B2B Buyers Are Using LLMs. Your Marketing Assumes They’re Not.

B2B decision-making has quietly shifted. Buyers are using large language models to research vendors, compare options, and build internal business cases before they ever contact a sales team. If your brand is not present in the outputs those models generate, you are not being considered, and you will not know it is happening.

This is not a future problem. It is a current one, and most B2B marketing teams are still optimising for a buying process that looks nothing like what their prospects are actually doing.

The Buying Process Has Changed Without a Press Release

I spent years running agency new business. We tracked every touchpoint we could: which campaigns drove inbound, which events produced qualified conversations, which pieces of content moved prospects along. We were reasonably good at it. But there was always a gap between when someone first heard of us and when they first raised their hand, and we could never fully see inside it.

That gap has widened. And it now contains an LLM query.

A procurement lead at a mid-market company is building a shortlist for a new marketing technology platform. They open ChatGPT or Perplexity and ask something like “what are the best B2B marketing automation platforms for a company with a 50-person sales team.” They get a synthesised answer. They share it with two colleagues. That list shapes the RFP. If your brand is not on it, you are not in the room.

This is not hypothetical. It is the logical extension of how B2B buyers have always behaved: they do private research before they make contact. LLMs have just made that private research faster, more structured, and more authoritative-feeling. A well-constructed AI response looks like a recommendation from a knowledgeable colleague. That is a significant psychological weight to carry in a buying process.

Why the B2B Buying Committee Makes This More Consequential

B2B is not a single buyer. It is a committee. That has always been true, but it makes the LLM dynamic particularly interesting. When one person in a buying group runs an AI query and shares the output, that output does not just influence them. It anchors the conversation for everyone who sees it.

Anchoring bias is well-documented in decision-making research. The first structured list of options a group encounters tends to frame the rest of the discussion. If an LLM produces a shortlist of four vendors and your brand is not among them, the committee’s default assumption is that you are either less capable or less established. Overcoming that assumption requires effort from your sales team that would not have been necessary if you had appeared in the output to begin with.

There is also a timing problem. In B2B, by the time a prospect makes contact, they have often already formed a strong preference. The LLM query may have happened weeks earlier, at the very start of the consideration phase. That is the moment your brand needed to be present. Everything after is harder.

If you want to understand the broader AI marketing landscape and how these shifts connect to other areas of your strategy, the AI Marketing hub at The Marketing Juice covers the full picture, from content and tools to commercial impact.

How LLMs Decide What to Surface About Your Brand

This is where marketers need to think carefully rather than reaching for a tactical checklist. LLMs do not index your website the way a search engine does. They are trained on large bodies of text, and they generate responses based on patterns in that training data, combined with retrieval mechanisms that pull in current web content for models with browsing capability.

What that means practically is that your brand’s presence in LLM outputs is a function of how consistently and credibly you appear across the broader information ecosystem. Third-party coverage matters. Industry publications that mention you, analyst reports that include you, comparison sites that list you, forum discussions where your brand comes up in a positive context: all of this feeds into the picture an LLM constructs of who you are and whether you are worth recommending.

Your own content matters too, but not in the way most SEO thinking would suggest. It is less about keyword density and more about whether your content is genuinely informative, consistently structured, and clearly authoritative on the problems your buyers are trying to solve. LLMs are better than search engines at recognising thin content dressed up as expertise. If your website is full of vague thought leadership that does not actually answer specific questions, it will not serve you well in this environment.

The SEMrush team has done useful work on tracking how brands appear in LLM prompts, which is worth reading if you want to understand the mechanics of measurement in this space. Ahrefs has also published material on improving LLM visibility that covers the structural side of how models evaluate content authority.

What B2B Marketers Are Getting Wrong About This

The most common mistake I see is treating LLM visibility as an SEO problem with a new name. Teams are talking about “GEO” (generative engine optimisation) as if it is a technical discipline that lives in the same box as meta tags and structured data. Some of it does. But the majority of the work is not technical. It is about credibility, coverage, and content quality at a level that most B2B marketing programmes have not historically prioritised.

When I was growing an agency from 20 to over 100 people, one of the hardest things to sell internally was the value of editorial quality in B2B content. There was always pressure to produce more: more blog posts, more white papers, more social content. Volume was the metric because volume was easy to count. The problem is that a large body of mediocre content does not build the kind of authority that makes a brand recommendable. It just creates noise.

LLMs are, in a sense, forcing a reckoning with that approach. If your content does not actually help a buyer understand something they needed to understand, it will not surface when they ask an AI to help them make a decision. The models are reasonably good at identifying useful information, and they are not impressed by volume.

The second mistake is ignoring third-party signals entirely. Some B2B marketing teams still treat earned media and analyst relations as nice-to-haves rather than core pipeline activities. In the LLM era, that calculus needs to change. A mention in a credible industry publication, a positive entry on a comparison platform, a quote in an analyst report: these are not just brand activities. They are inputs that shape what AI models say about you when a buyer asks.

The Measurement Problem and How to Start Solving It

One of the things I learned from years of managing large ad budgets across multiple channels is that measurement gaps do not mean you stop trying to understand what is working. They mean you build honest approximations and iterate. Perfect attribution has never existed in marketing. The answer has always been to triangulate from multiple signals and make commercially sensible decisions.

LLM visibility is measurable, even if imperfectly. You can run structured queries across major models, ChatGPT, Perplexity, Claude, Gemini, and track whether your brand appears, in what context, and with what sentiment. You can do this manually or with tools that are now emerging specifically for this purpose. SEMrush has published guidance on LLM monitoring tools that covers the current landscape of options.

The discipline is straightforward: define the queries your buyers are likely to run at each stage of the buying process. Early-stage queries tend to be category-level (“what are the best platforms for X”). Mid-stage queries get more specific (“how does [your category] handle [specific use case]”). Late-stage queries often include competitor comparisons. Run those queries regularly. Track your presence. Note the framing.

If you are not appearing in early-stage category queries, the problem is usually one of two things: insufficient third-party coverage, or content that does not clearly establish your category positioning. If you are appearing but being described inaccurately or unfavourably, the problem is likely a combination of outdated information in the training corpus and insufficient corrective content on your own properties.

Neither problem is unsolvable. Both require sustained effort rather than a one-time fix.

Content Strategy in the LLM Era: What Actually Matters

Early in my career, I built a website from scratch because there was no budget for an agency to do it. I taught myself enough to get it done. What that experience gave me, beyond the technical skills, was a very clear sense of what a website actually needed to do: answer questions that real people had, clearly and quickly. Everything else was decoration.

That principle applies directly to content in the LLM era. The content that surfaces in AI responses tends to be content that answers specific questions completely, uses clear language, and is structured in a way that makes the answer easy to extract. It is not content that hedges, waffles, or buries the answer in three paragraphs of preamble.

For B2B specifically, this means building content around the actual questions your buyers are asking at each stage of their process. Not the questions you wish they were asking, and not the questions that make your product look best. The questions they are actually typing into AI systems at 9pm when they are trying to get their thinking straight before a meeting.

Moz has published useful analysis on how AI tools are changing content production and research workflows, which is relevant context here. The shift is not just in how buyers consume content. It is in how content teams need to think about what they produce and why.

Practically, this means prioritising depth over breadth. One genuinely comprehensive piece of content that answers a complex B2B buying question thoroughly will outperform ten shallow pieces that gesture at the same topic. It means investing in formats that LLMs can parse easily: clear headings, direct answers, structured comparisons where relevant. And it means keeping content current, because models with retrieval capabilities will surface fresher information.

Positioning and Brand Clarity as Competitive Advantage

There is a commercial argument here that goes beyond content tactics. Brands with clear, consistent positioning are more likely to be accurately represented by LLMs. Brands that have tried to be everything to everyone, or that have shifted their messaging frequently, are more likely to be described vaguely or omitted.

I have seen this pattern in agency pitches for years. The clients who were hardest to win business for were not the ones with bad products. They were the ones who could not articulate clearly what they did and who they did it for. If your own team cannot answer that question in two sentences, an LLM certainly cannot.

Sharp positioning is now a visibility mechanism, not just a brand exercise. If you are the best platform for mid-market B2B companies managing complex sales cycles, say that, consistently, across every piece of content you produce and every external context where your brand appears. The more consistently that specific claim appears across the information ecosystem, the more likely an LLM is to reproduce it accurately when a relevant query comes in.

Vague positioning gets vague AI coverage. Specific positioning gets specific AI coverage. That specificity is what gets you onto the shortlist.

What Sales Teams Need to Know

This is not purely a marketing problem. Sales teams are encountering the consequences of it every day, even if they are not framing it in these terms. When a prospect comes into a sales conversation already knowing your competitors’ positioning in detail, already having formed a view on the category, already asking specific questions that suggest they have done structured research, there is a reasonable chance an LLM was involved somewhere in that process.

Sales teams need to understand this because it changes how they should handle early-stage conversations. The prospect may have been told something about your brand by an AI that is partially inaccurate or outdated. Correcting that without making the prospect feel foolish requires a degree of awareness that most sales training does not currently cover.

It also means that sales enablement content needs to be designed with LLM research in mind. If a buyer has already used an AI to understand the category, the sales conversation needs to go deeper than the AI went. It needs to bring specificity, case studies, and commercial nuance that a general AI response cannot provide. That is where human expertise still has clear value, and it is worth designing your sales process around it.

There is more on how AI is reshaping marketing functions across the board in the AI Marketing section of The Marketing Juice, including pieces on tools, strategy, and commercial application.

A Practical Starting Point for B2B Teams

If I were setting up a programme to address this today, I would start with three things, in this order.

First, run a baseline audit. Take the 10 to 15 queries your ideal buyer is most likely to run when researching your category, and run them across the major LLM platforms. Document what comes back. Note whether you appear, how you are described, and who else is being recommended. This gives you a starting point and, if you do it quarterly, a trend line.

Second, audit your third-party presence. Where does your brand appear outside your own properties? Industry publications, analyst coverage, review platforms, forum discussions, comparison sites. If the answer is “not many places,” that is the most important thing to fix. Earned media and analyst relations are no longer optional in B2B. They are pipeline infrastructure.

Third, review your content against the questions your buyers are actually asking at each stage of the buying process. Be honest about whether your existing content answers those questions completely and clearly, or whether it is primarily designed to look impressive rather than be useful. The gap between those two things is where most B2B content programmes fall short, and it is exactly the gap that LLMs expose.

None of this requires a new technology platform or a significant budget increase. It requires clear thinking about how your buyers are actually making decisions, and the discipline to build your marketing around that reality rather than the one that was true five years ago.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions