Incident-only scope: Guardrailed to workplace incidents; consent-first, optional anonymity where policy allows.

Native to Teams: Uses adaptive cards, chat nudges, calendar, and files — no new tools for HR, managers, or investigators.

Agentic loop: Intake signal → assess severity/policy → propose plan → act (assign, schedule, gather evidence, notify) → review weekly.

Policy engine: HR/EHS-defined templates, SLAs, escalation paths, and compliant prompt libraries for investigators/managers.

Privacy & audit: Minimal data retention, immutable activity log, chain-of-custody for evidence, outcome tracking (triage speed, closure time, recurrence).