Most AI-visibility trackers ask each question once per engine per day and print the result as a score. This page sets that pattern next to a repeated measurement — without naming anyone, because the arithmetic is the same on every price list.
| AnswerRange | Single-sample tracker | |
|---|---|---|
| How often each question is asked | Two or three times per platform per run | Once per engine per day, on most published price lists |
| What the number means | Share of answers citing you, with a 95% interval | Whether you appeared in one answer, shown as a score or rank |
| Day-to-day movement | Reported monthly; a move is reported only if the test supports it | Daily; single draws move on their own by several points |
| Who chooses the questions | You, or your Search Console; the corpus is labelled on the report | The tool, from a keyword list; provenance not shown |
| Before-and-after | Same questions, same repeats, pre-registered test | Two scores compared by eye |
| Sources shown | Every answer and every cited URL, exportable | Varies; often a mention count |
| Method | Published open access, three studies with data | Proprietary |
| Price basis | Per measurement, public | Per prompt-engine-day, public |
The right-hand column describes the common pattern across tools we examined while writing the measurement study. Run the check yourself: prompts × engines × responses on any pricing page, divided by prompts × engines, is how many times each question is asked. If it comes to one, the daily score is a single draw.
Watching many keywords cheaply, brand-mention alerts, a daily feel for a volatile market. Treat every daily move as weather.
Any number you will show a client, a board or a budget owner; deciding whether work changed anything; choosing which lost questions to fix.
Alerts from the tracker, decisions from the measurement. The two are not competing for the same sentence in your report.