Share of Voice · AI Visibility · LLM SEO

Share of Voice in AI — the metric that replaces impressions

In classic media, share of voice measured how loud your brand was on the airwaves. In the age of AI assistants, it measures something far more decisive: how often ChatGPT, Gemini and Google AI Overviews name and cite you when a customer asks their question — versus how often they recommend a competitor instead.

How we calculate it

We run a stratified library of ~120 buyer-intent prompts against every major assistant, every day. For each answer we detect who is mentioned, who is cited (linked), in what order, and with what sentiment. Your share of voice is your slice of that AI conversation — reported weekly with a 95% Wilson confidence interval, so a "move" only counts when it clears sampling noise.

Why it beats raw impressions

• An LLM answer is one answer — there is no "position 8" to console yourself with.
• Non-deterministic models mean a single query proves nothing; only aggregated share holds up.
• It compares directly: your slice vs each named competitor, on the same prompts, the same day.
• It's actionable: gaps map to specific prompts you can win with content.

Related

Share of Voice is one of four signals we track — see the full method on our LLM SEO page, or start on the AI Visibility Engine overview.