FRAUD DETECTION SYSTEM
better identification of insurance fraud
using advanced prediction models
Better detection of insurance fraud via an automated machine learning system enables a two-thirds more successful follow-up investigation to be achieved compared to the traditional approach. The result is more than double the profitability of the process, achieved by more efficient use of investigator capabilities and higher protected value. Correct identification of suspicious claims enables a decision about the fulfilment to be made within 3 minutes.
PosAm Fraud Detection System is focused on tracking suspicious facts throughout the life of an insurance contract, from creation to termination. It searches for non-standard behaviour of clients, salespeople, and partners.
The unique predictive model for fraud detection has a success rate of 80%, according to the methodology for assessing the success of a classification test. Under this methodology, the PosAm prediction model ranks as excellent, as it reached the area under the ROC curve of 0.91.
The automated machine learning system is significantly better in identifying potential fraud. The result is an increase of success of the follow-up investigation by two thirds. These results are achieved thanks to the optimization of rules using prediction models. They are being constantly dynamically modified on the
basis of feedback, which allows the system to learn and improve
its results.
A 66% increase in the amount of actually confirmed fraud has a significant positive impact on the efficiency of investigative work. Expenses incurred are balanced by higher protected value and higher overall profitability.
Automatic detection is not subjected to the human factor, which gives it a high degree of credibility and reliability. Honest clients are not unnecessarily harassed, which positively affects their customer experience. It is a great benefit for the insurer as well as the client if the claim can be decided within a few minutes.
PosAm Fraud Detection System significantly facilitates identification of fraud and increases the efficiency of fraud investigation. It helps to ensure quick fulfilment in legitimate cases and minimize the number of illegitimate ones. It positively impacts the profitability, reputation and competitiveness of the insurance company.
Maximal amount of detected fraud with minimal burden
2/3rds higher success rate
of fraud investigation
Double profitability
of the process
It is possible to decide on the fulfilment in 3 minutes
Protecting the values of the insurance company for greater competitiveness
80%
of uninsured property owners try to
take out retroactive insurance
65%
of life insurance clients will not tell the
truth about their health
37%
of motor vehicle damage claims can be
classified as insurance frauds
20%
14%
of insurance frauds are committed by contractual partners of insurance companies
of all damage claims can be
classified as insurance frauds