What share of voice means in AI answers
Share of voice (SOV) began as an advertising metric: your slice of a category's total ad spend or media presence. In generative engine optimization it has a sharper meaning — your share of the brand mentions that AI engines make when answering your category's questions.
If the engines you track produce 200 brand mentions across a month of answers to "best CRM for small business"-style prompts, and 46 of those mentions are your brand, your share of voice is 23%.
How it's calculated
SOV is computed over a fixed panel of tracked prompts, re-run on a schedule across engines:
share of voice = your brand's mentions ÷ all brand mentions × 100
Two design choices matter for the number to be trustworthy:
- The prompt panel must stay fixed. Changing the questions changes the denominator; trends only mean something on a stable panel.
- All brands are measured on the same answers. SOV is inherently comparative — it's computed from the same set of engine responses for you and every competitor, which is what makes it fair.
Share of voice vs mention rate
The two are easily confused:
- Mention rate asks: in what fraction of answers do I appear? It's a measure of your own presence, independent of competitors.
- Share of voice asks: of all the brand names given out, how many are mine? It's a measure of competitive position.
A brand can have a high mention rate and a shrinking share of voice — engines still name it, but they've started naming three rivals alongside. That combination is an early warning that a category's AI answers are getting more crowded.
Why SOV is the metric executives watch
Mention rates describe you; share of voice describes the race. It answers the question leadership actually asks — are we winning the category? — in one number, and it maps cleanly onto the older marketing intuition that share of voice leads share of market. When a competitor's SOV rises in AI answers, they are being introduced to buyers you never see, in conversations no analytics tool records.
Tracking the spread between your SOV and each competitor's — the competitor gap — over time turns AI visibility from anecdote into a managed KPI.
How to improve share of voice
Because SOV is zero-sum, improving it means winning mentions that currently go elsewhere:
- Find the prompts where you're absent but rivals aren't — that's the recoverable ground. A per-prompt view like Brandflare's prompt matrix makes these gaps explicit.
- Study what the engines cite for those prompts and get your brand into those sources — the comparison posts, review sites, and category roundups engines actually read.
- Publish direct, citable answers to the underlying questions on your own site (content strategy for AI search).
- Re-measure on the same panel and watch whether the gap closes.
Share of voice in AI answers is tracked automatically for every brand and competitor in Brandflare's competitor tracking.