Share of Model

Share of voice measured your slice of advertising or media presence; share of model measures your slice of AI-generated answers. The shift matters because AI engines act as a recommendation layer between buyers and brands: when a model answers "best CRM for startups" with three names, those three brands captured 100% of that answer's consideration set. Share of model aggregates this across the prompts that define your category, making it the natural successor to AI share of voice metrics.

The term emerged as marketers realized model outputs, not search rankings, were becoming the contested surface. Unlike a results page with ten blue links, an AI answer typically names only a handful of brands, making the metric brutally zero-sum.

Define a prompt set representing real category demand, 20 to 50 buyer-style questions, and run it daily across the engines that matter to your audience. Share of model is the share of those answers mentioning your brand, computed overall and per provider. Refinements weight by mention position, since being the first recommendation is worth more than a trailing footnote, and by prompt importance.

Daily repeated sampling is non-negotiable. Because of answer volatility, a single run is a coin flip, not a measurement; a stable share figure only emerges from aggregating many runs over time. Comparing share across providers also reveals where the gap is concentrated.

Share of model is the headline number for AI search programs: it compresses visibility, competition and category coverage into one trendable figure that executives grasp instantly. Use the per-prompt breakdown to direct work, prompts where competitors appear and you do not are your highest-leverage content gaps. Track share after each intervention to attribute movement. Platforms like Geonimo compute this continuously via multi-model monitoring, sampling your prompt set across providers daily so share of model becomes a managed KPI rather than an occasional manual audit.