Care plan updates, risk reassessments, and policy changes route to the correct named role.
AI proposes. A registered professional disposes.
CareAs never auto-applies a clinical or safeguarding change. The platform proposes, the registered manager or nurse approves, and the audit ledger records the human reviewer, the reason, and the version of the model used.

Care plan updates, risk reassessments, and policy changes route to the correct named role.
Approvers record reason text and clinical rationale before a change persists.
Configurable dual approval for medication, DoLS, and safeguarding domains.
If a model proposes and a human is unavailable, the record waits. It does not drift.