~ by |
1/1/1970
There's a number that should change how every B2B marketing team thinks about their discovery strategy.
94% of B2B buyers used large language models during their buying journey in 2025. Not some buyers. Not early adopters. 94%.
They asked ChatGPT which vendors to consider. They used Perplexity to compare categories. They consulted Google's AI Overviews to understand what questions to ask before getting on a demo call. And in most of those interactions, a relatively small number of brands appeared consistently the ones with the kind of content that AI systems cite while the majority didn't appear at all.
If your brand isn't showing up in those AI-generated answers, you're losing pipeline before you can measure it. The buyer has already formed a view of the landscape, already developed preferences, already shortlisted vendors and you weren't part of that process.
This is not a distant concern. It's happening now, in every B2B category, with every buyer who is digitally fluent enough to use AI tools in their research, which, at this point, is most of them.
To understand what this means for marketing strategy, it helps to be precise about what has changed in how B2B buyers research and evaluate.
The traditional model assumed a fairly linear process. Buyers became aware of a problem, conducted research through Google and vendor websites, engaged with content, entered a sales process, and eventually made a decision. Marketing's job was to show up at each stage with the right content in the right format.
That model was always an oversimplification. But the AI-driven buying journey breaks it more fundamentally, in three specific ways.
First, the research phase now happens in a conversation rather than a search. When a buyer types "what's the best way to improve pipeline velocity for a 150-person SaaS company" into ChatGPT, they get a synthesised answer that draws on dozens of sources without visiting any of those sources directly. The buyer gets an informed view of the problem and potential approaches without a single website visit. By the time they start visiting vendor websites, they already have a framework for evaluating what they find.
Second, the evaluation phase is increasingly peer-informed and AI-assisted simultaneously. Buyers use AI to generate questions for sales conversations, to compare vendor claims against each other, and to stress-test the business case they're building internally. Sales teams who aren't aware of this are regularly walking into conversations with buyers who are better informed than they expect and sometimes better informed than the sales rep.
Third, the shortlisting process often happens before any vendor engagement. A buyer who has asked AI tools about their problem, read the synthesised recommendations, and seen certain brands cited consistently has already formed a preliminary view about which vendors are worth talking to. The first touchpoint with your sales or marketing team is no longer the beginning of the evaluation. For many buyers, it's somewhere in the middle.
The reason most B2B brands don't appear consistently in AI-generated answers is not that their content is bad. It's that their content isn't structured in a way that makes it useful to AI systems looking for something to cite.
AI systems that generate answers to research queries are looking for three things: clarity, specificity, and proof. Content that provides a direct, clear answer to a specific question. Content that supports its claims with data, examples, or named case studies. Content that is structured in a way that allows the relevant section to be extracted and referenced without the surrounding context.
Most B2B content fails on at least two of these dimensions. It's written for a general audience rather than a specific question. It makes assertions without backing them up with anything specific. And it's structured as a flowing narrative rather than as a series of clearly delineated answers.
The result is content that ranks fine in traditional search because traditional search rewards well-structured, keyword-optimised content on topics with reasonable search volume but doesn't get cited in AI answers, because AI systems are looking for something more specific and more evidenced.
This is the core of what Generative Engine Optimisation GEO is about. Not a different set of tactics layered on top of SEO. A different question driving the content strategy: what would a buyer ask an AI about their problem, and does our content answer that question clearly and specifically enough to be cited?
Based on what's working for B2B companies that are building visibility in AI-generated research responses, five content types consistently perform better than others.
The skeptical response to GEO from B2B marketing teams is understandable. It's a new channel, the attribution is difficult, and the results are hard to measure in the ways traditional marketing metrics work.
That skepticism is valid but misses the bigger point.
The question isn't whether you can directly attribute pipeline to AI citations. The question is whether the buyers you want to reach are using AI to research your category and the answer to that is clearly yes and whether being consistently visible in those research moments builds the kind of early trust and familiarity that makes every subsequent marketing and sales touchpoint more effective.
When a buyer has encountered your brand multiple times in the AI answers they were consulting before they ever visited your website, the first impression your homepage makes is fundamentally different. They're not encountering you cold. They've already seen your perspective cited. They've already formed a view that you understand the space. The sales conversation starts from a different place.
This is the same dynamic that has always made consistent thought leadership valuable in B2B — the compound effect of being the voice a buyer has been reading before they're ready to buy. AI-driven discovery is simply the new medium through which that dynamic plays out.
There is no six-month roadmap that needs to be designed before you start building AI visibility. There are three things worth doing in the next four to six weeks.
Look at your five most important pages homepage, key service pages, primary landing pages and add a direct answer block to each one. The first 100 words should answer the core question the page is supposed to address, clearly and specifically, before any supporting context. This is the single fastest way to make existing content more citable.
Identify the ten questions your ICP asks most during the buying process and build a dedicated page for each one. These pages don't need to be long. They need to be direct, specific, and structured around the question rather than around what you want to say about the topic.
Review your three best case studies and make every outcome specific. Replace "significant improvement in pipeline" with the actual number. Replace "a mid-market technology client" with a description specific enough to be meaningful industry, size, challenge, timeframe. The specificity is what turns a case study from a credential into a citable reference.
None of this requires a new tool, a new hire, or a major strategic initiative. It requires a decision to write content for the buyer doing research at 9pm before they've talked to anyone — and to give AI systems the specific, structured information they need to cite your brand when that buyer asks a question your expertise should be answering.
At VORD, we help B2B technology companies build content strategies that generate visibility where buyers are actually looking in search, in AI answers, and on the platforms where their peers are sharing what they know. If you want to understand where your brand stands in the AI discovery landscape, let's talk.
Tags: AI Buying Journey, B2B Marketing, GEO, Generative Engine Optimisation, AI Search, SaaS Marketing, Content Strategy, Demand Generation
Discover how our integrated approach can deliver predictable growth in engagement, authority, and revenue for your business.
