June 10th, 2008. Sergei Ananyan of Megaputer Intelligence presents at the Fourth Annual National Workers’ Compensation Subrogation Strategies Executive Summit.
Insurance companies have to timely process overwhelming volumes of claims, frequently quite complex and lengthy. In this challenging environment, they can easily miss valid subrogation opportunities and thus not recover the money they are entitled to. Subrogations help recover substantial funds, but only a small percentage of cases represent valid subrogation opportunities. Currently, insurance companies task either individual adjusters or special recovery teams with determining the subrogation potential of handled claims. However, there is a general agreement in the industry that manual analysis of subrogation potential is rather inaccurate and time consuming, and a better solution is being sought.
Intelligent automated systems combining text mining and data modeling techniques proved to be efficient in identifying valid subrogation opportunities. Such systems exploit the joint use of linguistic and semantic text analysis techniques, advanced machine learning algorithms, and data cleansing and manipulation capabilities. Intelligent systems learn patterns characteristic for claims with high subrogation potential through the analysis of textual notes of historical claims with known outcomes. Performing joint analysis of patterns extracted from textual notes and structured attributes associated with the claim, the system develops a model for predicting subrogation potential of each claim.
The created model enables the automated ranking of claims with respect to their subrogation potential. The ability to focus on claims with the highest subrogation potential enhances the efficiency of work of human analysts and increases the recovery rate. An automated subrogation prediction system ensures the consistency of the analysis across all claims and across time. It helps minimize the number of missed subrogation opportunities.
Furthermore, one can task a subrogation prediction system to look at historical claims, the subrogation potential of which was already assessed by a human analyst. An automated subrogation prediction system can discover a significant number of good subrogation candidate claims, which had been missed in the previous manual analysis.
- Claims with high subrogation potential can be easily missed by a human analyst
- Systems based on data and text mining help automatically predict subrogation opportunities
- Such intelligent systems build predictive models based on historical data with known outcomes
- They improve the accuracy of subrogation prediction consistently across all claims
- They can detect valid subrogation opportunities even on claims previously discarded by a human