Predictive Mortality Model Qorus Innovation in Insurance Awards 2019 - Winner

FNB Life felt there was a need for disruption in the underwriting space of long-term insurance. Underwriting is a process that has been around for many decades and those same processes are around today. The normal process is to ask customers numerous questions and then to validate these questions with medicals. Customer have to get time with nurses, find a convenient time and place to be able to do medicals and not to mention how invasive the tests are. This needed to change and FNB Life embarked on the journey to look for an innovative technique to improve the underwriting process. This brought about the development of the decisioning engine which is a predictive mortality model. It uses big data and machine learning techniques in order to make proactive decisions about underwriting for customers. The model uses customer’s demographics, transactional data and home affairs (Government data) death data to predict the chances of death. The model looks at around 2 500 different factors, for example medical spend, fast food spends, and alcohol spend to mention a few. This predictive mortality model is used on our multiple channels such as FNB App, Online Banking and Contact Centre. This makes the sales process massively easy and efficient. We can now offer customers Pre-approved and Pre-qualified offers. Pre-approved allows the customer take up to R1.5 million covers with no underwriting questions or medical tests. Pre-qualified allows a customer to take up to R3 million cover with four underwriting questions and no medical tests.