When Buyers Ask AI Instead of Searching, What Happens to B2B Demand Generation?

By Victor Hoang · Updated Oct 11, 2026

Picture a VP of Operations who needs a new platform for her team. Five years ago she'd have typed a query into Google, opened six tabs, downloaded two PDFs and filled in a form or two. Your marketing team would have seen her. A cookie, a form fill, a campaign source.

Today she asks an AI assistant, "Which tools should a 200-person company shortlist for this, and what are the trade-offs?" She gets a structured answer in thirty seconds. She may never open a tab. Your dashboard records nothing, and yet she may still end up on a demo call, already holding an opinion about you formed by a machine you don't control.

That's the question for an executive: if the first, most influential stage of buying happens inside an answer you can't see, what is demand generation actually for?

Our view is that the job changes from earning traffic to earning trust at the source. The goal moves from getting buyers to your pages to being the thing the AI reaches for when it answers. And the way you measure it has to change too, or your team will cut the very work that's producing demand because the dashboard can't see it.

Key Facts

• Pew Research Center tracked 900 U.S. adults and found that users clicked a traditional search result in 8% of visits when a Google AI summary appeared, versus 15% when it didn't. Only 1% of visits included a click on a link inside the summary itself.
• In the same Pew data, the browsing session ended after 26% of visits with an AI summary, versus 16% without one.
• Ahrefs' analysis of 300,000 keywords found that AI Overviews reduce the organic click-through rate for position-one content by 58%, based on December 2025 data compared with December 2023. Part of that figure is modeled rather than directly observed.
• G2's March 2026 survey of 1,076 B2B decision makers found that 51% start software research with an AI chatbot more often than with Google, 69% chose a different vendor than they'd planned because of AI chatbot guidance, and 33% bought from a vendor they'd never heard of before.
• Seer Interactive's study of 3,119 informational queries across 42 organizations found brands cited in an AI Overview had an organic click-through rate of 0.70% versus 0.52% for brands that weren't cited. The authors say they cannot prove citation causes the higher rate.

What Actually Changes

Two things shrink at once, and executives tend to blame only one.

Clicks shrink. The Pew numbers are the cleanest picture of the mechanism. When an answer sits at the top of the page, fewer people leave the page, and almost nobody clicks the sources inside the answer. Ahrefs, working from a very different data set, lands on the same direction. The click you used to earn by ranking first is worth less than it was.

Visibility in the buying process doesn't shrink. It moves. This is the part that gets lost in the traffic panic. G2's survey suggests the shortlist is increasingly formed in the chatbot, not in your funnel. If a third of buyers purchased from a vendor they had never heard of before, then someone was introduced to those vendors by an AI answer. If 69% changed their plan because of chatbot guidance, the chatbot is doing the job a first sales call or an analyst briefing used to do.

So you're not looking at a smaller market. You're looking at the same demand, formed in a place your tracking can't follow.

Be careful with the evidence here. G2 is a software review marketplace, so its sample comes from people who research software, and a vendor-run survey leans toward the topic the vendor cares about. Pew measures general Google behavior, not B2B behavior specifically. None of the studies is a clean measure of what happens to your pipeline. Treat them as direction, then test it against your own data.

Why Attribution Breaks First

Most B2B attribution assumes a visible path: ad, click, visit, form, opportunity. Remove the click and the path loses its first and most influential steps.

What your reports will show:

• Direct and branded traffic rise. A buyer who learned your name from an AI answer later types it straight into the browser or searches for it. The credit goes to "direct" or "brand search," and the answer that created the intent gets nothing.
• Top-of-funnel content looks like it's failing. Informational articles lose clicks, so a naive reading says cut them. But the same article may be what the AI draws on to recommend you.
• Self-reported source gets truer than tracked source. "How did you hear about us?" answered with "ChatGPT" or "I asked an AI" is now a real data point, and it never appears in your UTM tables.

This isn't a new problem. Dark social and word of mouth always moved pipeline without leaving a trail. What's new is the volume and the speed. If your team already struggles to defend spend that isn't clickable, read our piece on attribution models both teams can trust and the practical guide to AI marketing attribution. And the argument we make in conversational ROI beyond attribution applies directly: when the influential moment can't be tagged, you need other evidence that it happened.

From Traffic Targets to Citation Targets

If clicks are a weaker signal, what replaces them? Not a single replacement metric, but a different question. Instead of "how many people did we bring to the site?", ask "when a buyer asks about our category, are we in the answer, and is what it says true?"

The Seer research hints at why this matters beyond vanity. Cited brands had higher click-through rates than uncited ones, on both organic and paid results. But read the caveat the authors themselves give: they can't establish that citation causes the lift. Brands with strong reputations may be both cited more and clicked more. That's actually the useful point for an executive. The thing you can't shortcut is being the kind of source worth citing.

Here's how the shift looks in practice.

Notice the last row. In G2's report, 45% of buyers say a review-site citation is the most confidence-inspiring signal inside an AI answer, and buyers treat consistency across chatbots as a trust signal. That tells you where the work goes: third-party presence, not just your own site.

What Becomes the Source AI Trusts

We'd be inventing a formula if we claimed to know exactly how any assistant picks its sources. The vendors don't publish it, and it changes. What you can do is invest in the things any sensible system, and any cautious human buyer, would weigh.

Say something only you can say. A generic explainer of your category is now a commodity, because the AI can write one. What it can't produce is your original data: a benchmark from your customer base, a survey of your market, a worked cost model. Publishing numbers other people have to cite makes you a primary source rather than a paraphrase. It's the same shift we describe in what marketing is for when content costs almost nothing.

Make the answer easy to lift. Lead sections with a direct answer. Use clear definitions, comparison tables and specific figures with named sources. Structure helps humans skim and helps systems extract. Our guide to AI SEO goes deeper on the mechanics.

Be present where independent opinions live. Review sites, analyst coverage, industry press, partner pages and credible communities. A claim about you that appears only on your site is weaker than the same claim appearing in three places you don't control. This is the old discipline of thought leadership, now with a machine audience in addition to a human one.

Keep your facts consistent. Pricing, positioning, integrations, who you're for. If your site, your listings and your sales deck disagree, a buyer comparing chatbot answers sees inconsistency and gets suspicious. Pick the correct version and fix every place that has the wrong one.

Watch for errors about you. G2 found 64% of buyers encounter AI inaccuracies at least a few times a month. When the AI describes your product wrong, the buyer may never tell you. Someone should ask the common category questions each month and write down what comes back.

A Decision Framework for the Budget Conversation

Your CFO will ask what to cut and what to protect. A simple sort helps.

None of this says "stop publishing." It says stop publishing things whose only reason to exist was a click that is no longer coming.

Measuring Demand You Can't See

You'll never get the clean funnel back. Replace it with a few measures that triangulate.

1. Answer share on a fixed question set. Pick 20 to 30 questions a real buyer in your category asks. Run them across the main assistants every month. Record whether you appear, what's said, and who else appears. It's manual and imperfect, and it's still more honest than guessing.
2. Self-reported attribution. Add a free-text "how did you hear about us?" field to demo requests and ask in the first sales call. Code the answers monthly.
3. Branded demand trend. Track branded search and direct visits against a baseline. A rise with flat non-brand traffic is the signature of influence happening off-site.
4. Win-rate by awareness path. Compare deals where the buyer already had you on the shortlist with deals where they didn't. G2's finding that buyers arrive with a formed opinion makes this comparison sharper than before.
5. Accuracy score. What share of AI descriptions of your product are right? Treat the number like any other quality metric with an owner.

Be honest about the limits. Answer share varies by phrasing, by day, and by assistant. Don't report it to the board as a precise figure. Report it as a trend and a set of examples. That framing also fits how we suggest measuring AI return more broadly: a small scorecard with stated uncertainty beats one number that looks precise and isn't.

What to Do Next

For a CMO, CEO or revenue leader, here's a sequence that fits in a quarter.

1. Run the audit this month. Ask the assistants the 20 questions your buyers ask. Write down who's recommended and what they say about you. This costs a few hours and changes the conversation.
2. Fix the factual errors you can control. Align pricing, positioning and integrations across your site, listings and sales material.
3. Pick one original data asset. A benchmark, a survey or a cost study. Commit the budget that would have gone to three generic posts.
4. Close the third-party gap. List the review sites, analysts and publications relevant to your category and where you're missing or thin. Assign owners.
5. Change the reporting. Add self-reported source and answer share beside pipeline sourced. Tell the board why traffic can fall while demand holds.
6. Protect the work that builds trust. Before cutting content for falling traffic, check whether it's cited, quoted or used in sales conversations.
7. Brief sales. Buyers will arrive with a chatbot-formed view. Teach reps to ask what they've read and correct it without arguing with a screen. If you're a smaller company working with a thin team, see why most SMB AI strategies are really shopping lists before you add another tool to the stack.

For the broader mechanics of generating pipeline, our demand generation reference remains the foundation. What changes is where the first impression happens, and who is making it for you.

Learn More

• Demand Generation: Strategy, Tactics, and Examples
• AI SEO Explained: How AI Changes Keyword Research, On-Page and Technical SEO
• Measuring AI ROI Beyond 'Time Saved'