Tim Sanders runs innovation at G2, the review site most B2B buyers pass through before signing anything. In April 2026, his team surveyed 1,076 business buyers and found something striking: 69% ended up choosing a different software vendor than they'd planned, because an AI chatbot pointed them elsewhere.

Not a different feature. A different company.

One in three went further and bought from a vendor they had never heard of before the conversation with the AI started. Sanders described the mechanism plainly in G2's writeup: the synthesis layer is doing the shortlisting now, and the synthesis layer is someone else's software.

For a category leader, that is the uncomfortable part. The same engine that pulls an unknown challenger into a deal can drop a ten-year brand off the list without anyone deciding to exclude you. No rep loses that deal. No one logs it. It never reaches a pipeline review, because from your side, it never existed.

That is what this issue is about: not whether buyers use AI, but what happens to the vendor they had already chosen when the AI answers. The signal below is the behavior. The build is a real company that rebuilt for it. The move is what you can run before next Thursday. About twelve minutes to read, and the exercise at the end takes under an hour.

The headline number from G2's Answer Economy report is the 69%. The more useful number sits underneath it. The same study found that 51% of B2B software buyers now begin their research inside an AI chatbot more often than in a search engine — up from 29% in April 2025. In twelve months, on the same survey instrument, the starting point of the buying process moved from "sometimes AI" to "usually AI."

Two facts together change the math. If most buyers start in an AI chatbot, and most of those buyers end up choosing a different vendor than they planned, then a large share of every category's deals are being silently re-decided at the research stage — before any vendor knows a deal exists.

The mechanism is shortlist formation. Many AI engines cite only a limited set of sources per answer, far fewer than a page of search results. G2's buyers named AI chatbots as the number one source influencing which vendors make the shortlist, ahead of review sites, analysts, and vendor websites. When the list is that short and the buyer trusts it, 85% said they think more highly of a vendor an AI names in its answer — being absent from the answer is not a soft cost. It is exclusion before evaluation.

An honest note on the source. G2 disclosed that it used generative AI models to help design the survey and analyze the results. That is worth knowing. It is also a 1,076-person survey across North America, EMEA, and APAC, spanning SMB to large enterprise, and the direction aligns with what other research has shown over the past year. Treat the 69% as a strong directional signal, not a decimal-precise law.

What a decision-maker does with this is not a marketing budget question. It is a question about whether your company is present at the moment that now sets the consideration set. If your buyers are asking AI which vendor to choose, and your brand is not in the answer, your sales team is competing for deals it will never be invited to.

Webflow sells website software to a crowded market. Like every company in that market, it watched buyers move their early research into AI assistants, where a short, cited answer decides who gets considered. Vivian Hoang, Webflow's SEO Lead, runs search across 44,000 site pages and more than a thousand blog posts. Her problem was not a lack of content. It was that refreshing, it was a manual grind — and the team could see buyers were now starting in ChatGPT and arriving with higher intent. As Hoang put it, refreshing content "required hours of keyword research, competitor analysis, and CMS updates," which "prevented us from scaling effectively and responding quickly to AI search changes."

Working with the content platform AirOps, Webflow rebuilt how it refreshed material — automating the time-intensive steps, integrating directly with its CMS so updates published in minutes instead of hours, and reshaping pages to be extractable by AI engines. The reported results: a 5x increase in refresh velocity, a 40% traffic uplift within days of publication, ChatGPT-attributed signups growing from 2% to nearly 10%, and AI-sourced traffic converting at six times the rate of traditional organic search.

The number that matters to a contract signer is not the traffic. It is the conversion gap. Visitors who arrive from an AI answer have already had the comparison done for them. The model narrowed the field, and the person who clicks through is closer to a decision than someone who typed a query into a search box. That is why the same companies report AI-referred traffic converting several times higher than organic across very different categories — the AI did the qualifying work before the click.

Two caveats, stated plainly because this newsletter does not hide them. First, timing: AirOps published this case in April 2025 and last updated it in November 2025, so the result is older than the last-quarter data this issue's Signal rests on — read it as an established pattern, not breaking news. Second, sourcing: the figures are reported by AirOps, the vendor Webflow worked with, not by an independent auditor or by Webflow's own published financials. Treat them as a directional, vendor-reported result from a named, verifiable company — not as audited fact. The pattern they describe holds across the wider 2026 data: structured, fresh, original content earns AI citations, and AI-sourced traffic can convert at a premium when the answer has already qualified the visitor.

What another company can take from it has nothing to do with publishing more. It is the opposite. Webflow's gain came from making fewer pages more extractable and more current, so that when an AI engine assembled its short answer, Webflow's material was the cleanest thing to quote.

Three steps. The first is free and needs no tools.

1. Run your own buying process as your buyer would. Open ChatGPT, Claude, Perplexity, and Gemini. Ask each the way a real buyer would: "What is the best [your category] for a [your buyer's company type]?" and "[Your company] versus [your top competitor] — which should I choose?" Write down which vendors each engine names, in what order, and whether you appear at all. This is the shortlist your buyers are seeing. Most executives have never looked at it.

2. Note where you are absent or second. For every prompt where a competitor is named and you are not, or where you appear below a vendor, you outperform — capture the exact wording. That gap list is your real AI-visibility problem, in priority order, and it costs you nothing to produce.

3. Trace one gap to its cause. Pick the single most valuable prompt where you are missing. Look at what the AI cited instead. Almost always, it is a structured, recently updated, third-party-corroborated page. That tells you what the engine rewards in your category and what your equivalent page lacks.

Run the three-step check at the end of this issue before next Thursday. Open ChatGPT, Claude, Perplexity, and Gemini, ask each the question your buyer would ask, and write down who gets named. That list is the shortlist your buyers already see. If you want the deeper version, Citation Audit at revenueexperts.ai maps exactly where you're absent and why.

Elizabeta Kuzevska Co-Founder, Revenue Experts AI https://revenueexperts.ai