AI Engine
AI Insights
Theft detection, predictive maintenance and optimisation models running continuously against the Data Lake.
Model · Theft Detection v3
AI Fuel Theft Detection
- PMP-021RISK 88Suspected siphoning eventconf 94%
- VEH-2210RISK 62Route abuse · 14% over baselineconf 81%
- TK-004RISK 76Ghost transaction · 980L unaccountedconf 89%
Model · RUL v2
AI Predictive Maintenance
- PMP-021P(fail) 71%9days remaining useful life
→ Schedule pump rebuild
- VEH-2305P(fail) 62%14days remaining useful life
→ Fuel pump service
- GEN-007P(fail) 41%28days remaining useful life
→ Injector inspection
Model · Optimization v1
AI Fuel Optimisation
- Consolidate Gauteng deliveriesR 218k/moconf 94%
- Re-route Fleet KZN via N3 corridorR 86k/moconf 88%
- Shift generator load Cape Town DCR 41k/moconf 82%
Incident Timeline · PMP-021
AI investigation: suspected siphoning
- 09:10:04Dispense 980L initiated · driver field = Unknown
- 09:10:06Flow rate anomaly detected (3.2σ over baseline)
- 09:10:08GPS correlation: no matching vehicle within 500m
- 09:10:11AI risk score raised to 88 · workflow triggered
- 09:10:12Dispenser PMP-021 auto-locked · manager notified