Share of model is the AI-era version of share of voice: out of the answers a language model generates for the prompts your buyers ask, what percentage mention your brand — and how does that compare to your competitors? Where share of voice measured presence in ads or press, share of model measures presence in the answers of ChatGPT and other AI engines.
The name captures the shift. There is no page of results to hold a position on; there is a model, and it either brings you up when it answers or it doesn't. Your "share" is how much of that answer space you occupy.
How share of model is measured
Measuring it honestly takes three things:
- A fixed prompt set — the buying questions that matter for your category, run consistently over time.
- Repeated sampling — the same prompt produces different answers run to run, so a single check is noise. Mention rates only stabilize across many samples; that's why we report sample counts alongside every number.
- Mention detection across all brands — yours and every competitor the model names, so the shares have a denominator.
From there, share of model is usually reported per prompt and as a portfolio number, and it moves meaningfully on a weekly cadence, not a daily one.
Share of model vs. AI share of voice
The two terms are near-synonyms, and many teams use them interchangeably. When a distinction is drawn, AI share of voice describes your presence across AI answer surfaces generally, while "share of model" emphasizes measurement against a specific model — your share of ChatGPT's answers, say, tracked engine by engine. Either way, the discipline is the same: fixed prompts, repeated samples, honest denominators.
Why it matters
Buying research is moving into AI assistants, and the brands a model consistently names capture consideration that never shows up in your web analytics. Share of model gives that invisible funnel a number you can track, compare, and move — and a falling share is an early warning that a competitor is winning the model's confidence in your category.
Related concepts
Share of model is the metric; generative engine optimization is the practice of improving it, and query fanout explains why presence across many related sub-queries drives it. TrackGPT measures your share of ChatGPT's answers for the prompts your shoppers ask — see the AI visibility platform to get your baseline.