What the 2026 survey measured
The most detailed primary source we checked is a July 2026 survey by Semrush. It surveyed 643 US B2B professionals in March and April 2026. After quality checks, 622 remained, and 519 of them (83%) used AI for work. Those 519 are the base for the findings. The article gives 83% in one place and 84% in another, so the figures below use the survey's own wording.
The survey covers US buyers only. It is not a UK study.
Where buyers use AI in the purchase
Across the B2B buyer journey, buyers use AI at several stages, and the stage changes what they ask for.
Early research
Most buyers start with AI. The survey found 72% use AI during early research, such as defining a need. 66% say they regularly use AI to research products.
Shortlisting and comparison
When buyers compare, AI does more of the work. 62% use it when comparing vendors, 61% ask it to compare vendors directly, and 48% use it to narrow the shortlist. In the same survey, 92% said AI shaped their vendor shortlist.
The final decision
The survey's strongest figure is for the last step. 83% said AI influenced their final vendor decision, and 32% called it a major influence. Only 45% said AI supported the decision itself, so the influence is real but not always the deciding factor.
Buyers check what AI says
Fact-checking AI answers is common practice. Buyers do not accept an AI answer without checking it. 75% said they fully or mostly trust AI vendor recommendations. Even so, 75% still use Google or another search engine for vendor research.
After an AI mention, the next steps are consistent. 71% visit the vendor's website, 63% search Google, and 38% check a review site such as G2. Only 14% ask a colleague. The order also varies: 41% start with AI and then validate in search, while 35% start with search and then turn to AI.
Vendors can be missing from AI answers
66% said they had noticed relevant vendors missing from AI results, and 26% said this happens frequently. Buyers name what makes a vendor visible. The top factors were a close use-case match (53%) and a clear, detailed description (50%). Brand recognition scored only 7%.
Buyers also described what frustrates them: 33% said AI recommendations were too generic, and 27% said they did not reflect real pricing or contract terms.
Traffic: fewer clicks, and conflicting figures
Zero-click search, where a query ends without a visit to any site, is rising as AI summaries appear. Search referrals are falling in some categories. Ahrefs reports that AI Overviews reduce clicks to websites by 34.5%. Seer Interactive reported a 61% fall in organic click-through on informational queries when an AI Overview appeared, in its September 2025 update.
These two figures do not measure the same thing, and neither page states a sample for the click figure. Use them to show the direction. Do not use either as a planning rate.
The zero-click B2B search guide turns that traffic concern into a page and outcome measurement plan.
A UK view, with a different method
A Gilroy analysis published in October 2026 covers UK B2B marketing revenue leaders. It is not a survey. It counts 224,003 online opinions from social platforms over twelve months. It reports that ChatGPT has a major influence on the AI-search conversation for 37% of those leaders, and that 54% are not confident their brand appears in AI search queries for their buying groups.
Because the method is social-listening rather than a sample of buyers, its figures cannot be compared directly with the survey above.
What vendors should change first
Start with the facts a buyer will check, because those are what an AI tool repeats and what a buyer tests.
- Use-case clarity. State the problem you solve and who it is for. A close use-case match was the most-cited reason a vendor was noticed.
- Detailed positioning. Say what you do and what you do not do, in plain terms.
- Real pricing and terms. Publish a range and what it covers, because buyers complain when AI answers ignore pricing.
- Dated, verifiable pages. Date your claims and cite the source for each number.
For a method to test whether you appear in AI answers, see the B2B AI search visibility audit. For definitions of AEO, GEO and SEO, see AEO vs GEO vs SEO for B2B.