Skip to content
[REGENT]
Governed Engagement Engine
This message was governed.
Four phases ran in a fixed order before dispatch. The patient sees only the left panel.
Phase 01 · Context — pre-judgment hard stops
Contactability confirmed
Active channel on the tenant connection · subject reachable
No crisis flag · no global opt-out
Non-overridable hard stops. Had either been present, the operation would have resolved here and the judgment engine would not have run.
Phase 02 · Judgment — the AI proposes
AI judgment: SEND
Lab results ready · follow-up due · model version pinned to this decision
Phase 03 · Governance — your rules run after the proposal
Proposal-specific consent confirmed
Consent scope covers this proposal · patient opted in Jul 14, 2026
Content approved
Results notification · permitted topic class
Frequency and window respected
First contact in 7 days · cap 1 of 3 used · inside the configured send window
Phase 04 · Action — the governed outcome
Dispatched · accepted by provider
Governed dispatch from the tenant's own Meta Business identity. Acceptance is not delivery confirmation.
Exact inputs, policy versions, model version, gates and recorded outcome are preserved and integrity-verifiable on demand. Your compliance team can inspect any decision at any time — the record is tamper-evident with governed redaction.
View full audit trace
[REGENT]
governed engagement engine
Run a Guided Eval on a public website →