CiteTrack separates a planned observation from a recorded result. The planned denominator is active buyer questions multiplied by enabled AI services. Recorded checks are results actually stored for that report window; the difference is shown as unrecorded coverage.
Crawler integrity
Measurement surfaces that claim verified AI crawler activity require a known bot and exclude spoofed rows. A matching user-agent alone is weak evidence, so claimed bot traffic is not allowed to inflate crawl totals or the opportunity funnel.
Prompt and method context
Reports preserve the latest check, service mix and plugin method version. Frozen baselines retain that context with the evidence. When comparing periods, first confirm that buyer-question coverage, enabled services and recorded/planned checks are comparable.