The job of this console
Read flagged patterns, open the underlying leave, and decide with a human. Tune rules via API/seed, not by deleting leave rows.
HRMS · Leave
Leave Anomaly & Fairness
Register view for anomaly rules, flagged reviews, and fairness snapshots — no self-serve write UI; seed via API. Flags never auto-reject leave.
Register view — seed via API. Flowcharts: rule → detect → review.
Fairness signals
Flags for review — they never auto-reject leave
Leave anomaly intelligence is rules, flagged reviews, and fairness snapshots. There is no self-serve write UI; seed via API. Flags never auto-reject leave.
Admin → Leave anomaly intelligence
Route
/admin/leave/anomaly-rules
HR Admin · Org Admin
Seat at the table
HR administrators. Employees are not scored in a public leaderboard here.
The working rhythm
Open the review list, inspect a flag, open the leave, act on the request using normal approval. Adjust rules only with care.
Moments that bring you here
A cluster of Friday-only leave. A fairness snapshot before peak season. After a policy change that looked like gaming. When someone asks why a flag appeared.
ⓘ
Signals, not a veto
Anomaly never auto-rejects. ApprovalAuthority and managers still decide.
Flowcharts
Each diagram auto-plays steps when it scrolls into view. Use Play / Pause, or click a step to jump.
1 — Anomaly detection
Rules scan leave activity and open review items.
flowchart TD
R[Anomaly rules] --> S[Scan leave data]
S --> F{Flag?}
F -->|No| Ok[No action]
F -->|Yes| Q[Review queue]
Q --> Act[HR action]
- 1Anomaly rulesThresholds for frequency, clustering, or balance drift.
- 2Scan leave dataScheduled or on-demand scan.
- 3Flag?Rule match opens a review item.
- 4No actionWithin expected patterns.
- 5Review queueAnomaly review console.
- 6HR actionDismiss, investigate, or adjust policy.
2 — Fairness analytics
Team-level views for approval latency and rejection rates.
flowchart TD
Data[Leave + approvals] --> Fair[Fairness analytics]
Fair --> Lat[Latency]
Fair --> Rej[Rejection rate]
Fair --> Dist[Distribution by team]
- 1Leave + approvalsSource operational events.
- 2Fairness analyticsAggregate dashboards.
- 3LatencyTime from submit to decide.
- 4Rejection rateApprover outcomes by team.
- 5Distribution by teamSpot skewed patterns early.
Overview
Anomaly rules, review, and fairness analytics help HR spot unusual patterns without auto-rejecting leave. Screens are registers only — no self-serve write UI; seed via API. Do not store raw sensitive medical content in analytics stores.
How to use
1. Open anomaly rules
Admin → Leave → Anomaly rules register.
2. Review rules
Inspect seeded rules; create/update via API.
3. Work the review queue
Inspect flags; dismiss or act via API.
4. Fairness snapshots
Inspect analytics registers for latency and reject patterns.