What is Model Share?
Model Share is the percentage of AI-generated answers, within a defined prompt set, in which a brand is mentioned. It is the generative-search analogue of share of voice, measured across ChatGPT, Gemini, Perplexity, Claude and Copilot. The metric exists because buyers increasingly stop at the AI answer rather than clicking through to a ranked list, so a brand can hold position one in Google and still be absent from the response a buyer actually reads. Measuring it requires running a category's real buyer prompts against each model on a schedule and recording three things: whether the brand was named, how it was described, and which sources the model cited. The third is what makes the metric actionable — it identifies the specific page, review or discussion that moved visibility, rather than merely reporting that visibility moved.
In detail
What it replaces
Share of voice, which counts impressions on channels a buyer may never reach if the AI answer resolves their question first.
How it is measured
A fixed prompt set is run against each model on a schedule. Mentions, sentiment and cited sources are recorded per run so the trend is comparable over time.
Why sources matter
Knowing a competitor was cited is diagnosis. Knowing which URL the model cited is a task.
Where it sits in MarkGrid
Model Share is the flagship module of the fifteen, and the one most teams adopt first.
Related terms
Stop Guessing. Start Executing.
Join the brands that let AI agents detect flaws and fix them, automatically. Talk to sales and we will map the platform to the gaps in your stack.
