Developed with the contribution of experienced claim handlers, Detector is a fast-growing, end-to-end platform that stops frauds and empowers experts to make the right decisions. Detector allows for greater precision, enabling the Insurer to concentrate investigative efforts on the most suspicious and complex claims.

Detector provides the Insurer with:

  • Scoring accuracy: Detector’s sophisticated multi-faceted approach, incorporating AI analytics, minimizes false positives and accelerates the decision-making process;
  • Easy to use platform: intuitive, self-explanatory and simple-to-use, Detector provides real-time proof of fraud;
  • Full automation: Detector claim scoring and investigation workflow ensures cost and productivity monitoring of case management processing.

The Add-on for Detector enables:

  • Seamless integration of Detector into Guidewire ClaimCenter; 
  • Instant scoring of the claim returned to ClaimCenter;  
  • Embedded UI into ClaimCenter, allowing in-depth, content-rich, analysis accessible without exiting ClaimCenter;   
  • Acceleration of the settlement process for genuine claims to enhance the customer experience; 
  • Compliance support on anti-fraud regulations. 

Scoring Accuracy

Detector includes hundreds of different criteria to establish how dangerous a claim can be.  It can ingest external data to further deepen the analysis. The criteria used by Detector are organized into five groups:

  • Heuristic, based on hands-on experience of claim handlers;  
  • Geographical, comprises analysis based on the geo-localization of each party (e.g. their socio-economic background, KP Risk-Map, etc.);         
  • Network Analysis, enables the creation of graphs that illustrate the relationship between all the parties involved in different claims;         
  • Machine Learning, a set of different approaches embedded in an all-encompassing predictor; 
  • Documentation & Photo Alteration Analysis, to establish that the documentation presented is authentic or whether it has been reused, downloaded from the internet, or altered.