Brand
What your buyer asks AI changes the shortlist of brands they see
November 2026
"Who leads this category?" and "Who can help with my problem?" don't return the same list of brands. If your AI visibility report only covers the first question, you're seeing half the picture.
TL;DR
- Different questions return different brands: IBM led Cube on Category Discovery. Cube led IBM on Personas Discovery.1
- A single AI visibility score can't show you that: Look at Category Discovery and Personas Discovery side by side.
- Market to the questions, too: Say plainly who you help and when, and make the proof easy to find.
What is the impact of what your buyer asks AI?
What your buyer asks AI shapes which brands appear in the answer. The situation gap compares how often AI systems name a brand for a category with how often they name it for a specific buyer's problem. It shows how AI presence depends on the question, not buyer preference or market share.
Be careful what you ask for
We're all buyers, and more and more of us start by asking AI. Ask for the best running shoes and you'll get one answer. Ask for trail shoes for someone training on dirt roads through a Vermont winter and you'll get another.2
B2B works the same way. Someone asking who leads data platforms and someone asking how to fix three conflicting versions of revenue aren't going to get the same answer.3
One category, two very different lists
The TOM Index is StudioNorth's ranking of how brands show up in B2B technology and healthcare categories. Category Discovery asks who the players are. Personas Discovery describes a specific buyer and the problem they're trying to solve.
Take Data & Analytics, one of our technology categories. On Category Discovery, IBM appeared in 18 of 75 AI responses and Cube appeared in three. On Personas Discovery, Cube appeared in 12 of 75 and IBM appeared in four.1
Same category. Same three AI systems. On one set of questions, IBM had a 6-to-1 advantage. On the other, Cube had a 3-to-1 advantage.
I've led marketing organizations, so I know why everyone wants a single AI visibility score to put on a slide. Buyers show up with a problem to fix, systems the answer has to work with, and a boss they'll have to convince.
What we compared
In the September research behind our October TOM Index, Category Discovery and Personas Discovery each had 25 questions, run across Claude, ChatGPT, and Gemini. That's 75 responses for each set of questions in each category.1
Snowflake appeared in 32 of 75 Category Discovery responses and 43 of 75 Personas Discovery responses.1
Each brand's shift went the same way in Claude, ChatGPT, and Gemini. This doesn't show any brand gaining or losing ground over time, and it doesn't tell us who buyers would pick.1
Why the list changes
One of our Personas Discovery questions in Data & Analytics described a company with two warehouses, three BI tools, and three conflicting versions of revenue. Another asked how a retailer could agree on what sales and margin mean across store brands it had acquired. Being a well-known platform and being the right fit for a specific problem are two different things.1
A word of caution, and it applies to our data, too. The questions you ask affect who shows up. Ask more about one specific problem and the vendors who solve that problem get more chances to appear. So the next time anyone hands you an AI visibility score, ours included, ask which questions produced it.
What this looks like in healthcare
Healthcare shows the same thing. In Clinical Documentation, one of our healthcare categories, Abridge appeared in 41 Category Discovery responses and 58 Personas Discovery responses. Ambience Healthcare appeared in 21 and 41. The Personas Discovery questions asked things like whether clinicians still used a tool 12 months later and who verified a generated note. Those are the questions that separate an interesting demo from a defensible decision. These are AI responses, by the way, not findings about physician preference or clinical performance.1
What does that change in how we market?
So we've got to market to the questions, too. That doesn't replace building a strong category brand. It means knowing the real problems buyers bring to AI, saying plainly who we help and when, and making the proof easy to find. Here's where I'd start.
1. Map the buying situations that matter
What's the buyer trying to change? What has to work with what they already have? What risk would kill the project? Your sales team's objection lists and lost-deal notes are a great place to find the real wording.
2. Read the answers, not just the count
A mention might be a recommendation, a comparison, or background context. Those aren't the same thing, and a tally can't tell them apart.
3. Check the answer against your proof
If implementation experience matters to the decision, can someone find a specific account of an implementation? If interoperability matters, do you explain what connects, what doesn't, and what the customer has to do?
4. Test against the same questions
None of this proves that publishing a particular asset will increase AI visibility. So test any change against the same questions, and record the AI systems and scoring method you used. Otherwise, you're just watching a number move.
Key takeaway
What buyers ask AI, and how they ask it, shapes which brands they see. Look at Category Discovery and Personas Discovery side by side. Market to the questions your buyers are really asking.
FAQs
What's the difference between Category Discovery and Personas Discovery?
Category Discovery asks AI systems about a market in general terms, such as who the players in a category are. Personas Discovery describes a specific buyer's situation, including their environment, constraints, or risks. In the TOM Index comparison described here, those two sets of questions returned different lists of brands.
Does appearing in more AI responses mean buyers prefer that brand?
No. These counts show how often a brand appeared in generated responses to a defined set of questions. They don't measure buyer preference, adoption, or product performance, and a mention can be a recommendation, a comparison, or background context.
How does the situation gap relate to the TOM Score?
The TOM Score brings a brand's Top of Mind and Top of Model results together into one view of its category position. The situation gap isn't a TOM Index score. It's a way to look beneath a category view and ask which buying situations produce a brand's appearances.1
Will publishing new content close a situation gap?
These results don't prove that it will. Test any change against a stable set of questions, with the AI systems and scoring method recorded.
Where should a brand start?
Start with the free TOM Index at TOMIndex.com, where each category ranking shows the brand labels AI systems surfaced with their Top of Mind, Top of Model, and TOM Score. From there, map the buying situations your category ranking can't show you on its own.
Sources:
1 StudioNorth. “TOM Index, Data & Analytics and Clinical Documentation comparison of Category Discovery and Personas Discovery.” September 2026 collection, October 2026 edition. Underlying research workbooks verified by the editorial team; public index: https://www.tomindex.com/
2 Buten, John. “B2B Buyers Make Zero-Click Number One.” Forrester, January 22, 2026. https://www.forrester.com/blogs/b2b_buyers_make_zero_click_buying_number_one/
3 Anthropic. “Prompt engineering for business performance.” February 29, 2024. https://www.anthropic.com/news/prompt-engineering-for-business-performance
See where your brand stands. Then ask what it's standing on.
