The COVID-19 pandemic has accelerated online purchasing across the board, including in insurance. Swiss Re confirms this trend in its most recent economic bulletin.
Insurers should accelerate their investment in machine intelligence in order to take advantage of this shift, Rajeev Sharan, senior economist at the reinsurer, explains.
The increase in online insurance purchasing has been particularly marked in Asia, Sharan continues. A Swiss Re survey shows that the rise in online purchasing is generating large volumes of new digital data about consumer behaviour and lifestyles.
“Insurers could potentially process this data to better service customer needs and underwriting excellence,” Sharan says. He mentions that the more diverse datasets now available online can enrich model training to improve exception handling capabilities. “This is where the exclusion terms of an insurance contract can be adapted to align with a policyholder's changing life circumstances,” he adds.
Huge volumes of consumer data
In the case of life insurance, MI models can facilitate continuous underwriting rather than the traditional at-time-of-sale approach, Sharan says. “The latter is based [on] past morbidity and behavioural data, and does not capture current-day changes in customer life circumstances. The huge volumes of consumer data available online are a window to such changes.”
Life insurers can access this data and process it using algorithms to provide ongoing underwriting services, whose policy terms and conditions are dynamically adjusted based on customer behaviour patterns. Suggested personalized actions can drive healthy behaviour, Swiss Re says.
John Hancock, a U.S. subsidiary of Manulife, has been using this strategy since 2018, Sharan notes. The insurer has added fitness monitoring to all policies. These “interactive” policies collect health data from wearable devices, and policyholders can earn discounts and rewards for meeting exercise goals.
Strengthening global resilience
The spread of digital technologies is also an opportunity to strengthen global resilience to natural catastrophe risks, Swiss Re continues.
“Wider acceptance of standardized data-formats for claims and exposures, and also for opensource information exchange can help improve predictive analysis and risk modelling capabilities,” Sharan says. He gives the example of the American company Terra Seismic, which has developed open source algorithms that use live broadcast data from satellite images and atmospheric sensors. These algorithms can potentially predict earthquakes anywhere in the world, Terra Seismic maintains.
Machine intelligence systems can also be deployed to reduce carbon emissions. For example, says Sharan, by using sensing devices and algorithms to gather and analyze data, MI can forecast the supply and demand of power, improve the usage and scheduling of renewable energy supplies according to need, and run predictive maintenance programs.
“This would support transition a low-carbon economy. As per the Global Commission on Adaptation estimates, every dollar invested in building climate resilience could result in USD 2-10 in net economic benefits,” the senior economist says.
