Data-Driven Innovation in Life and Health Insurance

We live in an increasingly interconnected world.

Everything from shoes to toothbrushes and more is connected to the newest bendy screen smartphone and to the cloud. Combine this data with information about spending habits gleaned from banking, credit card, and mobile wallet transactions, and it’s easy to see how pictures of an individual’s wealth and health can be generated.  

As a result, every company worldwide is highly interested in understanding how measurable behavior can impact core business operations and future strategic directions. Most of these efforts are focused on loyalty, customer engagement and streamlining the customer purchase journey.  

Life and health (re)insurance companies are no different. Companies have spent the past several years organizing and digitizing existing data assets, making it easier for advanced analytics teams to create management information dashboards and statistical models.  

In Asia, life and health insurance company data modeling efforts have mainly been focused on two goals: streamlining underwriting processes to speed policy issuance as well as simplify consumer needs analysis and cross-sell, and improve non-disclosure/fraud detection to maintain premium rates and protect the bottom line. Given the massive underinsured populations in many Asian countries, as well as the fraudulent activity that occurs in several markets across the region, these efforts are strategically wise.  

Efforts to date to achieve these goals have been admirable. At this point, however, our industry is merely scratching the surface of what potentially can be achieved by using existing data sources, gaining access to new permissioned data, and leveraging advanced statistical techniques such as machine learning and deep learning.  
Let’s delve a bit further into the data sources that life and health (re)insurers are interested in as well as the ways that this data can be utilized in the operations of life and health risk carriers. 

Novel Data Sources 

Existing life and health insurer data assets are still extremely useful, both in terms of understanding customer preferences and in pricing and managing risk. The competitive hunger to acquire more and richer data, however, is steady and increasing. Data-related partnerships are being developed by insurers with telecommunications companies, wearable device companies, and social networking companies, business partnerships involving data-matching and sharing are being made on a continuous basis, and this is all expected to accelerate. 

• Agency customer data: Data gathered from existing distribution channels can be further enhanced if the correct incentives could be given to sales agents to collect more comprehensive customer financial needs analysis data as a part of the new business acquisition process. These data can improve our understanding of customers and at the same time assist regulator goals of ensuring products sold are aligned with customer needs and circumstances.
• Bancassurance customer data: Most insurers in Asia have long-standing distribution relationships with banks. Banks have long viewed the insurance underwriting process as unnecessarily cumbersome and a barrier to greater sales. With appropriate customer permissions, bank-specific customer data could also become available for utilization.
• Wellness program data: These programs offer opportunities for insurers to gather novel data by enabling more frequent engagement with their policyholders. Built-in incentives reward customers for behaviors deemed healthy, such as exercising, getting medical checkups, and watching one’s food intake. These programs can be win-wins, as they allow insurers to collect more detailed permissioned customer data from biometric health checkups, wearable tracking devices, or other activities. Such programs need to be engaging so that customers would want to share usable personal data in exchange for valuable monetary incentives and health insights.   
• Credit data: For many years, a multitude of non-biometric data sources have been shown to be correlated to mortality and other insurance-related risks. Twenty years ago, for example, credit behavior emerged as a metric that could simplify buying a car, insuring a house, or acquiring a credit card. More recently, insurers in the U.S. and India have begun utilizing custom-built scores based on credit behavior in order to streamline underwriting.
• Driving data: Driving behavior has also been shown as correlative with both accident and sickness mortality risk. Indeed, individuals with infractions on their motor vehicle reports experience higher all-cause mortality than those with clean driving records. This data is currently being used in the U.S. to refine life and health insurance pricing.

Applications 

These new data sets have great potential to:

• Streamline operational processes such as underwriting and claims adjudication
• Increase customer engagement and loyalty by creating incentives for lower-risk health behaviors
• Reduce anti-selection, fraud, and non-disclosure by comparing applicant and customer disclosures against electronic health records
• Allow insurers to assess and price risk more accurately by finding new premium rating factors

While we have already seen insurers using Artificial Intelligence (AI) in claims processes, there’s a high probability it could also become part of underwriting processes as well. In that event, data from an ever-increasing number of sources would be used in conjunction with a variety of machine learning and deep learning techniques to arrive at underwriting decisions that are currently made by automated rules based underwriting engines and in some cases human underwriters.  

Integration of permissioned bank data into the underwriting process, for example, could allow insurers to cut down significantly on medical and financial information requirements, enabling simpler and more customized application experiences. Similarly, product recommendations could be more precisely tailored, based on data outlining customer financial needs, creating value for all parties involved.   

Permissioned electronic health records data, might also allow insurers to complete the underwriting process with limited applicant disclosure and input. Insurers would have more ability to check disclosures against third-party data, which would help avoid anti-selection, non-disclosure, and fraud.  

Conclusion 

While the future for new data acquisition and utilization looks bright, insurers and their partners must ensure they work within the confines of the latest data privacy and discrimination regulations to ensure consumer interests are respected. At the same time, regulators would want to ensure that data and information privacy regulations permit innovation.  

Opportunities are real and growing for life and health insurers to utilize data and advanced analytics to streamline operational processes in ways that both improve the customer experience and the bottom line. Many insurers today are reaping the benefits of such work while preparing the way for future partnerships and utilization of new data sources. The power of insurers’ existing data assets will be intensified when mortality and morbidity risk can be linked and combined with new sources of data pertinent to the risks presented.