Every answer you are absent from names somebody else. That list is the most actionable output of tracking AI visibility at all, because it tells you which pages the engine is actually reading — and most of them are not yours.
Share of voice, with its margin of error
Across 40 prompts, with 90% intervals.
Share of voice is measured from a finite number of mentions, so it carries sampling error — and at the sample sizes this category works with, that error is large. Two brands separated by a few percentage points are frequently not separated at all.
So we show the range rather than the point. When three brands are statistically tied, the page says they are tied and names them, instead of printing a confident 1–2–3 ranking that would be noise. Every other tool in this category will give you that ranking. It is not that their maths is different; it is that they are not showing you the uncertainty.
Blindspots — the prompts you are missing from
Rivals named. You are not.
A blindspot is a question where competitors appear consistently and you do not appear at all. Those are the clearest possible targets: the demand is proven, the engine is already answering it, and you are simply not in the answer.
Each one carries who was named instead, on which engines, and how consistently — so you can tell a rival who holds that ground from one who happened to appear this week. That distinction is Core vs Carousel applied to someone else's brand.
The source gap
Beyond who is named, we record which pages the engines cite. When a comparison article gets read for your category every single cycle and names four competitors and not you, the highest-leverage thing you can do is get added to that article — not publish a fifth page on your own domain.
That is frequently the single most useful row in the whole product, and it is the one least like traditional SEO: the page that decides your visibility is one you do not own and cannot edit.
How many competitors
Free tracks 3 per brand, and the paid plans more. Ones that appear repeatedly in your answers are detected automatically and offered — you do not have to guess the list up front, which is useful because the rivals AI names are often not the ones you would have written down.
Questions people ask
Why show a range instead of a number?
Because the number has a margin of error and hiding it invents precision. At the sample sizes this category works with, two brands a few points apart are frequently not separated at all — and a confident 1–2–3 ranking would have you chasing noise.
How do you decide who my competitors are?
You name the ones you care about, and brands that keep appearing in your own answers are detected and offered. Both matter: the second group is where the surprises are.
What is a blindspot exactly?
A question where competitors are named consistently and you are not named at all. Proven demand, an engine already answering it, and you absent from the answer — which makes it the clearest kind of target there is.
Can I see which pages the engines cited?
Yes, where the engine returns them. That list is often the most useful output of the whole product, because the page deciding your visibility is frequently one you do not own.