Share of model is your AI visibility broken down by individual AI platform, rather than combined into one figure. A brand can be strongly present in ChatGPT and almost invisible in Perplexity, and share of model captures that platform-by-platform variation. Share of voice, by contrast, is usually reported as an aggregate across platforms, so the two answer different questions: how you are doing overall, versus how you are doing on each specific model.
In short
- Share of model is the per-platform view of your visibility, while share of voice is typically the aggregate across platforms.
- Brands often perform very differently on different platforms, because each model uses different data and sources.
- The per-platform view tells you where you are strong, where you are weak, and where to focus next.
- It is a useful lens rather than a rigidly standardized metric, so confirm how any given tool defines it.
What does share of model mean?
Share of model describes how visible your brand is within a single AI platform, measured separately for each one you track, so you can compare your standing in ChatGPT against your standing in Claude, Gemini, Grok, Perplexity, Copilot, and Google's AI Overviews and AI Mode.
It is worth being straight about the terminology: share of model is a newer phrase and is not as standardized as share of voice. Different tools use it in slightly different ways, some to mean your share of mentions within a specific model, others more loosely to mean your visibility profile across models. The underlying idea is consistent and is what matters here: treat each AI platform as its own surface and measure your presence on each one, rather than assuming a single number describes them all. When a tool reports share of model, check exactly how it defines it.
How is share of model different from share of voice?
The difference is aggregation. Share of voice is your portion of all brand mentions in your category, normally rolled up across the platforms you track into one competitive figure. Share of model unrolls that, showing the same kind of information one platform at a time.
The two are complementary. Share of voice answers, in one number, how much of the category conversation you own. Share of model answers where that conversation is going your way and where it is not. A healthy aggregate share of voice can hide a serious weakness on one important platform, which only the per-platform view reveals.
Why do brands perform differently on different AI platforms?
Brands perform differently across platforms because each AI system is built differently: it learned from different data, retrieves from the live web in its own way, and trusts different sources.
A few factors drive the variation. Models are trained on different datasets and updated on different schedules, so what each one knows about your brand from memory differs. When they search the live web to ground an answer, they use different retrieval pipelines and lean on different sources, which is why one platform might surface community discussions heavily while another leans on established reference sites or news. The result is that the same brand, asking the same question, can be a top recommendation on one platform and absent on another. This is not noise to average away; it is signal about where your presence is landing and where it is not.
How do you use share of model in practice?
You use share of model to target your effort, by finding the platforms where you are weakest and understanding what those platforms reward.
Start by comparing your visibility platform by platform to find the gaps, then look at what each weak platform tends to cite for your category. If a platform you are weak on consistently draws on a type of source you are absent from, that tells you where to work, whether that is a reference site, a community, a review platform, or a particular kind of content. Prioritize the platforms your own audience actually uses, since strength on a platform your buyers do not touch is worth less than strength on the one they do. Much of the work that follows is the same content and off-site authority work covered in Answer Engine Optimization and Generative Engine Optimization, aimed at the sources a specific platform trusts. For more on how the platforms differ, see the AI platforms guide.