Notification Scenarios

Ad-hoc notification dispatch without a scenario framework leads to inconsistent alert volumes, coordinator fatigue, and missed critical events. HLF's workshop feedback directly cited poor peer mentor follow-up as a top dissatisfaction driver - scenario-based reminders systematically address this without requiring coordinators to manually track individual timelines. For Blindeforbundet, automated reminders after 10 days of no contact on an open assignment are a compliance requirement, not a nice-to-have. Scenario configurability per organization prevents alert fatigue by letting each tenant tune thresholds to their operational cadence, reducing unsubscribe rates and maintaining notification channel health.

Scenarios are evaluated by a scheduled job running on the Next.js backend (Vercel Cron or an external scheduler), querying the database for entities matching trigger conditions and cross-referencing the notification_rules table for per-organization thresholds and enabled scenarios. Each scenario evaluation produces a notification event record before dispatch, enabling idempotency checks to avoid re-triggering the same alert within a cooldown window. The rule engine is data-driven rather than hardcoded - new scenarios are added by inserting rows into the scenario registry table, referencing the trigger entity type, condition expression, cooldown period, and default enabled state. Scenario evaluation logic must handle tenant isolation rigorously, never cross-joining data across organizations.