Why your B2B buyers are using AI before they talk to you
AI has earned trust as a research tool faster than most B2B businesses have updated their go-to-market strategy. Here's where it sits in the buying journey.
Criteo became the first ad-tech partner in OpenAI’s advertising pilot for ChatGPT Free and Go in the US earlier this year. It was a signal that AI advertising is becoming operationally real - platforms are building the infrastructure, partners are connecting, budgets are starting to move.
It also prompted a set of conversations I have been having with marketing leaders and CEOs about whether to start buying ads in AI environments, and how to think about it relative to search, social, and everything else already competing for budget.
The honest answer is: those are not quite the right questions yet. The more useful starting point is understanding where AI actually sits in your buyer’s journey - and that picture is already clearer than most of the commentary suggests.
AI recommendations and AI advertising are not the same thing
Most coverage of ChatGPT advertising conflates two things that are worth separating.
AI recommendations - the contextual, trusted responses that influence how buyers research, build shortlists, and develop purchase confidence - are already happening at meaningful scale. Buyers are using AI to research problems and evaluate options right now, in ways that shape what they think before they ever visit your website.
AI advertising is different. It is paid placements beginning to form around that behaviour. Operationally real, early in development, and not yet proven at scale as a commercial channel.
They require different strategic responses, operate on different timelines, and have different implications for where your attention should go. Treating them as one question - “should we be in AI?” - misses the more useful one: “where in the buying journey does AI already matter for us, and where is it not yet relevant?”
Where AI has already earned trust
The IAB published research this year showing that AI is now the second most influential shopping source for consumers who use it - behind search engines, ahead of retailer websites and personal recommendations.
That is a faster shift than most B2B marketing teams have accounted for.
The pattern in B2B mirrors the consumer data in the early stages. Buyers use AI to research categories, understand options, build evaluation criteria, and validate decisions before they are ready to talk to a vendor. The AI session happens before the website visit. Before the inquiry. Often before the buyer has named the problem in a way they would phrase to a sales team.
This is not a prediction about where AI is going. It is a description of what is already happening. The question is whether your business is present in those AI-mediated research moments - and if so, what it is saying.
Why the buying journey distinction matters
Consider what a typical B2B buyer journey looks like now.
A CEO or senior leader becomes aware of a problem - growth is slower than it should be, the sales cycle has extended, marketing and sales data are not telling the same story. They are not yet ready to talk to a consultant or a vendor. But they are trying to understand the landscape.
That research increasingly happens in AI. Not because AI is more accurate than other sources, but because it synthesises faster and answers follow-up questions in context. A 20-minute AI conversation at the problem-definition stage can compress what used to be an afternoon of research.
What AI says in that moment shapes the shortlist. It shapes the vocabulary the buyer uses when they do talk to vendors. It shapes whether your firm is in the frame at all.
This is the point where AI recommendations matter - and where most B2B businesses have done the least work. Not because they have ignored AI, but because they have been thinking about it as a channel (like search or social) rather than as a research environment that operates upstream of the channel.
What the conversion data tells us
Walmart ran a test earlier this year with purchases completing directly inside ChatGPT. Those transactions converted at three times lower rates than purchases that clicked through to Walmart’s own site.
That single data point is informative for B2B in several ways.
Native AI checkout is not yet a reliable commercial environment, even for low-friction consumer purchases. For considered B2B purchases - where cycles are longer, stakeholders are multiple, and trust is a significant factor - the gap is almost certainly larger.
AI earns trust in research mode. It does not yet transfer that trust effectively to commercial action in the same session.
The practical implication: the value of AI presence right now is upstream. It is influence over consideration, shortlist formation, and the framing of the problem - not direct conversion. Marketing investment that treats AI like a performance channel is likely to be disappointed. Investment that treats it as a brand and credibility environment is more likely to compound.
AI advertising: real but early
Criteo’s move into ChatGPT advertising is a real development, not a gimmick. Ad-tech infrastructure is being built. Publishers are joining. The plumbing is going in.
But there are structural reasons why B2B budgets should not move aggressively into AI advertising yet.
Anthropic has explicitly positioned Claude as an ad-free alternative to ChatGPT. That means the AI advertising environment is, at least for now, concentrated in one platform - and the audience that has opted for a different platform because they prefer not to see ads is not available.
Even within ChatGPT, not all users see ads. The Free and Go tiers see ads; paid subscribers may not. The total addressable reach for AI advertising is a subset of the AI research audience - and the most research-intensive, influential buyers tend to be the ones with paid subscriptions.
The Walmart conversion data suggests that even where ads are visible, the action they drive happens off-platform. If the conversion still requires a click-through, you are essentially paying for a different kind of awareness - not a new conversion path.
None of this means AI advertising is not worth watching. It means the timing for large budget reallocation is not now.
Where to focus instead: being present in AI research
The higher-leverage question for most B2B businesses right now is not whether to buy AI ads. It is whether AI surfaces them credibly when buyers are researching.
This is the emerging discipline of GEO - generative engine optimisation - and it is meaningfully different from traditional SEO, though not unrelated.
The core question in GEO is: when a buyer in your target market asks an AI assistant about the problem you solve, does your business appear in the answer? And if it does appear, does it appear correctly - with the right positioning, the right evidence of credibility, the right framing?
The factors that make AI likely to cite a business are becoming clearer: authoritative content that answers questions directly, structured information that is easy to parse, external references and mentions that signal credibility, and consistency between what different sources say about you. These are not entirely different from what makes content rank in search - but the emphasis shifts toward being citable over being clickable.
For most B2B businesses, the gap between their current content and what would make them consistently present in AI research sessions is the more immediate opportunity. It does not require a media budget. It requires the right content, structured the right way, saying something worth citing.
What this means for your marketing mix
The practical question most CEOs and marketing leaders need to answer is: where should AI sit in our go-to-market thinking right now?
A reasonable framework for 2026:
AI recommendations (now): Treat this as a credibility and presence question. Audit whether your business appears when buyers in your target market research the problems you solve. Invest in content that answers those questions directly and authoritatively. Think of it as the research-stage equivalent of being on the right shortlist.
AI advertising (watch, small test): If you have the capacity to run a small test in ChatGPT advertising, the learning is worth having. But do not reallocate significant budget from proven channels on the basis of current performance data. The infrastructure is early and the conversion dynamics are still resolving.
Your own site and channels (protect): The Walmart data reinforces that conversion still happens off-platform. A buyer who encounters you in an AI research session and then finds a weak website, thin content, or a slow sales response is not necessarily converted. The existing funnel still matters.
Final thought
AI has already earned trust as a research environment. That happened faster than most marketing strategies have adjusted for.
The businesses that get ahead in the next two years are not necessarily the ones that buy ads in ChatGPT first. They are the ones that become consistently present in AI-mediated research sessions - the conversations that happen before a buyer ever reaches out.
That is a content and credibility challenge, not a media buying challenge. And the window to get ahead of it, before the market catches up, is right now.
Frequently asked questions
Should B2B companies advertise on ChatGPT?
Not as a primary channel yet. The conversion data (Walmart’s 3x lower native conversion vs. click-through) suggests AI advertising currently drives awareness rather than direct action. For B2B, where cycles are longer and trust is a larger factor, the focus is better placed on being present in AI research sessions - through credible, citable content - than on early media buying.
How do buyers use AI in the B2B buying process?
Research shows AI is now the second most influential shopping source for consumers who use it, behind only search. In B2B, buyers use AI to define problems, research categories, build evaluation criteria, and validate shortlists - typically before they are ready to speak to a vendor. AI sessions increasingly happen before website visits or inquiries.
What is GEO and how is it different from SEO?
Generative engine optimisation (GEO) refers to practices that increase the likelihood of your content being cited when AI assistants answer relevant questions. It overlaps with SEO but shifts emphasis from being clickable to being citable - direct answers, structured information, external credibility signals, and authoritative sourcing tend to perform better in AI responses than keyword-dense pages optimised purely for search rankings.
How do I know if AI is influencing my buyers?
Ask five recent customers how they first heard about you and how they validated the decision. Specifically ask whether they used AI for research at any point. Most teams that have done this exercise find AI was involved earlier than they expected - and that buyers arrived at conversations with more pre-formed views than they used to.
Connected Paths works with CEOs on AI go-to-market strategy, including GEO presence and content that compounds. Start here.