What does competitor share of voice show you?
It shows who AI engines talk about when your buyers ask a question, and how much of that talk goes to you. Being mentioned is a start. Being mentioned first, and recommended, is what tends to lead to a visit.
Most teams come to this with a simple worry. A rival seems to be named everywhere, and they want to know if that is true and why. Share of voice turns that worry into a number you can track, per engine and per prompt.
How is AI share of voice worked out?
It is worked out from every stored answer, with each mention weighted by where it appears.
- Read every answer. Competitor intelligence goes through each stored answer from your prompt set.
- Record every brand. Every brand named is logged, whether or not you listed it as a competitor.
- Note the position. The order in which each brand first appears is recorded.
- Weight by position. Each mention gets a weight of 1 divided by its position, so first place counts most.
- Record the recommendation. Whether each brand was recommended, and the reason given, is stored too.
- Add it up. Your weighted mentions are compared with everyone else's, over time and per engine.
The weighting looks like this for the first few places.
| Position in the answer | Weight |
|---|---|
| Named first | 1 |
| Named second | 1/2 |
| Named third | 1/3 |
| Named fourth | 1/4 |
Alongside share of voice sits a 0-100 prominence score. It folds in whether you were recommended, whether your pages were cited and the sentiment of the answer.
What is share of recommendation?
Share of recommendation counts only the times an engine actually recommends a brand. It is a stricter measure than share of voice. A brand can be named in many answers and still rarely be the one the engine suggests.
The reasons are stored with each recommendation. So you can read why an engine prefers a rival, in its own words. Common reasons might be price, range, reviews or location. Those reasons are often the most useful part of the whole report.
What do you see on screen?
You see a league table of brands for your prompt set, with your brand picked out. Charts show how each brand's share moves over time. You can switch between engines to see where a rival is strong.
You can also drill into any prompt to read the answers behind the numbers. If you run the same prompts in several places, local and multi-market AI visibility splits share of voice by geography.
Why does share of voice change from week to week?
Some movement is normal, because AI answers vary from run to run. A single week's drop is rarely worth acting on by itself. A steady change over several weeks is more likely to be real.
It also helps to look at where the change happened. A shift in one engine only often points to a change in the sources that engine uses. AI citation tracking shows those sources, so you can see what a rival gained or you lost.
Why does this matter for AI search?
In a normal search result, a buyer sees a page of options. In an AI answer they often see only a handful of names, and the first one tends to get the most attention. So the gap between first and fourth matters more than it used to.
Share of voice also gives you a fair way to set goals. You can't control how often a question is asked. You can track whether your part of the answer grows. Our guide to AI share of voice covers how to set that up and read it.
Who uses it?
Marketing leads use it to report on the brand's position against rivals. Agencies use it to show clients where they stand, and it fits into client reports with the client's own logo. Firms in crowded categories use the recommendation reasons to shape their messaging.
Take a dental group with several practices. It might find that the engines name it often but recommend a rival for evening appointments. That is a clear lead. It might be worth making opening hours easier to find on the site and on the directories the engines cite, then watching the recommendation share over the next few runs.
Where it fits
Share of voice is built on AI visibility tracking. AI citation tracking shows the sources behind each rival's mentions. Local and multi-market AI visibility splits the numbers by place, and client reports turn them into a monthly PDF.