Who AI systems recommend instead of you.
AI answers have no positions 1 to 10. An AI names three providers, perhaps recommends one and links to another. If you only know your own mention rate, you cannot see who is taking that spot.
OFTEN CITED AUTHORITY
IN THE MARKET TALKED ABOUT,
NOT LINKED BARELY
PRESENT
Measured against competitors you have chosen.
It includes competitors confirmed in the company profile or added manually. The many domains from SERP runs only count once they have been confirmed.
This keeps the comparison group commercially meaningful instead of expanding it with every brand detected.
View company profile →Four recommendation states.
A stacked bar for each brand shows the distribution, so a mention is never confused with a recommendation.
The brand is the first answer to the question.
Mentioned with explicit endorsement.
Mentioned without judgement.
Mentioned with a caveat or warning.
Brand shares across measurement days show how the competitive field changes – the only view in this area with a timeline. It reveals whether a brand newly appears, gains citations or loses them.
Where your brand loses out.
An action list of the questions in which competitors are mentioned or cited and your brand is not. It also pins down the exact question in which the difference showed up.
Unlisted providers are a suggestion, not a fact.
By design, brand detection only recognises what has been stored. That is why the analysis also reports providers that appear in answers but are not yet on the list. They are not counted, but put forward for you to add.
If you add a provider as a competitor, the domain is required. All stored answers are then re-evaluated retroactively.
Favicons are loaded directly from each domain, never through a favicon service. Such a service would receive the customer's competitor list with every page view.
Not just: who gets mentioned? But: who does the AI recommend – and where?
An honest picture of the competition in AI answers, with a concrete list of the questions worth targeting.