Audit Dashboard
AI-powered fraud detection & investigation tools
Debug: Cases loaded: ,
Active Tab:
Total Cases
504
Open + In Review
High Risk (ML 80%+)
127
Priority investigations
Avg Resolution Time
4.2 days
ā 15% vs last month
Recovery Amount
$48,350
This month
š¤ ML Model Performance
Model Accuracy
94.2%
Gradient Boosting Classifier
Training Dataset
50,482
Historical cases
False Positive Rate
5.8%
ā 2.3% vs baseline
Top Risk Indicators:
- ā¢ Variance % above route baseline (42% avg importance)
- ā¢ Conductor historical pattern (28% avg importance)
- ā¢ Void/refund anomaly detection (18% avg importance)
- ā¢ Depot cluster analysis (12% avg importance)
Recent Activity
John Banda closed CASE-2025-087
$1,200 recovered ā¢ 15 minutes ago
Mary Phiri picked CASE-2025-093
High-risk cash variance ā¢ 1 hour ago
ML Model flagged CASE-2025-094
92% confidence ā¢ Conductor C-112 ā¢ 2 hours ago
š„ Top Priority Cases
š¤
Unassigned
Case Queue
Sorted by ML Risk Score (highest first)
š¤
ā¢
š« E-Ticket
ML Confidence
š
No cases found
Cases will appear here when flagged by the system
Cases Submitted to Director
Track approval status of your high-risk case submissions
ā¢
Variance: $
Pattern:
Auditor:
Director Notes:
ā³ Awaiting review
ā
Approved
ā Rejected
š¤
No submitted cases
High-risk cases you submit will appear here for tracking
Auditor Performance
JB
John Banda
Senior Auditor
47
Cases
42
Resolved
3.8d
Avg Time
MP
Mary Phiri
Auditor
38
Cases
35
Resolved
4.1d
Avg Time
TM
Thomas Moyo
Junior Auditor
29
Cases
23
Resolved
5.2d
Avg Time
Recent Team Activity
John Banda
resolved
CASE-2025-087
15 min ago
Recovered $1,200 ā¢ Cash variance investigation complete
Mary Phiri
picked
CASE-2025-093
1 hour ago
ML Score: 88% ā¢ High-risk cash variance
Thomas Moyo
added progress note to
CASE-2025-091
2 hours ago
"Cross-referenced e-ticket data. Found 12 missing transactions."
John Banda
submitted to director
CASE-2025-089
3 hours ago
High-risk case requiring director approval ā¢ $1,250 variance
High-Risk Cases for Director
Only director can approve high-risk audit findings and investigations
HIGH RISK
Findings Summary:
Case details
Summary
š¤
AI Risk Assessment
Top Contributing Risk Factors:
Model: Gradient Boosting Classifier ā¢
Accuracy: 94.2% ā¢
Trained on: 50K+ cases
š
Investigation Progress
š
Investigation Tools
For highly flagged conductors, print trip tickets and compare with conductor's submitted logbook
Conductor
Trip ID
This will generate official ticket records for comparison with conductor's physical logbook
Workflow actions
Wire these to audit-service endpoints later.