Clinical safeguards

Each safeguard, and what enforces it.

A safeguard is only as good as the mechanism behind it. Here is each one, beside the part of Faurenis that holds it in place.

The dentist accepts, or nothing becomes record
AI drafts enter a clinician-review lifecycle. No AI output becomes part of the official record until a dentist accepts it.
Enforced by the record lifecycle on the server. Acceptance is limited to the dentist role, and the app cannot skip a step.
No AI output accepts itself
Asking Faurenis for a draft is never a sign-off on it. Generating a perio report files it for review; it does not approve it.
Enforced by the record lifecycle, which has no path from a draft to accepted without a clinician.
Uncertain output is held back
A draft below the confidence floor, or flagged by an output rule, goes to review. A draft containing personal data that should not be there is blocked and never created.
Enforced by the output validators and the personal-data scanner on the AI routes.
Every decision leaves a reason
Accepting, rejecting or changing an AI draft requires a reason, stored with the decision.
Enforced by the record's transition log, mirrored into a hash-chained audit trail.
Patients see only accepted records
Draft or review-pending content is never included in a patient-facing document or export.
Enforced by patient-facing documents, which read the record's state from the server before they render.
A signed record cannot be rewritten
Once locked, a record cannot be changed. A correction is a new version linked to the original, and both are kept.
Enforced by a database trigger that rejects edits to locked records.

These describe how Faurenis works today. They are not a regulatory classification. Intended use and regulatory status.

Clinical safety

Ask how a safeguard works in your clinic.

We will walk your clinical lead through any of these, on your own workflow.