Detect Fuel Spend Anomalies Before They Cascade
ML-driven weekly briefs for WEX/EFS/Fleetcor fuel card programs, built for solo operators running 5-100 vocational vehicles.
$pip install fuelguardiq && fuelguardiq init Weekly Fuel Spend Cascade Briefs
ML-driven detection of unusual fuel purchase patterns across your vocational fleet with ranked intervention sequencing.
- Statistical outlier detection across vehicle clusters
- Configurable sensitivity thresholds
- Weekly digest via email or webhook
```python
import fuelguardiq as fgq
# Initialize with your WEX/EFS/Fleetcor CSV export
client = fgq.Client(api_key=os.environ['FGQ_API_KEY'])
# Run weekly anomaly scan
brief = client.weekly_brief(
fuel_csv='./transactions/export.csv',
threshold=1.5, # 1.5 standard deviations
vehicle_segment='vocational'
)
print(f"Detected {len(brief.anomalies)} anomalies across {brief.vehicle_count} vehicles")
```Intervention Sequencing Briefs
Ranked action items for fuel card optimization based on anomaly severity and estimated cost impact.
- Prioritized by financial impact
- Includes estimated recovery amounts
- WEX/EFS/Fleetcor card-level granularity
```python
# Access ranked intervention list
for idx, intervention in enumerate(brief.interventions, 1):
print(f"{idx}. {intervention.card_id} | {intervention.type}")
print(f" Est. recovery: ${intervention.estimated_savings}")
print(f" Confidence: {intervention.confidence_score:.0%}")
```CSV Upload Workflow
Direct import from WEX/EFS/Fleetcor fuel card exports. No API access required—upload CSV, get brief.
- Standard WEX/EFS/Fleetcor export formats supported
- Automatic date/unit normalization
- Historical baseline comparison
```bash
# 1. Export fuel card transactions from WEX/EFS/Fleetcor portal
# 2. Upload to FuelGuardIQ
fuelguardiq upload \
--card-program wex \
--file transactions-q4.csv \
--vehicles 47
```FAQ
How does anomaly detection work?
FuelGuardIQ applies statistical outlier detection to your transaction history. It computes per-vehicle baseline profiles using rolling 8-week windows, then flags transactions exceeding configurable standard-deviation thresholds. For vocational fleets (construction, delivery, HVAC), the model adjusts for seasonal load patterns and geography-specific fuel price variance.
What CSV formats are supported?
FuelGuardIQ parses standard export schemas from WEX, EFS, and Fleetcor fuel card platforms. Required columns include transaction date, card ID, vehicle ID (if available), gallons dispensed, total spend, and station location. Optional fields like odometer and driver ID improve detection precision. The upload wizard validates schema completeness before processing.
Do I need API access to my fuel card portal?
No. FuelGuardIQ operates on CSV exports only. WEX, EFS, and Fleetcor all provide manual or scheduled CSV exports from their operator portals. This approach avoids API integration complexity and per-call fees, though it requires a weekly manual export step. API access remains on the roadmap for enterprise tiers.
What fleet sizes does this target?
FuelGuardIQ is designed for fleets of 5-100 vocational vehicles with $500K-$10M annual operating budgets. The model performs optimally with at least 12 weeks of transaction history per vehicle. Larger fleets or mixed asset types may require segmentation configuration. Fleets below 5 vehicles may lack sufficient baseline data for reliable anomaly detection.
Do I need a fuel analyst or credentials?
No professional credentials are required. FuelGuardIQ is built for solo fleet operators managing without a dedicated fuel analyst. The system surfaces anomalies and ranks interventions, but decision authority remains with the operator. No JD, CPA, or fleet management certification required.
Start Your Weekly Fleet Brief Today
Ship MVP in 2-6 weeks. No dedicated analyst required.