Share of voice in AI search is the percentage of relevant AI answers that include a brand as a cited source or named mention. It is the single most useful headline metric for AEO because it captures position relative to competitors on the queries that matter, instead of tracking absolute citation counts in isolation.
How share of voice is calculated
The standard calculation:
- Define a query set — the prompts the brand should appear on.
- Run each query against each engine on a fixed schedule.
- For each result, mark whether the brand was cited or mentioned.
- Divide the count of brand appearances by the total responses across all queries and engines.
A brand cited in 30 of 100 monitored responses has 30% share of voice across that prompt set.
Why it beats absolute citation counts
A brand cited 50 times last month and 60 times this month has gained citations — but if the prompt set tripled in size, share of voice has dropped. Absolute counts move with prompt set size, AI traffic volume, and seasonality. Share of voice normalizes against those.
Share of voice also surfaces competitor pressure. A brand at 30% with the closest competitor at 5% is in a different position from a brand at 30% with two competitors at 28% each.
How to segment it
Track share of voice in slices, not just an overall figure:
- Per engine. Performance often differs sharply between Perplexity, Gemini, ChatGPT, and Bing Copilot.
- Per query type. Branded queries, category queries, and competitor queries each tell a different story.
- Per topic cluster. Group queries by sub-topic to see which areas are strong and which need work.
- Per content type. Citations split by which page on the site is being cited reveals which content formats are doing the work.
The headline number is the trend; the slices are the diagnostic.
What moves it
Share of voice moves when:
- New cite-worthy content is published.
- Authority signals (Wikipedia, press, schema) improve or degrade.
- Crawler access changes.
- An engine ships a model or retrieval update.
- Competitors publish or improve their cite-worthy content.
It does not move meaningfully on day-to-day swings. Treat weekly numbers as noise; treat monthly trends as signal.
Targets
Realistic targets depend on category competition:
- Niche category, few competing publishers — 40–70% share of voice is achievable for an established brand.
- Competitive category, many publishers — 10–30% is a strong position; market leaders sit around 25–40%.
- Highly competitive consumer categories — single-digit share of voice can still produce meaningful traffic.
Setting a target before measuring leads to bad targets. Run the baseline first, then set ambition relative to where competitors sit.
Implementation example
AwesomeShoes Co. wants to know whether its new footwear-fit content is actually increasing market presence in AI answers. The insights lead builds a share-of-voice model across branded, category, and competitor query sets.
Implementation discussion: the analyst tracks weekly appearance rates by engine and topic cluster, the SEO lead investigates sudden declines tied to crawl or indexing issues, and content owners prioritize clusters where share remains low despite strong demand. This gives leadership a clear signal on whether visibility gains are real and where to focus next.
