Also written as: AI answer market share, share of AI voice
Last updated: August 18, 2026 - Reviewed by the DataEase AI editorial team
Want the full breakdown? Read the guide to AI market share ->
Share of model is a brand's slice of the answers an AI assistant gives in its category. It is rank-weighted, so a brand named first counts for more than one named tenth, and every answer's credit adds up to one.
"We were named in plenty of answers and still had a small share of model. Every time we looked closely, we were the fourth or fifth name in a list of eight. Being mentioned and being the answer turned out to be very different things."
Two properties do most of the work here, and both of them exist to stop the number from flattering you.
The first is rank weighting. Position inside an answer carries meaning. A buyer reading a recommendation treats the first name as the suggestion and the rest as hedging. Counting a first mention and a tenth mention as the same event would produce a metric that rewards being listed anywhere, which is not what anyone is buying. So the earlier you appear, the larger your slice of that answer.
The second is that the credit for a single answer always adds up to one. One answer is one unit of category attention, no matter how many brands are crammed into it. This matters because a list of ten is not ten wins. It is one answer divided ten ways, weighted toward the front of the list.
Two exclusions keep the denominator honest. Branded questions and direct questions about you are left out, because asking an assistant "what is DataEase AI" and being told about DataEase AI proves nothing about category demand. What remains are the questions a buyer asks when they do not yet know who you are.
No. The credit for a single answer always adds up to one, however many brands appear in it. A list of 10 brands shares that one answer between them, so the first name takes the largest slice and the tenth takes very little.
| View | What it answers |
|---|---|
| By brand | How your slice compares with each tracked competitor in the same category |
| By AI assistant | Whether you are strong in one assistant and absent in another, across the assistants on your plan |
| Over time | Movement across the last 6 scans, so a single noisy scan does not read as a trend |
| By question type | Whether you win discovery questions but lose comparison or recommendation questions |
The by-question-type view is usually the one that changes a roadmap. Winning discovery and losing recommendation is a different problem from the reverse, and the fix is different too.
Unclaimed answers are the share of category answers where no tracked brand is named at all. It is often the largest single slice in an early category, and DataEase AI reports it alongside the 4 share views so you can see the headroom.
Unclaimed share is the most useful number on the page for a young company, because it is the only one that tells you whether you are fighting for scraps or walking into an empty room. A category where most answers name nobody is a category where the first brand to build a real evidence trail takes a disproportionate slice. See the cold-start problem for why that room is empty in the first place.
| Concept | What it covers |
|---|---|
| Share of model | Your rank-weighted slice of category answers, relative to competitors |
| Mention rate | How often you are named at all, without regard to position or to who else was named |
| AI Visibility | The broader score covering mentions, recommendations, citations, and sentiment |
| Traditional share of voice | Your slice of media coverage or ad impressions, not of generated answers |
| Brand Presence Score | The rollup across AI assistants, citations, search, answer engines, and perception |
Answer Engine Mention Rate AI Visibility Brand Alias Cold-Start Problem Generative Engine Optimization Brand Intelligence Brand Presence Intelligence
For the full treatment with the views side by side, read the guide to AI market share. To see how your slice is affected by name resolution, read brand alias, since an unregistered alias quietly hands your credit to nobody. For the wider score that share feeds into, see the AI Visibility Score, and for question-level competitor comparison, see the Branding app.