AI Search · 6 min
There is no rank position in an AI answer, so visibility is sampled rather than measured: a fixed set of buyer questions, asked repeatedly across the assistants your market uses, recorded over time.
How it works
Twenty to forty questions someone would actually type or say, covering the decision: what a thing is, how to choose, who provides it, what it costs. Fix the list, because changing it breaks comparability.
The third is the most useful and the most often skipped.
Responses vary between runs for the same question, and personalisation and model updates both move results. Anyone reporting AI visibility to a decimal place is describing noise with unearned precision.
Adding one question to your enquiry form about how someone found you produces better evidence than most tooling.
If your server logs show AI crawlers receiving errors or thin HTML, sampling will confirm absence without explaining it, and the fix is technical rather than editorial.
Because generation is probabilistic and personalisation, region and model version all affect it. That is why this is sampling: read the trend across repeated runs, not any single response.
Which sources are cited instead. That tells you who the systems currently trust on your subject, which is far more actionable than your own absence.
There is no rank position in an AI answer, so visibility is sampled rather than measured: a fixed set of buyer questions, asked repeatedly across the assistants your market uses, recorded over time.
Connected
You have read what we think. If you want to know what it means for your case specifically, describe it and we will tell you which parts of ai visibility measurement guide actually apply — and which do not.
