”	B2B SaaS GEO	Not stated
Omniscient Digital	“Full-service engagements start at $120,000/year.”	B2B search and AI search	Not stated
Ignite Visibility	“between $2,000 and $50,000+ per month”	Ecommerce marketing overall	Not stated

Source: each agency’s own pricing or service page, checked Nov 5, 2026. Links are in the Sources list.

Key Findings

• Three of the seven monthly floors in the chart sit between $3,000 and $3,500.
• The B2B specialists in the table (Grow and Convert, Omniscient Digital, Obility) start higher than generalist WebFX.
• Siege Media and Ignite Visibility price content or ecommerce marketing as a whole, so their figures are not GEO-only.

GEO Pricing by Engagement Model

Agencies sell GEO in five main ways, and each model changes what you pay for. The model also sets how long you are committed.

Model	Published example	What the fee covers
Monthly retainer	WebFX from $3,000 per month; RevenueZen $3,000 to $15,000 per month	Ongoing content, technical fixes, mentions, and reporting
Term-discounted retainer	Breaking B2B Gold at $8,500 (6 months) or $8,000 (12 months)	The same scope at a lower monthly rate for a longer term
Project or audit	WebFX guide: $5,000 to $50,000 per project	A fixed deliverable such as an AI visibility audit
Hourly consulting	WebFX guide: $50 to $300 per hour	Advice for a team that does the work itself
Software only	Peec AI from $95 per month; Ahrefs Brand Radar AI from $199 per month	Prompt tracking, with your team making the changes

Source: agency and tool pricing pages, and the WebFX 2026 GEO cost guide.

Key Findings

• A retainer pays for work. A tool subscription pays for measurement only.
• Breaking B2B’s 12-month term saves $6,000 a year on each tier it lists.

What AI Visibility Tracking Tools Cost

Most retainers include tracking, but in-house teams buy it on its own. Tool prices scale with how many prompts you track and on how many AI platforms.

Tool	Plan	Price	Included
Peec AI	Starter	$95/mo	50 prompts, 3 models, daily tracking
Peec AI	Pro	$245/mo	150 prompts, 3 models, daily tracking
Peec AI	Advanced	$495/mo	350 prompts
Ahrefs	Brand Radar AI	From $199/mo	Brand research across Ahrefs’ prompt database plus your own prompts
Ahrefs	Custom prompt packages	$50, $100, or $250/mo	2,500, 7,000, or 25,000 checks per month
WebFX	Prompt tracking add-on	$75/mo per 50 prompts	Added to a WebFX GEO plan, up to 500 prompts

Source: Peec AI pricing, Ahrefs pricing, and the WebFX GEO service page, checked Nov 5, 2026.

Key Findings

• Dividing each plan’s price by its prompts, Peec AI’s per-prompt cost falls from $1.90 at Starter to about $1.41 at Advanced.
• Ahrefs bills custom prompts per check, so tracking one prompt on two platforms uses two checks.

What Drives the Cost of a GEO Agency

The rate cards above show where the money goes. Five inputs show up again and again in them.

Content volume

RevenueZen‘s $3,000 Foundation tier includes 4 articles a month, and its $15,000 tier has no hard cap at 12 to 15+ articles.

Off-site mentions and links

Breaking B2B‘s Light plan includes 2 link credits a month, and its Gold tier includes 6 link or brand mention placements.

Prompts and platforms tracked

WebFX charges $75 per month for each extra 50 prompts. Ahrefs counts one prompt on one platform in one location as one check.

Contract length

At Breaking B2B, a 12-month term lowers the monthly rate by $500. WebFX asks for 6 months, and Siege Media asks for 12 months with a 30-day out clause.

Business complexity

Ignite Visibility says its ecommerce fees scale with “your goals, catalog size, and the channels involved.” A brand with thousands of products or several markets needs more pages covered.

GEO Agency Cost vs. an In-House Team

Running GEO in-house means paying for staff and tools yourself. The median annual wage for US marketing managers was $166,790 in May 2025, according to the Bureau of Labor Statistics.

That wage excludes benefits and software. A single Peec AI Advanced plan adds $5,940 a year at $495 per month.

Option	Published cost	Source
Marketing manager salary (median)	$166,790 per year	BLS, May 2025
Marketing manager, information industry (median)	$192,600 per year	BLS, May 2025
Grow and Convert B2B SaaS GEO	From $10,000 per month	Grow and Convert
Omniscient Digital full service	From $120,000 per year	Omniscient Digital
WebFX GEO	From $3,000 per month	WebFX

Key Findings

• One median marketing manager salary costs more than a year of Omniscient Digital‘s entry engagement.
• An in-house hire still needs content and tracking budgets on top of salary.

Questions to Ask Before You Sign a GEO Contract

Two quotes with the same monthly fee can buy very different work. These questions make proposals comparable.

1. How many new or rewritten pages do we get each month?
2. How many off-site mentions, placements, or link credits are included?
3. Which AI platforms do you track, and how many prompts?
4. What is the minimum term, and is there an out clause?
5. Who owns the content and the tracking data if we leave?
6. What does the report show each month, and how do you tie AI visibility to leads?

Be wary of any agency that guarantees AI placements. WebFX‘s own service page says that “no generative engine optimization agency, contractor, or freelancer can guarantee placements.”

Frequently Asked Questions

How much does GEO cost per month?

Among established agencies that publish a price, managed programs start between $3,000 and $10,000 per month. WebFX‘s 2026 guide puts the full market range at $1,500 to $50,000+ per month depending on company size and scope.

How much does a GEO agency cost compared with SEO?

Many agencies sell the two together. WebFX‘s SEO pricing guide puts SEO retainers at $1,000 to $5,000 per month on average, and its GEO services start at $3,000 per month.

Is there a cheaper way to start with GEO?

Some teams start with a tracking tool and do the work in-house. Peec AI starts at $95 per month for 50 prompts, and Ahrefs sells custom prompt packages from $50 per month.

Do GEO agencies require long contracts?

Many do. WebFX asks for a 6-month commitment, Breaking B2B offers 6- or 12-month terms, and Siege Media uses 12-month contracts with a 30-day out clause.

Why do so few GEO agencies publish prices?

Scope varies with content volume and the number of markets covered, so many agencies price each engagement by proposal. The Digital Elevator‘s pricing guide says “even well-established agencies rarely publish a rate card.”

Talk to Genevate

If you are budgeting for GEO next year, write down how many pages and tracked prompts you need, then ask each agency to price that same scope. You can then compare quotes line by line.

Genevate is a New York City GEO agency that works with mid-market and enterprise brands. You can review our generative engine optimization services, compare agencies in our list of the best-value GEO companies in the US, or contact Genevate to scope a program.

Sources

1. WebFX, AI search optimization services page
2. WebFX, “How Much Does Generative Engine Optimization Cost in 2026?” (Oct 15, 2026)
3. WebFX, SEO pricing guide
4. SEO.com (WebFX), AEO services
5. RevenueZen, pricing
6. Breaking B2B, pricing
7. The Digital Elevator, AEO and GEO pricing guide
8. Obility, pricing
9. Siege Media, content marketing services
10. Grow and Convert, B2B SaaS GEO agency
11. Omniscient Digital, answer engine optimization
12. Ignite Visibility, ecommerce marketing
13. Peec AI, pricing
14. Ahrefs, pricing
15. U.S. Bureau of Labor Statistics, Occupational Outlook Handbook, Advertising, Promotions, and Marketing Managers

Our analysts reviewed 20 public sources released between July 2025 and October 2026, including company disclosures from OpenAI, Google, and Microsoft, surveys from Pew Research Center, and traffic data from Comscore, Similarweb, StatCounter, and Ahrefs. We set out to measure how many people use AI search in 2026, which assistants they use, and how often an AI answer now sits on top of a Google result.

Marketing teams spent most of 2025 asking whether AI search was big enough to plan for. By late 2026, the usage numbers have settled that question, and the harder question is what those users do after they read an AI answer. That second question is the reason generative engine optimization (GEO) exists as a discipline.

In the following sections, we compare user counts by platform, adoption by age, how often AI answers appear in search, market share between assistants, and what happens to clicks and website traffic.

How Many People Use AI Search in 2026?

No single census counts AI search users, so the answer has to combine surveys with company disclosures. Surveys show what share of people use AI tools, while company figures show how large each product has grown.

• US surveys: 49% of US adults use AI chatbots, and 42% use them to search for information, according to Pew’s survey of 5,119 adults in February 2026. About a quarter (24%) use a chatbot daily.
• AI summaries in search: 60% of US adults say they read AI search engine summaries, more than the share who use a chatbot at all.
• Worldwide: Microsoft measured generative AI use at 18.8% of the world’s working-age population in June 2026, with the UAE at 73.3% and Singapore at 64.3%.
• Product scale: ChatGPT has more than 1.2 billion weekly users, while Google reports more than 2.5 billion monthly users of AI Overviews.

These figures do not add up to a total. OpenAI counts weekly users and Google counts monthly users, and nearly every AI Overviews user is simply a Google Search user who was shown one.

AI Search Users by Platform

Each company reports its audience in its own unit, so side-by-side comparisons need care. The table lists the latest figure each company has published, with the metric type in its own column.

Latest published user counts for AI assistants and AI search features

Product	Latest figure	Metric	Date	Source
ChatGPT	More than 1.2 billion	Weekly users	Oct 7, 2026	OpenAI
Google AI Overviews	More than 2.5 billion	Monthly active users	May 19, 2026	Google I/O 2026
Google AI Mode	More than 1 billion	Monthly active users	Jul 22, 2026	Google Q2 2026 remarks
Gemini app	More than 1 billion	Monthly users	Aug 11, 2026	Google
Microsoft 365 Copilot	More than 30 million	Paid seats	Jul 29, 2026	Microsoft FY26 Q4
GitHub Copilot	50 million	Users	Jul 29, 2026	Microsoft FY26 Q4
ChatGPT (business)	More than 9 million	Paying business users	Feb 27, 2026	OpenAI

Source: company disclosures from OpenAI, Google, and Microsoft, 2026.

ChatGPT’s count moved fast this year. OpenAI reported more than 900 million weekly active users on February 27, 2026, and more than 1.2 billion by October 7. Google said AI Mode queries were more than doubling every quarter since launch as of May 2026.

Anthropic’s Claude, Perplexity, and Meta AI do not appear in the table because we found no current user count from those companies themselves. Third-party data still tracks them: venture firm a16z reported that Claude has now passed DeepSeek and Perplexity in traffic, and Pew found that 6% of US adults use Claude and 14% use Meta AI.

Key Findings

• Google runs three AI surfaces that each report at least 1 billion monthly users, and two of them, AI Overviews and AI Mode, sit inside Google Search itself.
• Paid use is still small. Just 4.5% of US consumers had a paid personal subscription to ChatGPT, Gemini, or Claude in August 2026, up from 2.1% a year earlier, according to a16z.

Who Uses AI Chatbots, by Age

Age is the clearest dividing line in AI adoption, and it shapes which buyers a brand reaches through AI answers. Pew asked the same national panel about chatbot use, daily use, searching, and AI summaries, so the four measures below can be read side by side.

US adults using AI chatbots and AI search summaries, by age group

Age group	Use AI chatbots	Use daily	Use chatbots to search for information	Read AI search summaries
18 to 29	66%	31%	54%	72%
30 to 49	61%	34%	53%	68%
50 to 64	42%	19%	37%	59%
65 and older	23%	7%	20%	38%
All US adults	49%	24%	42%	60%

Source: Pew Research Center, How opinions and use of AI differ by age and Americans and AI 2026, Jun 17, 2026. Survey of 5,119 US adults, Feb 17 to 23, 2026.

Key Findings

• Adults aged 30 to 49 are the heaviest daily users at 34%, slightly ahead of adults under 30 at 31%. Daily use drops to 19% for adults aged 50 to 64.
• Every age group reads AI search summaries at a higher rate than it uses chatbots. Among adults 65 and older, 38% read AI summaries and 23% use chatbots.

How Often Google and Bing Show an AI Answer

Published rates for AI Overviews range from about 14% to over 40% of searches, and the gap is mostly about which searches each study counted. A panel of everyday US desktop searches gives a different answer than a set of trending-topic queries.

Share of searches that show an AI-generated summary, by study

Study	Search engine and sample	Period	Share with an AI answer
Similarweb	US Google searches	2026	43%+
Comscore	US Google desktop searches, panel	June 2026	39.4% (25.8% in July 2025)
Comscore	US Bing searches with Copilot Search	June 2026	17.3% (11.8% in July 2025)
Pew Research Center	Google searches by 900 US adults	March 2025	18%
Washington University in St. Louis	55,393 trending-topic Google searches	Mar 13 to Apr 21, 2026	13.7%

Source: Similarweb, Jul 29, 2026; Comscore, Sep 22, 2026; Pew Research Center, Jul 22, 2025; Washington University in St. Louis, Sep 29, 2026.

Question phrasing explains a lot of the spread. The WashU researchers found that nearly 65% of question-form queries produced an overview, compared with 9.5% of other searches. Pew found the same pattern in 2025, when 53% of searches with 10 words or more returned an AI summary, against 8% of one- or two-word searches.

Key Findings

• Comscore’s panel shows Google’s AI Overview rate climbing from 25.8% to 39.4% in eleven months. Long, question-style searches are the most likely to get an AI answer first.
• Accuracy is uneven. WashU found that about 11% of verifiable claims in AI Overviews lacked support from, or contradicted, the sources cited, and an average overview cited about eight sources.

ChatGPT, Gemini, and Claude Market Share

Market share is where AI statistics disagree the most, because each measurement firm counts something different. The table puts four current measures side by side.

ChatGPT, Gemini, and Claude share under four different measures

Measure	ChatGPT	Gemini	Claude	Source
Share of tracked chatbot traffic, worldwide, Sep 2026	80.48%	10.91%	0.95%	StatCounter
Share of tracked chatbot traffic, US, Sep 2026	75.17%	13.26%	1.81%	StatCounter US
Share of generative AI website visits, May 2026	About 53%	Around 27 to 28%	Close to 9%	Similarweb
Share of US AI prompt volume, Jun 2026	50%	30%	11%	Comscore

Source: StatCounter worldwide, Sep 2026; Similarweb, Jul 29, 2026; Comscore, Sep 22, 2026.

StatCounter measures visits from its network of tracked websites, Similarweb measures web visits to the AI sites themselves, and Comscore measures prompts in a consumer panel. On every measure ChatGPT leads, but its lead shrank in 2026: Comscore found its share of prompt volume fell from 70% to 50% between January and June.

Key Findings

• Gemini is the main gainer. Similarweb puts it at roughly 27 to 28% of generative AI web visits, up from under 9% a year earlier.
• ChatGPT handled half of US AI prompts in Comscore’s June 2026 panel, so a brand that only checks ChatGPT sees about half the picture.

How AI Answers Change Clicks and Website Traffic

Usage figures affect a marketing budget only when they change where people end up. The clearest behavioral data comes from Pew’s browsing panel and from referral benchmarks that count visits arriving from AI assistants.

What people do when an AI answer appears

Behavior	With an AI summary	Without one	Source
Clicked a regular search result	8% of visits	15% of visits	Pew, Mar 2025 browsing data
Ended the browsing session on the results page	26% of pages	16% of pages	Pew
Clicked a link inside the AI summary	1% of visits	Not applicable	Pew
Always or often click through to news sources	4% from AI	19% from search	Reuters Institute, Jun 2026 report

Source: Pew Research Center, Jul 22, 2025; Reuters Institute Digital News Report, Jun 16, 2026.

Referral data tells the same story from the website side. Across 115,254 sites, Ahrefs counted 0.43% of September 2026 traffic from AI assistants against 26.82% from search engines, with ChatGPT the largest AI source at 0.34%.

Google disputes the idea that AI features cost publishers clicks. Sundar Pichai said in July 2026 that AI Mode is driving an incremental increase in Search queries overall and that Google is sending billions of clicks to websites every week through AI features in Search.

Key Findings

• AI answers are starting to point to sources more often. Similarweb found that the share of US ChatGPT prompts showing citations rose from about 1.6% to roughly 6.8% between June 2025 and May 2026.
• Referral counts miss part of the effect, since a buyer can read an AI summary that mentions a brand and arrive later through a branded search or a direct visit.

Is AI Search Replacing Google?

Google search is still growing, and AI tools are growing faster. Both trends show up in the same quarter of company results.

• Google said Search queries reached an all-time high in the quarter before I/O 2026, and Alphabet reported 17% revenue growth in Search and Other in Q2 2026.
• Graphite, an SEO firm, estimated from Similarweb data that monthly AI sessions are 56% the size of search worldwide and 34% in the US. The same vendor study puts search-style “asking” prompts at 28% the size of search worldwide and 17% in the US, because many AI sessions are for writing or coding.
• Comscore found that 35% of US desktop users and 29% of mobile users visited an AI assistant in June 2026, up from 25% and 12% in January 2025.

AI search is adding a second place where buyers form opinions before they ever see a list of blue links. Brands now need to be accurate and visible in both places.

What These Numbers Mean for Marketing Leaders

For mid-market and enterprise brands, many buyers now meet a category inside an AI answer the brand did not write. When 60% of US adults read AI summaries and only 1% of visits include a click inside one, the summary itself can serve as the brand’s first impression.

That shifts measurement. Teams that judge AI search by referral traffic will see well under 1% of visits and conclude it is small, even though those buyers may already have read an AI description of the brand. Genevate’s GEO services track how AI assistants describe a brand and improve the sources they draw on, and our search engine optimization work covers the clicks that still come from Google.

Search is also splitting across more assistants. Gemini and Claude both gained prompt share in 2026 while ChatGPT kept adding users. So if you are setting next year’s search plan, add a monthly check of how ChatGPT, Gemini, and Google AI Overviews describe your brand next to your Google rankings report.

Frequently Asked Questions

How many people use AI search?

About half of US adults (49%) used AI chatbots in February 2026, and 60% said they read AI summaries in search results, according to Pew Research Center. Worldwide, ChatGPT reports more than 1.2 billion weekly users and Google reports more than 2.5 billion monthly users of AI Overviews.

What percentage of Google searches show an AI Overview?

Comscore measured AI Overviews on 39.4% of US Google desktop searches in June 2026, up from 25.8% in July 2025. Other studies range from 13.7% to over 43% because they measure different sets of searches, and long question-style searches trigger AI Overviews most often.

Which AI assistant is most popular?

ChatGPT leads on every major measure. StatCounter put its share of tracked chatbot traffic at 80.48% worldwide in September 2026, and Comscore put its share of US AI prompt volume at 50% in June 2026, with Gemini at 30% and Claude at 11%.

Who uses AI search the most?

Younger adults use AI chatbots most often. Pew found that 66% of US adults aged 18 to 29 and 61% of those aged 30 to 49 use chatbots, compared with 23% of adults 65 and older, while adults aged 30 to 49 have the highest daily use at 34%.

Does AI search reduce clicks to websites?

In Pew’s 2025 browsing study, people clicked a regular search result on 8% of visits when an AI summary appeared, compared with 15% without one. Ahrefs found that AI assistants sent 0.43% of website traffic in September 2026, while Google says its AI features send billions of clicks to websites every week.

Talk to Genevate

Genevate is a New York generative engine optimization agency that helps mid-market and enterprise brands understand and improve how AI assistants describe them. If you want to know where your brand stands in ChatGPT, Gemini, Google AI Overviews, and Perplexity, reach out through our contact form.

About the author

Brett Kleinberg is Founder & CEO of Genevate, a New York generative engine optimization agency. He has more than a decade of experience across PR, digital strategy and reputation management. About Genevate · LinkedIn

Sources

1. OpenAI, “GPT-6 and Intelligent UI for everyone,” Oct 7, 2026. https://openai.com/index/gpt-6-for-everyone/
2. OpenAI, “Scaling AI for everyone,” Feb 27, 2026. https://openai.com/index/scaling-ai-for-everyone/
3. Google, Sundar Pichai, I/O 2026 keynote remarks, May 19, 2026. https://blog.google/innovation-and-ai/sundar-pichai-io-2026/
4. Google, “100 things we announced at I/O 2026,” May 20, 2026. https://blog.google/innovation-and-ai/technology/ai/google-io-2026-all-our-announcements/
5. Alphabet, Q2 2026 earnings remarks by Sundar Pichai, Jul 22, 2026. https://blog.google/company-news/inside-google/message-ceo/alphabet-earnings-q2-2026/
6. Google, Gemini app passes 1 billion monthly users, Aug 11, 2026. https://blog.google/innovation-and-ai/products/gemini-app/one-billion-monthly-users/
7. Microsoft, FY26 Q4 earnings call, Jul 29, 2026. https://www.microsoft.com/en-us/investor/events/fy-2026/earnings-fy-2026-q4
8. Microsoft, “The continued state of global AI diffusion in 2026,” Sep 21, 2026. https://blogs.microsoft.com/on-the-issues/2026/09/21/the-continued-state-of-global-ai-diffusion-in-2026/
9. Pew Research Center, “Americans and AI 2026: chatbots, smart devices and views on impact,” Jun 17, 2026. https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/
10. Pew Research Center, “How opinions and use of AI differ by age,” Jun 17, 2026. https://www.pewresearch.org/internet/2026/06/17/how-opinions-and-use-of-ai-differ-by-age/
11. Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results,” Jul 22, 2025. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
12. Andreessen Horowitz (a16z), “Top 100 Gen AI Consumer Apps,” 7th edition, Oct 5, 2026. https://a16z.com/100-gen-ai-apps-7/
13. Similarweb, “AI Search Stats 2026,” Jul 29, 2026. https://aisearch.similarweb.com/blog/gen-ai-stats/
14. StatCounter Global Stats, AI chatbot market share, worldwide, September 2026. https://gs.statcounter.com/ai-chatbot-market-share
15. StatCounter Global Stats, AI chatbot market share, United States, September 2026. https://gs.statcounter.com/ai-chatbot-market-share/all/united-states-of-america
16. Comscore, Q2 2026 AI Intelligence Report press release, Sep 22, 2026. https://ir.comscore.com/news-releases/news-release-details/comscores-latest-ai-intelligence-report-shows-competition
17. Washington University in St. Louis, AI Overview citation study, Sep 29, 2026. https://source.washu.edu/2026/09/washu-research-finds-gaps-in-google-ai-overview-citations-claims/
18. Reuters Institute, Digital News Report 2026, “Emerging uses of AI chatbots for news,” Jun 16, 2026. https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026/emerging-uses-ai-chatbots-news-and-what-it-means-journalism
19. Ahrefs, AI vs search traffic benchmark, data through September 2026. https://ahrefs.com/benchmarks/ai-vs-search
20. Graphite, “AI Is Much Bigger Than You Think,” Mar 5, 2026. https://graphite.io/five-percent/research/ai-is-much-bigger-than-you-think