AI model that leads to fraud elimination in an electric power distribution network

,

fraud detection
Discover Suspicious Electric OfftakeDiscover Suspicious Electric OfftakeDiscover Suspicious Electric Offtake
Ing. Jarmila Verešová

“Analytical tool from BigHub helped us to detect suspicious electric offtakes. Moreover, the fusion of isolated data sources into one place created an option for us to further analyze the data and work on behavioral segmentation of our customers.”

Ing. Jarmila Verešová
Ing. Jarmila Verešová
Head of the Inspection team, VSD, a.s.
Project background

Electricity supplier, VSD, wants to discover unauthorized energy offtakes by „Analytics First“ approach. This helps to decrease in-fields inspectors allocations and let them focus on the most relevant cases.

Our approach

Unauthorized power offtake is accompanied with consumption drop. The solution is analytical tool based on AI algorithm which analyzes data from consumption meters and detects „black offtakes“.

The data provided by the client allowed model to reconstruct past frauds and train AI to find other similar cases based on key auto-learned symptoms.

This helps Inspection team to make more data driven choices.

Part of the analytical tool is a dashboard with ability to visualize the results.

Benefits of our approach

Cost savings due to fraud elimination
Fast anomaly detection with AI model
Visualization of the results
Detection of suspicious cases „a priori“ as oppose to „ex post“
Integral view on energy consumption in context of business type, turnover, ownership structures and peer benchmarking
Advance tool to analyze data fused from different sources on one place
Discover Suspicious Electric OfftakeDiscover Suspicious Electric OfftakeDiscover Suspicious Electric Offtake
Detection of Fraud Offtakes

Efficient suspect targeting

  • Discover sudden consumption drops
  • Elimination of natural anomalies
  • More accurate inspections
Inspection Team Tool

Cost savings on inspections

  • AI modelling based on past frauds
  • Finding similar cases with machine learning
  • Comprehensive conclusions to present
Next solution
+420 720 732 783Vaclavske namesti 802/56, 110 00 Nove Mesto
ID 05388066
VATCZ05388066