How often, and how favorably, a brand gets mentioned inside AI-generated answers, as opposed to how often it ranks in a traditional search results page. For agencies, it is becoming the metric clients ask for before anyone has a fully reliable way to measure it.
Share of model is a proposed successor to share of voice, adapted for a world where people increasingly get answers from ChatGPT, Gemini, Perplexity, and Google’s AI Overviews instead of scrolling a list of links. Instead of measuring how much of the media landscape a brand occupies, share of model measures how often a brand shows up when an AI model answers a relevant question, and whether it shows up as the recommended option or an afterthought.
The metric matters because AI answers only cite a handful of sources, often two to seven, per response, compared to ten blue links on a search results page. Being one of those two to seven is a much narrower target, and a much bigger prize, than ranking somewhere on page one used to be.
Share of model is where client demand is currently running ahead of measurement tooling, which is exactly the kind of gap agencies get asked to fill.
It reframes what winning looks like. AI Overviews have cut click-through rates for top-ranking content by more than half in some measurements, so a brand can lose website traffic while actually gaining ground on the metric that’s coming to matter more: getting named inside the answer itself.
It’s genuinely hard to measure consistently. AI models don’t return the same answer twice, don’t expose a ranking the way search engines do, and differ from platform to platform, so agencies reporting share of model today are mostly building their own sampling methodology rather than pulling a number from an established tool.
It rewards a different kind of content work. Improving share of model leans on being factual, well-structured, and citable, the same qualities answer engine optimization targets, rather than the keyword and backlink tactics that used to move traditional rankings.
A financial services client wants to know why a competitor keeps getting recommended when people ask ChatGPT about the client’s exact product category. The agency runs the same set of twenty representative questions across ChatGPT, Perplexity, and Google AI Overviews weekly, logs which brands get named and in what order, and reports it to the client as a share of model tracker sitting next to the usual paid and organic numbers. It’s rougher than the click-through data the client is used to, but it’s the closest thing available to a read on how the brand is doing in the conversations that now happen instead of a search.
The automations and agents module of the workshop teaches you how to build AI workflows that compress the busywork without taking the craft out of the studio.