What is share of voice in AI search?

Two businesses can both get mentioned by ChatGPT every week and still be in completely different positions. One might be the first name that comes up nine times out of ten; the other might scrape in as an afterthought, mentioned but rarely leading. A plain mention count won't tell you which is which. Share of voice is the number that does.

The metric

It's borrowed from advertising, where share of voice originally meant your slice of total ad spend or airtime in a category. The AI search version keeps the same idea and swaps spend for mentions. Across all the questions you track, engines mention a set of brands — yours and your competitors'. Share of voice is the percentage of those mentions that are yours. If, say, you're named in 3 out of 10 total brand mentions across your tracked prompts, your share of voice is 30%. It's a share of a fixed pool, not a raw count, which is what makes it useful: it answers "how much of the conversation is mine" rather than just "how often do I show up."

Why it beats counting mentions

A raw mention count goes up the moment you start tracking more prompts — track twice as many questions and, all else equal, you'll rack up roughly twice as many mentions, with nothing about your actual standing having changed. Reported on its own, a rising mention count can look like progress when it's really just a bigger sample size. Share of voice doesn't have that problem: it's a proportion, so it stays comparable as you add prompts, add competitors, or come back to check again next month. It's also the only version of the number that lets you compare directly against a named competitor rather than against your own past self, which is usually the comparison that actually matters to a business.

Reading it per engine

Share of voice isn't one number — it's one number per engine, and they don't move together. A business can dominate ChatGPT's answers in its category while barely appearing in Perplexity's, because the two engines search differently, trust different sources, and draw on different corners of the web when they build an answer. Look at the aggregate and you'll miss which engine is actually worth the effort, and end up spreading attention evenly across engines that don't deserve equal weight.

Moving the number

Share of voice moves when you close the specific content gaps that are costing you mentions, and when you earn citations on the sources engines already trust in your market. Both of those start with the same input: knowing where the gaps are, rather than guessing at them. Opportunity tracking surfaces the prompts where competitors are named and you aren't, which is where the number actually moves — chasing prompts you already win does little for the total.

See yours

Visibility tracking shows your share of voice per engine, tracked over time, against the competitors you choose to watch, so you can see whether the work you're doing is actually shifting the balance.