Tenant submits urgent request
A tenant submits a request with emergency keywords (“fire,” “flood,” “gas leak”) and Urgent priority. The emergency detection system activates automatically.
Two-vote classification
The system uses two signals: keyword regex detection and AI risk assessment. If both agree it’s a true emergency, the request is classified as EMERGENCY with a red pulsing badge — not just yellow “Urgent.”
Emergency notification
You receive an EMERGENCY email with critical severity. Your dashboard shows a red EMERGENCY badge and HABITABILITY RISK card with matched keywords, confidence score, and recommended response time.
30-minute escalation
If you don’t respond within 30 minutes, an automatic escalation re-ping is sent. This fires once per request — it’s a single safety net, not a loop.
Non-emergencies filtered out
Urgent requests without emergency keywords (like “squeaky door”) remain as normal URGENT priority with standard notifications. The two-vote system prevents false positives from cluttering your emergency queue.
Common questions
What keywords trigger emergency detection?
Fire, flood, gas leak, no heat, no water, sewage, electrical hazard, carbon monoxide, and similar habitability-related terms. The AI assessment also weighs context, not just the presence of a keyword.
Can I turn off emergency alerts?
No. Emergency alerts are always active for habitability safety and cannot be disabled. You can manage other notification types in your settings.