We are talking about AI models: ChatGPT, Claude, Gemini, and their derivatives. The way they describe your company, your services, and your positioning depends almost entirely on something most companies have never designed for this purpose.
According to an analysis of approximately two million LLM sessions published by Branded Agency in 2026, ChatGPT now handles 84.1% of AI discovery traffic in B2B journeys. For a growing share of buyers who use AI models to explore the market before building a shortlist, the first point of contact with your company is not your website, your sales rep, or an event. It is a response generated by a model that has processed everything you have published and decided how to summarize it.
The question is not whether this is happening. It is whether you have an answer to what the model says when it is asked about you.
How LLM models build a reputation
To understand the problem, it is worth understanding the mechanism. Language models do not "search" for information the way a search engine would: they process it, summarize it, and return it as a response. When a buyer asks ChatGPT "who are the best B2B Go-to-Market partners in Italy?", the model does not show a list of links. It builds a response based on everything it has processed on the topic.
This processing privileges some things over others.
It privileges consistency. If a company's positioning is clear and repeated consistently across multiple sources, website, articles, interviews, downloadable content, it is transmitted accurately. If it is vague, contradictory, or absent, the model interprets it by filling the gaps with inferences that do not always match how the company wants to be perceived.
It privileges concrete evidence. Data, case studies, measurable results. A positioning statement that says "we are the best" without evidence is flagged, explicitly or implicitly, as generic. A specific, verifiable positioning anchored to real contexts is treated as such.
It privileges distinction. In a market where, as documented by April Dunford in one of the most cited B2B positioning analyses, the number of vendors per category has doubled or tripled since 2020, AI models struggle to differentiate between companies that communicate similarly. If your positioning is indistinguishable from your competitors', the AI, just like a human buyer, won't know why to choose you.
What this means for corporate positioning
The first implication is that positioning is no longer just a matter of human perception. It is also a matter of machine readability: how easily a language model can understand who you are, what you do, for whom, and why you are different.
This does not change the core of good positioning, clarity, specificity, proof. However, it adds a new variable: cross-channel consistency. A company that communicates differently on its website, blog posts, sales decks, and downloadable assets presents a series of contradictory signals to the model. The result is a summary that looks like none of the original versions. And often, it is not the one the company would have chosen.
"A vague positioning did not stand up before. It stands up even less when an AI model summarizes it for a buyer who has not yet had any direct contact with your company."
The second implication concerns the type of content that produces value in the context of AI discovery. Generic content like "we are B2B GTM experts" or "we offer integrated solutions" contributes little to the model's ability to describe a company accurately. Specific content, an article that takes a stance on a precise topic, a clearly explained proprietary framework, a recognizable viewpoint on how the market works, is the type of signal a model processes and returns reliably.
The third implication is that positioning must be proof-based, not just story-based. The boundary is subtle but important. Story-based positioning says "how we were born, what we believe, what our vision is." Proof-based positioning says "who we have worked with, what problems we have solved, with what measurable results." AI models process the second much better than the first. And interestingly, human buyers do the exact same.
Practical implications for brand managers
For those responsible for corporate positioning, the starting point is not "how do we optimize for AI models." It is a more fundamental question: is our positioning clear, specific, and consistent enough to withstand any type of mediation, human or artificial?
If the answer is yes, AI discovery is an amplifier. If the answer is no, it is a revealer of weaknesses that already existed.
In practice, there are three concrete checks to make.
The first is the consistency of the message across channels. Take the website homepage, the last three blog articles, the most recent sales deck, and the founders' LinkedIn bios. Is the positioning that emerges the same? Do they have the same language, the same priorities, the same differentiators? If not, the AI model will receive mixed signals and produce a summary that does not match any of these materials.
Want to learn more?
Do you know what ChatGPT says when a B2B buyer asks about your company? We analyze your brand's positioning across leading LLMs and optimize your digital footprint to ensure you stand out in B2B AI discovery. Discover our AI services for Go-to-Market