Health systems have long used note quality rubrics — completeness of the history, presence of an assessment, documented reasoning — to audit clinical documentation. Those rubrics were designed to catch the failure modes of a tired human writing at the end of a shift.

Generated notes do not have those failure modes. They are complete, structured, and consistently include every element the rubric asks about. Average scores at systems with mature ambient deployments have risen sharply, and quality directors are increasingly skeptical that the rise means anything.

A pilot audit at one system attempted to score notes on a different axis: whether a covering clinician reading the note cold could reconstruct the treating clinician's actual reasoning. Generated notes scored worse than the human-written baseline, primarily because the reasoning was frequently implied by the assembled findings rather than stated.

The proposed rubric revision adds a criterion for explicit clinical reasoning attributable to the author. It is harder to score, requires a clinician reviewer rather than a checklist, and is the only element of the audit that current tooling cannot inflate.