AI search now shapes B2B buying. 94% of B2B buyers use AI while buying, AI summaries cut clicks roughly in half, and for B2B software, 79% of AI citations come from independent content and vendor comparison pages. Every number here has a named, dated source.
Key Takeaways
- AI is where B2B research starts. 94% of B2B buyers used AI during their buying process (Forrester, 2026).
- AI summaries take clicks. Searches with an AI summary led to a click 8% of the time, versus 15% without one (Pew, 2025).
- Being cited rarely sends traffic. Only 1% of searches led to a click on a link inside an AI summary (Pew, 2025).
- Independent content wins citations. It and vendor comparison pages make up 79% of AI citations for B2B software. Review sites are 7% (Growfusely, 2026).
- B2B SaaS keeps the door open. Only 21 of 442 companies block any AI crawler, and only one blocks the bots that cite it (Growfusely, 2026).
AI search: Search that returns an AI written answer instead of, or alongside, a list of links. It includes ChatGPT, Perplexity, Gemini, Google AI Overviews and Google AI Mode.
How Did I Choose These Statistics?
Most AI search statistics pages repeat the same numbers, quoted from each other, with no origin you can check. I set one rule for this page.
A statistic appears here only if it traces to a named source with a date: a research organization, a platform operator, or one of my own studies. Figures I could not trace are listed separately at the end.
Statistics marked Growfusely come from my own 2026 research. Everything else names the organization that published it.
On this page
- Which AI search numbers come from my own research?
- How many people use AI search, and how much traffic does it send?
- How much do AI Overviews reduce clicks?
- What do AI engines cite?
- How do B2B buyers use AI?
- How do AI crawlers treat your website?
- What I could not verify
Which AI Search Numbers Come From My Own Research?
These come from five studies I ran in 2026, and you will not find them anywhere else.
If you cite one thing from this page, cite one of these.
What AI cites for B2B software
1. Independent content and vendor comparison pages make up 79% of AI citations for B2B software
I analyzed 1,739 citations that ChatGPT and Perplexity produced for B2B software buying questions. Independent content and vendors’ own comparison pages accounted for 79% of them.
Source: Growfusely citation study, 2026
2. Review platforms are only 7% of AI citations for B2B software
G2, Capterra and similar platforms made up about 7% of citations. If your AI visibility plan is mostly “get more reviews,” you are working on the smallest slice.
Source: Growfusely citation study, 2026
3. Independent content alone is 61% of what AI cites for B2B software
Blogs, editorial round ups and independent comparison pages were the single largest source type.
Source: Growfusely citation study, 2026
4. Vendors’ own comparison pages show up in every category, at 12% to 24% of citations
Across 16 B2B software categories, vendor comparison and alternatives pages were cited consistently. It is one of the few AI visibility levers you fully control.
Source: Growfusely citation study, 2026
5. ChatGPT and Perplexity cite different kinds of sources
ChatGPT cited vendor pages and tech media two to three times more often than Perplexity. Visibility in one engine tells you little about the other.
Source: Growfusely citation study, 2026
6. Reddit appeared in just 11 of 1,739 B2B software citations, all from Perplexity
That is very different from studies of general queries, which often put Reddit near the top of AI citations. For B2B software buying questions, community content barely registers, and ChatGPT did not cite Reddit once in my sample.
Source: Growfusely citation study, 2026
Who blocks AI crawlers
7. Only 21 of 442 B2B SaaS companies (4.8%) block any AI crawler
The predicted wall against AI never went up in B2B SaaS. The other 421 leave every AI crawler I checked free to read their sites.
Source: Growfusely crawler study, 2026
8. The most blocked AI crawler is TikTok’s, not OpenAI’s
Bytespider, run by TikTok’s parent ByteDance, is blocked by 13 sites. OpenAI’s GPTBot is blocked by 5.
Source: Growfusely crawler study, 2026
9. Exactly one of 442 B2B SaaS companies blocks the bots that cite it
Thirteen companies block a training bot by name. Only Figma blocks the search bots, including OAI-SearchBot and PerplexityBot, that put a brand in live AI answers.
Source: Growfusely crawler study, 2026
10. 53 B2B SaaS companies explicitly invite AI answer engines in
They name OAI-SearchBot and PerplexityBot in robots.txt and allow them by name. The list includes Apollo, Chargebee, Okta, PostHog, Retool and Zapier.
Source: Growfusely crawler study, 2026
llms.txt adoption
11. 51.8% of B2B SaaS leaders publish an llms.txt file
264 of 510 reachable companies publish one, fast adoption for a file first proposed in September 2024.
Source: Growfusely llms.txt study, 2026
12. Only 10.2% publish the fuller llms-full.txt
52 of 510 companies went beyond the basic file. Most are developer facing: Atlassian, Cloudflare, Databricks, GitHub and Grafana among them.
Source: Growfusely llms.txt study, 2026
13. SEO and content tools have the lowest llms.txt adoption of any category, at 23%
Design and no code tools lead at 70%. The sample for SEO and content tools is 13 companies, so treat the gap as a signal.
Source: Growfusely llms.txt study, 2026
AI readiness in fintech and cybersecurity
14. Fintech averages 82 out of 100 on AI readiness, and cybersecurity averages 80
I scored companies on four public signals: structured data, meta tags, citation crawler access and llms.txt. Fintech is the most machine readable category I have measured.
Source: Growfusely fintech readiness study, 2026; Growfusely cybersecurity readiness study, 2026
15. Ten of 95 fintech companies and 6 of 97 cybersecurity companies scored a perfect 100
In fintech, Alloy, Brex, Column and Sift are among them. In cybersecurity, Auth0, Cloudflare, Descope, Sumo Logic, WorkOS and Zscaler.
Source: Growfusely fintech readiness study, 2026; Growfusely cybersecurity readiness study, 2026
16. Structured data is already common: 83% of fintech and 80% of cybersecurity companies use it
The signal companies skip most is llms-full.txt, published by only 12% of fintech and 10% of cybersecurity companies.
Source: Growfusely fintech readiness study, 2026; Growfusely cybersecurity readiness study, 2026
How Many People Use AI Search, and How Much Traffic Does It Send?
Hundreds of millions every week, and the numbers keep climbing.
For B2B teams, the more useful question is how fast AI is becoming a research channel for buyers, covered further down.
Usage
17. ChatGPT reached 900 million weekly active users in February 2026
That is up from 400 million a year earlier. It is the most recent figure OpenAI has confirmed; reports of 1 billion weekly users have not been officially confirmed.
Source: OpenAI, February 2026
18. The ChatGPT app passed 1 billion monthly active users in June 2026
Sensor Tower estimates put it among the fastest apps in history to reach that scale. This is an app estimate, separate from OpenAI’s weekly figure.
Source: Sensor Tower via Reuters, June 2026
19. OpenAI’s annualized revenue passed $25 billion by February 2026
That is up from about $10 billion in mid 2025. AI search is now a large commercial business, which means its behavior will keep changing quickly.
Source: OpenAI via Reuters, February 2026
20. 88% of organizations use AI in at least one business function
Yet only about 6% qualify as high performers that capture significant value from it. Adoption is nearly universal; the advantage comes from execution.
Source: McKinsey, The State of AI, 2025
Referral traffic to B2B websites
21. ChatGPT referrals to B2B websites grew 303% in a year
Monthly visits from ChatGPT to the B2B sites Demandbase measures rose from about 645,000 in June 2025 to 2.6 million in June 2026. The analysis covered more than 11 billion website visits across 1,584 companies.
The base is still small compared with total B2B web traffic, so measure it separately before drawing conclusions.
Source: Demandbase, August 2026
22. Perplexity referrals to B2B websites declined over the same year, while Gemini and Claude stayed flat
ChatGPT drove almost all of the growth in AI referrals to B2B sites. It is another reason to plan for each engine separately.
Source: Demandbase, August 2026
How Much Do AI Overviews Reduce Clicks?
Substantially. The best studies show clicks falling by roughly half when an AI summary appears. Treat any single “traffic drop” number with care, since methods vary widely.
23. Searches with an AI summary led to a click 8% of the time, versus 15% without one
Pew tracked the real browsing of 900 U.S. adults. This is the cleanest click impact figure available.
Source: Pew Research Center, July 2025
24. Only 1% of searches led to a click on a link inside the AI summary itself
Being cited in an AI answer rarely sends traffic. The value is presence and influence in the answer.
Source: Pew Research Center, July 2025
25. People ended their browsing session after 26% of searches with an AI summary, versus 16% without one
An AI summary often ends the search entirely.
Source: Pew Research Center, July 2025
26. An AI Overview now means 58% lower clickthrough for the top ranking page
Ahrefs compared 300,000 keywords using December 2025 data. The same study measured a 34.5% drop in April 2025, so the effect has grown.
Source: Ahrefs, 2026
27. Estimates of how often AI summaries appear range from 18% to about 48% of searches
Pew saw summaries on 18% of searches in real browsing data. BrightEdge reported AI Overviews on about 48% of tracked queries in February 2026. Different keyword sets produce very different answers.
Source: Pew Research Center, July 2025; BrightEdge, February 2026
28. 58.5% of U.S. Google searches end without a click to any website
Zero click search predates AI Overviews. AI summaries are accelerating a trend that featured snippets started years ago.
Source: SparkToro and Datos, 2024
What Do AI Engines Cite?
Mostly independent content, and each engine has its own preferences.
B2B software queries behave differently from general queries.
29. 88% of Google’s AI summaries cite three or more sources
Only about 1% cite a single source.
Source: Pew Research Center, July 2025
30. Wikipedia makes up about 48% of ChatGPT’s ten most cited sources
ChatGPT leans on encyclopedic sources for general queries. For B2B software questions, my own data shows independent comparison content matters more.
Source: Profound, 2025
31. Studies disagree sharply on how many AI Overview citations come from top ten results
BrightEdge put it at 17% in February 2026. Ahrefs earlier found a much higher share. Ranking well still helps, but it no longer guarantees you get cited.
Source: BrightEdge, February 2026; Ahrefs, 2025
32. Review platforms receive 8.5% of all links cited in Google AI Overviews
SE Ranking analyzed 30,000 commercial keywords and more than 211,000 cited links. Review platforms appeared in 34.5% of AI Overviews but made up a small share of total links.
That is close to the 7% I found across ChatGPT and Perplexity. Two different methods reach the same conclusion: review platforms matter, but they are a small part of what AI cites.
Source: SE Ranking, April 2026
33. Five review platforms account for 88% of review platform citations in AI Overviews
Gartner Peer Insights leads at 26%, followed by G2 at 23.1% and Capterra at 17.8%. If you invest in review profiles, focus on the platforms your category is actually cited on.
Source: SE Ranking, April 2026
How Do B2B Buyers Use AI?
As their starting point. Your buyers form opinions and shortlists inside AI tools long before they contact your sales team.
Where B2B research happens
34. 94% of B2B buyers used AI during their buying process
Forrester surveyed nearly 18,000 business buyers. Generative AI and conversational search now rank as a more meaningful information source than vendor websites and sales reps.
Source: Forrester, The State of Business Buying 2026
35. B2B buyers spend only about 17% of the buying journey with any vendor
Most of the journey is self directed, and much of that research now happens in AI tools.
Source: Gartner
36. 61% of B2B buyers prefer a buying experience without a sales rep
Buyers want to research on their own terms, for at least part of the journey.
Source: Gartner, June 2025
Who is involved in a purchase
37. A typical B2B buying decision involves 13 internal stakeholders and 9 external influencers
Each one may research separately, often inside AI tools.
Source: Forrester, The State of Business Buying 2026
38. A complex B2B purchase involves 6 to 10 decision makers, each with four or five pieces of information they gathered on their own
Every one of those people can form a view of your company from AI answers before your sales team meets them.
Source: Gartner
Trust and verification
39. 69% of B2B buyers check AI generated insights with a sales rep
Buyers use AI to start, and people to confirm.
Source: Gartner, May 2026
40. 19% of buyers who use generative AI feel less confident in a purchase because of inaccurate information
Accuracy is a real problem. When AI describes you poorly, it can cost you trust as well as visibility.
Source: Forrester, October 2025
How Do AI Crawlers Treat Your Website?
They visit constantly and send back very little traffic. Controlling them is a per bot decision.
How much AI crawlers take
41. Bots now generate 57.5% of HTML web traffic, and people 42.5%
It is the first time automated traffic has passed human traffic on Cloudflare’s network.
Source: Cloudflare, June 2026
42. Google crawls about 5 pages for every visitor it sends; AI crawlers crawl hundreds to thousands
In one July 2026 reading, OpenAI crawled about 217 pages per referral and Anthropic about 2,237. These ratios swing sharply from month to month.
Source: Cloudflare Radar, 28 day window ending July 21, 2026
How to control AI crawlers
43. OpenAI runs three separate crawlers, each with its own robots.txt rule
GPTBot collects training data, OAI-SearchBot supports citations, and ChatGPT-User fetches pages when a person asks. Blocking one does nothing to the others.
Source: OpenAI documentation
44. Google-Extended and Applebot-Extended are opt out tokens, not crawlers
Blocking them keeps content out of model training. It has no effect on Google Search, AI Overviews or Siri.
Source: Google and Apple documentation
45. Google has said it does not use llms.txt
John Mueller wrote in June 2025 that no AI system currently uses the file, and Google’s 2026 guidance lists it among tactics site owners can skip.
Source: Google, 2025 and 2026
46. robots.txt does not stop every crawler
In August 2025, Cloudflare published evidence that Perplexity used undeclared crawlers to reach blocked pages. Perplexity disputed the findings. Either way, robots.txt only works when a bot honors it.
Source: Cloudflare, August 2025
Which Popular AI Search Statistics Could I Not Verify?
These figures circulate widely, but I could not trace them to a primary, dated source. I am not saying they are wrong, only that I will not publish them as fact.
- “AI Overviews reduce organic traffic by 70% or more.” The defensible figures are Pew’s 8% versus 15% and Ahrefs’ 58% drop for the top result.
- A single “share of searches that are now AI.” Estimates range from 18% to about 48% depending on method.
- Precise ChatGPT versus Google market share percentages. Most compare different things, such as users, visits or queries.
- Reddit as the most cited source across all AI answers. Studies of general queries say so. My B2B software data found Reddit in under 1% of citations.
- User counts for Google AI Mode, Gemini, Perplexity and Copilot beyond what each company has officially confirmed.
If you have a primary source for any of these, email hello@growfusely.com and I will verify it and add it with credit.
How Should You Use These Statistics?
- Measure AI referrals properly. Check how your analytics classifies traffic from ChatGPT and Perplexity, since much of it arrives without a referrer.
- Track citations, not just clicks. With 1% of searches leading to a click inside an AI summary, presence in the answer is the real goal.
- Invest in comparison content. Your own honest comparison pages are among the few sources you fully control.
- Plan for each engine separately. ChatGPT and Perplexity cite different kinds of sources.
- Keep the citation bots reachable. Check your robots.txt against the AI Crawler Directory.
- Make your descriptions consistent. If engines describe you vaguely, you may have entity drift.
How to Cite This Page
Dholakiya, P. (2026). AI Search Statistics 2026. Growfusely. https://growfusely.com/blog/ai-search-statistics/
You are welcome to use any statistic or chart with a link back to this page. For the original Growfusely figures, please link to this page or the individual study.
Why Partner with Growfusely for AI Search Visibility?
Every number on this page leads back to the same question: when your buyers ask AI about your category, are you named?
- Citation measurement across ChatGPT and Perplexity, so you know where you appear and where you do not.
- Original research, like the five studies on this page, built for your own category.
- Comparison and alternatives content, the source type AI cites most for B2B software.
- Crawler access and entity consistency audits, so engines can read you and describe you accurately.
If you want your own numbers, schedule a call.
FAQs
How many people use ChatGPT?
OpenAI reported 900 million weekly active users in February 2026, up from 400 million a year earlier. Sensor Tower estimated the ChatGPT app passed 1 billion monthly active users in June 2026.
How much do AI Overviews reduce clicks?
Pew found searches with an AI summary led to a click 8% of the time, versus 15% without one. Ahrefs measured 58% lower clickthrough for the top ranking page when an AI Overview appears.
What do AI engines cite for B2B software?
In my analysis of 1,739 citations, independent content and vendor comparison pages made up 79%. Review platforms were about 7%.
How many B2B buyers use AI?
Forrester found 94% of B2B buyers used AI during their buying process, based on a survey of nearly 18,000 buyers.
Do B2B SaaS companies block AI crawlers?
Rarely. Only 21 of 442 B2B SaaS companies I checked block any AI crawler, and only one blocks the bots that cite it in AI answers.
Read Next
- What AI cites for B2B software. My analysis of 1,739 AI citations.
- The AI Crawler Directory. Every AI bot and what blocking it costs.
- Entity drift. Why AI engines hedge when describing your company.
Pratik Dholakiya is the Founder of Growfusely, a SaaS SEO and AI Search Visibility agency.
