AI is giving insurers godlike powers, says Sompo chief

Artificial intelligence and cutting-edge data analysis software mean that underwriters can now make predictions about the weather, natural disasters and senile dementia that previously “only God knew about”, the president of one of Japan’s biggest insurance companies has claimed.

The bold assertion by Mikio Okumura, head of Sompo Holdings, comes as the company prepares to roll out Japan’s first dementia prevention insurance package — a product designed for the world’s oldest society and based on analysis of the heartbeats, appetite and sleeping patterns of thousands of nursing home residents.

The move by Sompo marks the latest insurance industry escalation of a battle to secure an advantage through tech. Okumura said this was an area of competition that would decide the survivability of individual companies as they moved away from their conventional business areas.

The “god” claim follows Sompo’s $500mn investment two years ago in Palantir, the US specialist in big data analysis, and its taking of a 22 per cent stake in a Japanese AI start-up, Abeja.

“We can now reveal things that, in the past, only God knew about, thanks to technology including AI,” Okumura said, outlining a plan for an insurance scheme that not only pays out when symptoms of dementia occur, but also attempts to delay the onset of the disease by incentivising customers to change their daily behaviour. That could include improving sleep, diet and exercise.

Sompo aims to make use of 500 sets of data obtained from 80,000 residents of its nursing houses, a business the company entered in 2015, to create preventive insurance catering to each individual’s lifestyle. In 2020, 6.3mn people were estimated to be dementia patients in Japan, and the figure is likely to reach 11mn by 2060, which will be equivalent to one in almost three of the elderly population, according to the Cabinet Office.

Palantir’s technology has so far enabled the insurer to analyse correlations of data and create a model to suggest improvements to care plans for each resident. Okumura hoped that these “big data” skills could now serve its core insurance business, helping the company to develop an insurance package that “avoids risks”.

“We will encourage our customers to change behaviour and such a solution could be attached to the medical insurance of people who are still at their working age,” Okumura said. “If they manage to delay symptoms of dementia [for two or three years], we can offer them cheaper insurance.”

“This will reduce the burden on the families who support them, extend the healthy life expectancy of the customer and reduce the insurer’s insurance payments . . . As a result, the national social security system will become more sustainable,” he added.

Sompo, in common with other insurers, is also eager to use ever more powerful tools to update the insurance business model for a world increasingly battered by the effects of climate change. Insurers around the world have been forced to confront a steep rise in natural catastrophe payouts, with events ranging from Europe’s worst drought in 500 years this summer to devastating floods in Pakistan.

In the wake of Hurricane Ian’s destructive rampage across Florida and South Carolina in September and October, risk-modelling company Verisk said in its initial reports that insurance companies were braced for a hit of up to $57bn as a result of estimated wind, storm surge and inland flood losses.

Combining historical data on damages caused by typhoons or floods with predictions of future climate change, Okumura believes Sompo can, for example, estimate the degree of natural disasters and economic losses, and give suggestions to customers to change the structure of buildings to mitigate damage.