Ingosstrakh and GlowByte create an analytical platform to counter insurance fraud

2022: Plan to establish an analytical platform to counter insurance fraud

Ingosstrakh, together with the consulting company GlowByte, is creating an analytical system for countering fraud based on open source technologies. GlowByte announced this on December 7, 2022.

The created platform will reduce the amount of payments under fraudulent schemes thanks to a more accurate assessment of losses in automatic mode. This opportunity will appear thanks to machine learning models and scoring using graph analytics.

Ingosstrakh decided to improve the existing anti-fraud scheme, as it was based on static expert rules and required additional resources when conducting specific investigations. In addition, due to the lack of new solutions, not all cases of fraud were confirmed.

To solve the problem, we need a single interface for conducting investigations, which allows us to visually analyze the relationships for all to data that are available to the company. The system we are working on with GlowByte reduces the time identifications of complex fraudulent schemes from several days to several hours and improves the effectiveness of investigations. Also an important aspect of the system is its constant adaptation to the new methods of fraud that regularly arise on the market. We see the possibility of applying graph analysis in the insurance business, - said Alexey Vlasov, Deputy General Director for Retail Business at Ingosstrakh.

Automated fraud risk assessment is based on machine learning methods. They are much more accurate than expert rules. By analyzing loss information, a mathematical model allows you to identify hidden patterns and statistical relationships in data, a certain combination of which indicates a high or low likelihood of fraud.

The graph analysis mathematical apparatus identifies cycles of connectivity between participants, ROAD ACCIDENT connections with known fraudsters, and also calculates various business indicators of the environment in which a loss is included. For example, the presence of people surrounded by a client with a refusal insurance on suspicion of fraud or a large number of losses with appeals to the court. These indicators can be used both in machine learning models and in expert rules.

Together with Ingosstrakh, we aim to create an advanced anti-fraud system based on open-source technologies. Automation of anti-fraud processes will increase the accuracy of loss analysis, reduce the time for processing a large array of data on the relationships between objects and insurance entities and, in general, increase the loyalty of bona fide customers by reducing the level of false positives and reducing the settlement time for them, - commented Evgeny Chernoburov, head of insurance practice at GlowByte.

In the future, the implemented graph solution can be scaled not only for motor insurance, but also for other Ingosstrakh insurance products.