Demand Gen Playbook

AI Chatbots Talked 57.5% of Buyers Out of a Purchase. Here Is the B2B Playbook

C By Charles Summers  ·  Sep 9, 2026, 9:08:41 AM
Chart-style illustration: AI chatbots talked 57.5% of AI users out of a purchase

AI chatbots are now vetoing purchases, not just recommending them. In a Semrush survey of 2,338 US consumers published September 7, 57.5% of AI users said a chatbot has talked them out of buying something at least once. The same study found 65% of AI users have replaced at least some product-related Google searches with chatbot conversations. If a buyer asks an AI about your product and the answer is lukewarm, you lose deals that never touch your pipeline.

The numbers come from Semrush's new consumer study, run with Exploding Topics across 2,338 US adults in July 2026 and published this week. Here is what it says, and what demand gen teams should do with it.

What did the Semrush study actually find?

The headline stat cuts both ways. 57.51% of AI users have bought a product a chatbot recommended, and 57.5% have decided against a purchase because of what a chatbot told them. Persuasion and veto run at almost exactly the same rate.

FindingShare
AI users talked out of a purchase by a chatbot57.5%
Consumers who research purchases with AI and have skipped a buy on AI advice80.93%
All consumers less likely to buy if a chatbot flags mixed or negative reviews74.15%
AI users who replaced some or all product-related Google searches with chatbots65.01%
Consumers who at least occasionally run an AI background check on a seller47.54%
AI users who discovered a brand they did not know through a chatbot59.27%

The veto effect concentrates among the buyers most worth reaching. Among consumers who name AI chatbots as one of their purchase research sources, 80.93% have decided against a purchase on AI advice, and 42.81% have done it multiple times.

Why should B2B marketers care about a consumer survey?

Because the same behavior is arriving in buying committees, and one detail in the study makes it sharper for B2B. ChatGPT only serves ads on its Free and Go plans. Plus, Pro, and business accounts stay ad-free. Your buyers, sitting on company ChatGPT seats, see only the organic answer. There is no paid escape hatch in front of them.

The buyer's first move is increasingly a background check. 47.54% of consumers at least occasionally ask an AI about a company before buying, and among people who have ever done it, nearly half did it within the past week. When a director of marketing asks a chatbot whether your product is worth it, the answer is assembled from reviews, comparison pages, forums, and documentation you may not have touched in a year.

Get the next analysis first

We break down studies like this one for demand gen teams. Grab the free resources and the next analysis lands in your inbox before it hits the blog.

Get the free demand gen resources

Should you buy chatbot ads instead?

The study argues no. 41.63% of consumers actively dislike ads in AI chatbots, and 66.05% of that group say ads make them doubt the integrity of the entire response, organic answers included. Seeing a chatbot ad made 23.12% of consumers think worse of the product, against 20.36% who thought better. That is a net negative before you spend a dollar.

Paid still has a role at the bottom of the funnel, and the ad machine is real. We covered the revenue side when ChatGPT's ad business crossed a $1 billion run rate and the mechanics in our guide to advertising in ChatGPT for B2B. The short version after this study: paid can supplement your AI visibility. It cannot substitute for the organic answer, and for B2B buyers on ad-free plans it never gets the chance.

What should you do about the AI veto?

1. Audit what AI already says about you. Ask ChatGPT, Gemini, Claude, and Perplexity the questions your buyers ask: best tool in your category for your ICP, whether your brand is worth it, your brand versus your top competitor. Run each prompt several times, note what is wrong, and note which sources the answers cite.

2. Fix the sources the answers draw from. Chatbots assemble verdicts from reviews, comparison pages, community threads, and your own documentation. Stale pricing pages, unanswered negative reviews, and thin comparison content read as evidence. The study found 74.15% of consumers are less likely to buy when a chatbot surfaces mixed reviews, so review response is now revenue work, not brand hygiene.

3. Track your share of AI recommendations like you track rankings. Dedicated trackers such as Rank Prompt monitor how often the major assistants recommend you for the prompts that matter, so you catch a slipping verdict before a quarter of lost deals does.

Want to know what the chatbots say about you?

Our SEO and AEO program audits your AI visibility, fixes the sources, and tracks your share of recommendations month over month.

See the SEO + AEO program   or talk to an expert

Frequently asked questions

What did the Semrush AI chatbot study find?

Semrush and Exploding Topics surveyed 2,338 US adults in July 2026. 57.5% of AI users have been talked out of a purchase by a chatbot, 57.51% have bought a product on a chatbot recommendation, and 65.01% have replaced at least some product-related Google searches with AI conversations.

Can AI chatbots really stop people from buying?

Yes. Among consumers who use AI as a purchase research source, 80.93% have decided against a purchase on AI advice at least once. Across all consumers, 74.15% say they would be less likely to buy if a chatbot flagged mixed or negative reviews.

Do ads in ChatGPT fix AI visibility problems?

No. 41.63% of consumers dislike chatbot ads, and 66.05% of that group say ads make them doubt the integrity of AI responses. ChatGPT also serves ads only on Free and Go plans, so many B2B buyers on paid or business accounts never see them.

How do I improve how AI chatbots describe my brand?

Audit the answers buyers get today, fix the sources those answers cite (reviews, comparison pages, documentation, pricing clarity), and track your share of AI recommendations continuously. This is the core of answer engine optimization, AEO.

C
Charles Summers

Founder of Hacking Demand. 12+ years building B2B demand generation, including 330+ B2B and B2C webinars produced.

Want us to build your demand engine?

Done-for-you webinars, AI agents, and pipeline plays, live in days.

Talk to an expertBrowse the hacks