Search results

In the current global economy, the survival of an insurance company depends on its ability to respond to the customer demands. One of the demands of customers is efficient settlement of insurance claims. All insurance companies face the conflicting goals of authenticating claims and settling claims quickly. The use of human adjusters in claim settlement process leaves room for subjective judgment and the use of discretion while finalizing a claim. It has been observed that the opportunity exists for claim adjustors to settle insurance claims in favor of the claimants simply by colluding with the claimant and sacrificing the monetary interest of the insurers. The increasing cost of human experts for authentication (fraud detection) has led many companies to develop technological solutions such as expert systems to assist in the authentication of processed claims.

We have used fuzzy math in combination with the expert systems technology to design this model system.

We develop a fuzzy logic based expert system that can identify and evaluate whether elements of fraud are involved in insurance claims settlement.

We could not obtain real life data from any one of the insurance companies even after various attempts in Canada. Canadian Privacy Legislation does not permit these organizations to share them with any one.

This expert system can help decide if settled claims are genuine or if an element of fraud might exist which needs substantive testing by an auditor. The proposed methodology has been illustrated with an example that tends to model insurance claims in general.

The model designed in this paper is original and carries a substantial value to internal/external auditing professional who has access to these data to train the inference engine.