What Is Share of Model?
Share of model is a measurement concept that quantifies how often a brand appears in AI-generated responses across a representative set of queries. The term draws from the traditional concept of share of voice, which measures a brand's presence in paid or organic media relative to competitors. Share of model applies the same logic to the AI answer layer: across a defined query set, what percentage of responses include your brand, either as a citation or as a named mention?
Calculating share of model requires defining a query set, selecting which AI tools to audit, and then running each query and recording outcomes. If a Johannesburg logistics business queries ten AI tools with five relevant questions each (50 total responses) and its brand appears in 12 of those responses, its share of model is 24 percent for that query set and tool mix. Tracking this number periodically reveals whether content investments and SEO efforts are increasing or decreasing the brand's presence in AI-generated answers.
Share of model can be broken down in several ways. Citation share of model measures only responses where the brand's website is hyperlinked as a source, reflecting retrievable and grounded mentions. Mention share of model captures all responses where the brand name appears regardless of citation, including mentions drawn from parametric knowledge embedded during model training. Distinguishing between these two sub-metrics helps brands understand whether their presence is driven by owned content quality or by general brand recognition that predates any AI optimisation work.
As a relatively new metric, share of model has no established reporting standard. Each marketing team must decide the query set, the AI platforms to include, the frequency of auditing, and how to weight different platforms based on their user base in the South African context. Consistency in methodology matters more than the absolute number; what is useful is the trend over time and the comparison against direct competitors.
Share of Model In Practice
The scenario below is an illustrative example, not a Juicy Designs client result. The figures indicate the scale of effect that share of model tracking typically produces, so treat them as indicative rather than measured.
Picture a Cape Town-based financial services firm running a digital marketing programme that defines a query set of around 20 questions prospective clients typically ask about retirement annuities, tax-free savings accounts, and unit trusts. Running those queries monthly across Perplexity, ChatGPT, and Google AI Overviews would give a consistent picture of how often the firm is cited or mentioned in AI responses, and how that compares to two or three direct competitors.
If the firm's share of model on citation-based queries is low but its mention share is higher, the likely explanation is that the brand is known but its web content is not being retrieved by AI tools. That would point to specific content and technical improvements: improving page structure, adding schema markup, publishing more authoritative long-form guides, and acquiring links from sources that AI tools treat as reliable. Each of those actions mirrors established SEO best practice, reinforcing that share of model grows through the same content investment that builds organic search visibility. Tracking share of model alongside traditional search rankings allows South African marketing teams to allocate budget across both channels with greater precision.
What share of model measures
Share of model is an emerging metric in the age of AI search that measures how often a brand appears in the responses of AI models, relative to competitors, for the relevant questions in its market. As people increasingly ask AI assistants and answer engines rather than only searching, being named, recommended or cited by these models becomes a form of visibility, and share of model attempts to quantify a brand's presence in AI-generated answers against its competitors, much as share of voice measures advertising presence or share of conversation measures presence in wider discussion. It is assessed by testing a set of relevant questions across AI models and recording how often each brand is mentioned or cited, then comparing brands' shares. The concept reflects the shift in discovery towards AI answers: a brand frequently surfaced by AI when people ask about its category holds a strong share of model, while one rarely mentioned is largely invisible in this growing channel, regardless of its traditional search rankings.
Why share of model matters
Share of model matters because AI-mediated discovery is a growing way people research, compare and choose, so a brand's presence in AI answers increasingly affects whether it is found and considered. As with share of voice, a strong share of model is a leading indicator of visibility and, by extension, of the awareness and consideration that later drive demand: a brand consistently named and recommended by AI models when people ask about its category is positioned to win business from that channel, while one absent from AI answers cedes that ground to competitors who appear. Measuring share of model, testing relevant questions across AI systems and tracking how a brand's presence compares with rivals over time, gives an early read on AI visibility and whether it is improving, complementing traditional metrics. Improving it rests on the same foundations as broader AI visibility: being an indexed, trustworthy, well-represented entity with clear, quotable content and consistent presence across the web that AI models draw on. As a nascent metric, share of model is still being defined and measured imperfectly, but it captures something real and increasingly important, how much of the AI-generated conversation about your market features your brand, which is why it is entering the vocabulary of AI-era marketing measurement.
FAQ
How do I calculate share of model for my brand?
Define a representative set of queries relevant to your industry, then query a selection of AI tools and record how often your brand is mentioned or cited in responses. Divide your brand's mention count by the total responses reviewed to get a percentage. Repeat periodically to track changes over time.
What is a good share of model benchmark for a South African business?
No universal benchmark exists yet, as share of model is still a relatively new metric. The most useful reference is your own trend over time and your share relative to direct competitors across the same query set. A rising share of model alongside stable or growing organic traffic suggests your AI visibility strategy is working.
How do you calculate share of model?
By testing a set of questions relevant to your market across AI models and recording how often each brand, yours and competitors', is mentioned or cited in the answers, then expressing your brand's share of those mentions as a percentage of the total. It is a comparative measure of AI-answer presence, tracked over time and across models.
How is share of model different from share of voice?
Share of voice traditionally measures advertising or media presence relative to competitors. Share of model measures presence in AI models' answers, how often a brand is named or cited by AI, relative to competitors. Both gauge comparative visibility, but one concerns advertising presence and the other presence in AI-generated responses.