How is the condition and quality of IT-systems insurance company affect its ability to withstand an insurance fraud? As with the help of modern IT-systems to detect suspicious losses and prevent unjustified payments?
The problem of insurance fraud today is relevant to most mature and emerging insurance markets, and the Russian insurance is no exception. Fraud can occur in various types of insurance, at all stages of the insurance cycle - from purchase of the policy to the calculation of service organizations. But the most pressing, of course, the issue of anti-fraud is on the stage of the settlement of losses.
Developing an effective strategy to combat fraud - one of the most urgent challenges facing today's insurance companies. The first step towards its solution is to develop efficient methods to detect fraud. Here to help insurers to come modern information technology. Specialized IT-system will not only help to quickly identify the illegal scheme, but also provide specialist security services all the information necessary to investigate the facts and the evidence of fraud. Thus, when a suspicious case, the system will mark the reasons for the "trip" will give full information on the occurrence, the insurance contract, will show similar cases in the past and can offer a step by step plan to investigate the event.
Modern IT-systems for the detection of fraud are based on powerful analytical packages that allow for regular review of available historical data (for contracts of insurance, losses, etc.) and build on the basis of the analytical model. Immediate implementation of fraud detection, preceded by the project, whose goal is a comprehensive analysis of historical data. In our experience, the duration of this project is approximately four months. As a result, insurance companies get ready to use analytical models, as well as a prototype system that was put into operation with minimum effort. The project prepared representative samples of historical data, carried out by their purification, enrichment and transformation, and then conducted a comprehensive analysis. It includes, as a rule, the various methods: validation of expert rules, statistical analysis, cluster analysis, modeling of social networks.
In our view, when working on the implementation of fraud detection systems, Russian insurers will inevitably be faced with a lack of quality of historical data. This is due to the Russian insurance market, in particular, the model of "off-line" sales (where the policy is issued to the agent of the paper form, and then introduced into the information system of teller of the insurance company), as well as the computerization of weak insurance sector as a whole. Insufficient quality of historical data makes the task of building adequate analytical models to identify poor-selling fraud.
What a way to solve the problem of insurance fraud for the Russian insurance market can be found in such conditions? One promising option would be to create a centralized information system for fraud detection systems in place within each local insurance company.
Using a centralized solution will bring a much greater effect for individual insurance companies, and for the insurance sector as a whole. The total investment in the implementation of a centralized system will be less than the creation of disparate tools within the insurance industry. Efficiency will also be higher in the first place, due to larger amounts of data, and secondly, through the ability to identify types of fraud affecting several insurance companies (eg, "double insurance"), and thirdly, because of the global blacklist unscrupulous insurers and counterparties, thereby preventing the possibility of "transition" from one scam to another insurance company. Such a centralized system may be, for example, is implemented on the basis of a professional association of insurers who will carry out its operation and provide the appropriate services to insurance companies participating.
Of course, a system of fraud detection at the level of the market - a task far from simple. In addition to purely technical aspects (choice of optimal technology platform and its configuration, development of data exchange interfaces, etc.) need to solve a lot of organizational issues, in particular, to develop a business model of the proposed solutions:
• calculate the investments in creating a system of cost recovery, to determine the conditions and cost of using the system by insurance companies that have invested in the creation of a system, and other insurers;
• calculate the expected economic effect of implementing this system for insurance companies, correlating benefits by reducing unnecessary insurance premiums to investment swindlers to create a system and the cost of its operation;
• Identify a model of operational cooperation between all parties during the operation of the system, describe the business processes, to prepare regulations, etc.;
• To ensure the interest of insurance companies not only in using the system, but also to provide historical data for its content;
• To ensure confidentiality of data - both in terms of personal data protection of policyholders and for the protection of commercial interests and portfolios of insurance companies;
• analyze the legal aspect of the project, its operation in terms of legislation and regulations;
• to evaluate the possibility of expanding the use of the system with the CTP (which is a type of insurance is a priority) to the hull and other public insurance.
All of these tasks, despite the apparent complexity can be addressed effectively provided a comprehensive approach to the implementation of the project, its clear organization and good management. It is very important pragmatic approach is needed to achieve "quick wins» («quick wins») in the shortest time frame, otherwise, the project could turn into "unfinished."
During the implementation and operation of the fraud the insurers may need to finalize their transactional systems, which was originally recorded data on insurance contracts and losses, in order to provide the necessary input attributes and the correctness of the data. However, there should be a balance between the volume of recorded information (and, accordingly, a time of specialists necessary for its entry), and the expected economic benefit. The cost of increasing labor costs for data entry, as well as the costs of the investigation identified instances of potential fraud should be borne by future reduction in benefits in terms of proven fraud.
Ultimately, the introduction of a centralized system to identify cases of insurance fraud will create conditions for a planned reduction in the proportion of unjustified payments associated with fraud, which now, according to various estimates, more than 10% of the total payments to insurance companies.
No comments:
Post a Comment