Crisis-Line Protocol.
Last reviewed · May 19, 2026
Summary
Active-risk language — suicidal ideation, overdose risk, immediate harm — triggers DIAL3D's crisis-line protocol. The agent stays on the line, drops to a slower trauma-informed cadence, and warm-transfers to a human within three seconds. After-hours fallback is 988 unless your facility has configured a different escalation path.
This page documents how the protocol works, what triggers it, and how customers can configure it for their facility.
Trigger detection
The agent runs a real-time classifier trained on substance-use-specific crisis language. Triggers include direct statements of suicidal ideation, planning language, overdose context, and immediate-harm signals. The classifier is tuned against false negatives — we accept some false positives to minimize the chance of missing a real crisis.
The classifier runs on every conversation turn. Trigger latency is under 400ms; total time-to-protocol-activation is under one second.
Activated behavior
On trigger: (a) the agent's voice immediately drops to a slower, gentler cadence; (b) the agent acknowledges the patient's statement without minimizing or amplifying; (c) the agent offers immediate human handoff; (d) a warm transfer fires to the configured target.
During business hours, the target is typically your facility's designated crisis-trained clinician. After hours, the default fallback is 988 (the national crisis line). Customers can configure alternative paths — on-call clinician, 24/7 nurse line, etc.
What the agent does not do
The agent does not provide clinical advice on managing the crisis. The agent does not attempt to talk a patient out of suicide or substitute for a clinical intervention. The agent does not document a clinical risk assessment.
Crisis-protocol activation is a routing decision, not a clinical decision. Clinical work happens after the warm transfer.
Customer configuration
During onboarding, your facility configures: (a) the primary crisis escalation target during business hours, (b) the after-hours fallback, (c) the language at which the agent introduces the human handoff (some facilities use specific phrasing tied to their clinical training), and (d) any special routing for adolescent patients or other populations.
Configuration changes take effect within 24 hours.
Documentation + audit
Every crisis-protocol activation is logged with: trigger language, timestamp, classifier confidence, target of warm transfer, time-to-transfer, and outcome of the transfer (accepted, missed, fallback engaged).
Documentation is available to the customer's clinical and compliance teams for audit. Aggregate (de-identified) metrics inform our model monitoring.
False-positive handling
If the agent activates the crisis protocol on a non-crisis call (a patient mentions a past suicide attempt in a routine intake, for example), the warm transfer still fires — we prefer over-routing to under-routing on this category.
False positives are tracked and inform classifier tuning. We do not punish patients with longer hold times or additional screening to reduce false positives.
988 integration
For after-hours fallback or when no human is available, the agent routes the caller to the 988 Suicide & Crisis Lifeline. Routing is direct-dial — the agent stays on the line until 988 picks up and confirms the handoff.
Customers can opt out of 988 routing in favor of a contracted 24/7 clinician service if they have one.
Need this document, a security questionnaire, or a deeper conversation?
Need to discuss crisis-protocol configuration for your facility, see the trigger taxonomy, or review activation logs? Our clinical-product team responds within one business day.
Contact our compliance team