Predictive Claim Complexity, Fraud and Subrogation

Whether analyzing claims for potential fraud, subrogation potential or triaging claims for complexity, time is of the essence. Shutting down fraud or pursuing recoveries depends on quick and accurate identification. In addition, any system that performs this function must provide transparency (and traceability) into key data elements that drive the decisions.

Predictive subrogation is the first platform that performs all these functions in a purpose-built solution that can be implemented in just a few days.

Predictive subrogation analyzes information that is locked up in claim notes, police reports, medical billing (or any other unstructured claim data). It finds patterns in the data using built-in search terms and fully automated, advanced text analytics (topic modeling). When combined with other structured claim data (accident codes, at fault indicators, etc.), the AI machine learning in Athens can accurately predict future claim fraud, subrogation potential or complexity with a very high degree of accuracy. In addition, Athens gives insight into how it makes predictions pointing examiners to the exact indicators that were used in the predictive process.

Predictive Subrogation can greatly increase your identification and recoveries with a very minimal investment of your time and money!

Click for full-sized image
Predictive subrogation not only predicts the subrogation/fraud/complexity potentials but it also gives insights into the key variables [claim data indicators] that support the prediction. This includes both positive and negative indicators. This allows claim examiners or the SIU to quickly focus investigations/reviews.
Click for full-sized image
Predictive subrogation comes preconfigured for several insurance lines (e.g. auto collision subro, auto PIP fraud, etc.). QueBIT’s deep experience has resulted in a compliation of highly effective search terms. In addition, predictive subrogation uses text analytics to find new search terms (or topics) that are automatically extracted from claim notes, police reports, etc. that add additional lift to the predictive models.

– Greater than 95% identification
– Low false positive rate < 10%
– Fully automated claim identification
– Ability to scale to large claim amounts with frequent claim re-evaluation
– 10x or better ROI (first year)



    Your Name*

    Your Email*


    Your Message

    This form collects your contact information so that we can correspond with you. For more information on how we are committed to protecting and respecting your privacy, please review our Privacy Policy.

    I consent to QueBIT collecting my contact information.