Ask the same question — "what's the best CRM for a small team?" — to ChatGPT, Gemini, and Claude. You'll get three answers with a similar tone and very different substance. One names five products. Another gives you two, wrapped in caveats. The third barely names a brand at all, talking instead about "criteria for choosing."
That difference isn't random. It comes from how each model was trained, what it retrieves in real time, and how each company tuned its assistant's behavior. And it matters a lot if you want to show up.
Why the Three Behave Differently
How many brands a model names depends on three things: what was in its training data, whether it runs a live search before answering, and how heavily it was tuned to be "cautious" versus "direct."
ChatGPT tends to be generous with proper nouns when a question asks for a recommendation — especially in its search-enabled modes, which pull recent sources and name what they find. Without search, it falls back on training memory, and the brands that surface are the ones with strong presence up to its cutoff date.
Gemini plays a home-field advantage: real-time access to Google's index. That makes it more likely to name brands with a solid, current web presence, because it literally finds them at answer time. When the topic is local or very recent, that edge shows.
Claude is usually the most reserved of the three about naming companies, particularly without search. It tends to explain criteria, offer frameworks, and only name names when the question makes clear that's what the user wants. It's not a lack of knowledge — it's a more conservative calibration about when to get specific.
"Who Cites the Most" Is the Wrong Question
It's tempting to look for a winner. But the model that names the most brands isn't necessarily where you should focus — because citing a lot also means diluting. If the AI lists eight tools, being on the list is worth less than being one of the two it mentions when it's more selective.
The useful question isn't "who cites most," it's "in what kind of question does each model cite, and how does it decide who gets in." A selective model that puts you at the top of a short answer can be worth more than a generous one that buries you mid-list.
What Changes With the Type of Question
Citation behavior varies less by brand and more by the intent of the question.
An open recommendation question ("what tool for X?") tends to generate more names across all models — it's where a brand has the best shot at appearing.
A direct comparison ("is A better than B?") usually pulls specific names, but only the ones already in the conversation or with enough standing for the model to bring them up unprompted.
A conceptual question ("how do I choose a CRM?") almost always yields few or no names — the model focuses on criteria. Here, appearing depends on your brand already being so associated with the topic that the model offers it as a natural example.
What This Means for Your Strategy
The practical takeaway isn't to pick one model and ignore the rest. It's to understand that each has a different entry point.
For Gemini, a current, structured web presence helps most — it searches live and rewards whoever is well-positioned in Google's index right now. For ChatGPT without search, what counts is having built consistent presence in the sources that made it into training. For Claude, and for any model in conceptual mode, the path is becoming such a clear reference on the topic that the AI uses you as an example even when it's being selective.
Notice that none of these doors opens with a trick. Underneath, they all ask for the same thing: authoritative content, presence in sources the AI respects, and clarity about what you solve.
Measuring All Three Is the Only Way to Know
The inconvenient detail is that your brand can have three realities at once: well-cited in one model, middling in another, invisible in the third. And there's no guessing which is which without checking.
Looking at one model gives you a biased snapshot. It's like judging your reputation by listening to a single person. Tracking how ChatGPT, Gemini, and Claude talk about your brand — across the real questions in your industry, side by side — is what turns that intuition into an actionable map.
That's exactly the view Genoma gives you: share of voice by model, so you can see where you're already well-positioned, where you're losing ground to a competitor, and which front is worth tackling first.