TECHNOLOGY
State-of-the-art technology in order to increase the chance of catching the fraudster.
Developments in the field of software and technology mean that insurers are increasingly carrying out projects to support the detection of potential fraud cases. The demand for detection software is increasing. An effective example of fraud detection is the use of automated detection rules as a filter during bulk processing. The requests that are filtered out deserve further investigation and are removed from the bulk flow. Checking these takes place in the background. The employees receive notification of this but do not experience any delay because of it and can continue with their tasks. In this way there is no risk for the productivity. As well as this form of detection there is also a type in which a risky part of the portfolio is analysed and isolated periodically. This can be effected by using batch-wise detection or with the help of data mining. Friss provides various software modules for supporting the process of fighting fraud. The most important modules are: Friss Detection, Friss Management, Friss Dashboard, Friss Case Management, Friss Data mining and Friss E-learning. A further explanation of these modules will be given in the following paragraph. In the current situation in insurance companies the recognition of fraud indicators is very often the responsibility of employees and there is hardly, if any, automated support. This means that it is at present a cumbersome and labour intensive process. Numerous indicators have to be checked manually, claim files completed or external sources consulted to check whether a possible fraud has been committed. In practice this is only done by employees who have an affinity with the subject of fraud. The pressure for production is otherwise too great and too important. Our thesis is that it is possible to increase the level of detection by several factors by employing an automated check of around thirty fully developed and validated indicators with the use of background data sources.