AI visibility tracking

See what AI systems actually answer about your products

We generate the questions shoppers really ask, run them on a schedule, and keep the full answer so every number is auditable.

Nine kinds of buyer intent

Discovery

best waterproof trail running shoes

Recommendation

which running shoe should I buy for wet trails

Comparison

Trail Runner GTX vs Speedgoat

Problem / solution

my running shoes soak through in the rain

Purchase

where to buy Trail Runner GTX

Price

best trail running shoes under $150

Brand

is Summit a good running shoe brand

Alternative

alternatives to Salomon trail shoes

Feature

trail shoes with a rock plate

Examples are illustrative. Your prompts are generated from your own product data.

Mentions and citations are not the same thing

A mention is your brand, product or store name appearing in the answer text. A citation is the answer pointing at your URL — either in the provider’s own source list or as a link in the response body. Tools that report both as “citations” inflate every number on the page.

We store them separately, each with the sentence or source that produced it, and weight citations more heavily because they are verifiable and a reader can follow them.

What we monitor, and how honestly

OpenAI, Google Gemini and Perplexity, through their official APIs with search grounding enabled where the API supports it. We label all of it as API-based monitoring, because it is not a capture of the consumer apps. Personalisation, memory, and interface-specific ranking are not reproduced, and answers vary between runs.

A provider without working credentials is shown in your dashboard as unavailable. We would rather show one real source than three invented ones.

Positions, where they exist

Some providers return an ordered source list; there, the position of your first citation is meaningful and we record it. Where the response has no ordering, we do not invent a rank. An empty position column is an honest one.

Competitors

Add competitor domains and every monitoring run tells you, per prompt, which of them the AI system cited and which it did not. This comes from the same stored responses — we do not crawl competitor sites to produce it.