The new solution uses predictive models to forecast the behavior of insured persons on the basis of more than 70 factors, such as the client's industry, office locations, routes from offices to clinics, etc. It takes the technology just a few minutes to assess medical care consumption and claim amounts. AI will help insurance companies to increase risk assessment accuracy 1.5 to 2 times and therefore reduce health insurance claims, which amount to RUB 40-60 million annually for a large insurer. Moreover, the technology will benefit corporate clients as well, as the underwriting accuracy provides a better and more objective insurance rates. Thanks to this AI solution a significant share of large customers is going to be able to save about 10% of their insurance spending.

Development of the AI-based model took more than a year. It began with historical data preparation, structuring and augmentation. This was followed by training of predictive models, search for and adaptation of significant factors. At the time being technology testing process is over, and the system has been commissioned for commercial operation.

Worldwide there is also a number of cases using big data in underwriting which had a success. For instance, U.S. health insurer Collective Health, according to 2018 data, made a dent in American companies' $1.2 trillion annual healthcare spend. The Russian market experience such a large-scale solution for the first time. Another advantage of the technology is its adaptability/ A trained model can analyze data not only for federal cities, but also for regions, where expert knowledge of city specifics is insufficient.

"While working on this technological solution, we looked at the needs of both the insurance market and our clients. AI will help to boost underwriting accuracy and forecasting speed, making it possible to detect clients with lower or higher risks. Those companies that make good use of preventive measures for preserving their employees' health and that have a balanced health insurance program are to be offered lower rates. This is especially relevant in the tough times of pandemic, when companies have to spend more on the health of their employees. Integration with external data, tracking of macro parameters, and regular additional learning make this tool "agile" and allow it to consider constantly changing environment"

Mains Lab CEO

publication date:

August, 10, 2021

Tags:

AI-underwriting

Machine learning

Health insurance

Share: