Compare the runs you planned with the answers you collected and analysed. A worksheet for measuring AI monitoring completeness before you read any visibility number.
A chart-review checklist for AI visibility reports: denominators, missing periods, collection changes and axis choices, so a small move is not drawn as a big one.
Share of voice is a ratio against the competitors you track, so it can climb while your own mentions do not move. Here is why, with a worked example and a two-metric reporting pattern.
Seasonal buying questions distort an always-on AI visibility trend when they are pooled in. Tag them, give them dates and their own comparison period, and use the calendar template.
There is no universal number of days. A readiness checklist shows when a new project's results move from a first observation to a provisional read to a baseline you can report.
Adding prompts changes what your AI visibility rate measures. Freeze a comparison cohort, keep growing the live set, and use a copyable cohort register to report both honestly.
App answers and API answers are different populations. A copyable segmentation convention and migration note template keep your AI visibility reports from mixing them.
Two teams can read the same AI answers and report different citation rates. Work through answer, citation, URL and domain denominators with a worksheet that fixes the definition beside every number.
Days when no AI answers were collected can bend a visibility trend without any real change. A worked example and a copyable missing-data checklist show how to choose a comparable window.
A missing AI result can mean your brand was not named, the engine showed no answer, or collection failed. A run-log method and legend keep the three apart.
A drop or jump in AI visibility often has nothing to do with your site. Work through a ten-step checklist covering windows, filters, engines, failed runs and prompt edits.
Log the edit date, then compare the same prompts, engines and window before and after. Improvement, no change and too little evidence are all valid results.
Follow visits from AI engines to a recorded conversion in GA4, understand what goes missing along the way, and fill in a worksheet that stops short of prompt-level sales claims.
Brand visibility, citation rate, and share of voice divide by different things. One small labeled dataset shows what each number answers, plus a reporting cheat sheet.