Each edition has its own scope
A research month identifies an edition of a study. Model output, source availability, and provider mentions may differ in later runs. Compare dates, response coverage, and methods before interpreting a change.
Changes concern the response set
An increase in recommendations does not establish better service quality. A decline does not establish worse delivery performance. Historical comparisons concern the collected AI responses unless another measure is explicitly documented.
Use the matching evidence
Archived articles and exported evidence should be read together. A SHA-256 hash identifies the relevant snapshot. Current provider information may differ from an archived edition.