Definition
AI share of voice (AI SOV) is the percentage of brand mentions in AI-generated answers that belong to your brand, measured against the competitors you compare with. If AI answers to your tracked prompts mention brands in your category 200 times and 50 of those mentions are yours, your AI share of voice is 25%. It borrows the idea of share of voice from advertising and PR and applies it to the answers ChatGPT, Gemini, Perplexity, Copilot and Google's AI features give.
The formula is simple: your mentions divided by the total mentions of all compared brands, times 100. The details decide what the number means:
- Mentions are usually counted once per brand per response. In Promptwatch's definition, an answer that names your brand five times counts as one mention, so one enthusiastic answer cannot inflate the share and the result reads as 'the share of answers we appear in' relative to rivals.
- The denominator depends on who is in the comparison. Comparing against your direct competitors gives one number; comparing against every brand detected in the answers gives a lower one. Adding a competitor to the set adds their mentions to the denominator.
- Only the brand's presence counts. A passing mention at the end of an answer counts the same as the headline recommendation. Prominence belongs to other metrics.
That last point is why AI share of voice is best read next to a prominence measure. Share of voice tells you how often you appear relative to competitors; an AI visibility score tells you how prominently. High share of voice with a low visibility score means you show up often but as an afterthought. Low share of voice with a high visibility score means the answers that include you treat you well, and the problem is coverage.
It also differs from two neighbors. Mention rate divides your mentions by the number of responses, so it can rise for everyone at once; share of voice is relative, so it can fall even while your mention count grows, if a competitor grows faster. Share of model is often used for the same idea, sometimes scoped to a single model. Citation share measures sources, not brand names.
AI share of voice is most useful when segmented. By platform, because a brand can hold 40% on one engine and 5% on another and the blended figure shows neither. By topic or product line, because a company can lead in one category and be absent from an adjacent one. By prompt type, because unbranded category questions, comparisons and branded questions produce different competitive sets. And over time, annotated with content launches, competitor campaigns and platform updates.
To grow it, look at the prompts where competitors are mentioned and you are not, and read which sources those answers cite. The gap is usually content that does not answer the question, missing third-party coverage on review sites and forums the models rely on, or a crawler access problem on one platform. Our explainer on what AI share of voice is covers the measurement setup in more detail.
Examples of AI Share of Voice
- Across 150 CRM prompts in one month, AI answers mention tracked brands 900 times; a vendor with 216 of those mentions has a 24% AI share of voice, second to a rival at 31%.
- A cybersecurity firm's share of voice falls from 22% to 18% even though its mention count rose, because a competitor that launched a threat report doubled its mentions over the same weeks.
- A fintech company reports 41% share of voice for 'startup banking' prompts but 6% for 'small business accounting', and shifts its content plan toward the second category.
- A marketing team switches its comparison from five direct competitors to all detected brands and sees its share drop from 28% to 15%, because answers also mention marketplaces and review sites as brands.
