Guide · AI visibility
AI visibility tracker
A tracker is the monitoring half of AI visibility: it runs the same questions on a schedule so you can see movement rather than a single snapshot. The value is entirely in the consistency — same questions, same engines, same cadence.
What a tracker should record
Every run should keep the question, the engine, the date, the full answer text, which brands were named, and which sources were cited. Without the answer text you cannot audit the score, and you cannot see who was recommended instead of you.
Reading the numbers honestly
Citation rate moves on its own. Engines rewrite answers, roll out models and change source preferences constantly, so a brand can gain or lose several points in a week with no change on its side. Any claim that a tool caused an increase needs a control group — a comparable set of brands that got no work — measured over the same window.
Cadence and sample size
A useful rule: at least twenty questions, at least three engines, checked at least weekly. Below that, week-to-week changes are mostly randomness.
See where your brand stands
Free, no account: the questions you lose and the competitor named instead.
Common questions
- How many questions should a tracker cover?
- Enough to cover how buyers actually phrase the need — typically 20 to 60 per category, including comparison and alternatives phrasings.
- Can I track competitors too?
- Yes, and you should. The most actionable output is not your own score but the list of brands engines name in your place.
Comparing products? See the ranked list of AI visibility tools, including our own measurement of which tools AI engines name.