How insurers will use AI to provide cover

And increasingly, market segmentation will reach the level of individuals, with a richer understanding of each person in the risk pool. 

Pros and cons

Mr McKenna says although big data and AI have huge potential to transform the way protection products are underwritten, he warns that this can bring both positive and negative consequences. 

He adds: “These techniques will, in time, enable insurers to predict far more accurately which individual customers have a greater propensity to illness and disease. This actually threatens the whole concept of pooled risk on which insurance is based.” 

Even if insurers agree to not use some of this data, people will be able to understand far more accurately their need for protection products. 

If only consumers who have potentially poor health seek cover, premiums could in the long run become unaffordable. 

This is an enormous commercial and moral issue, which insurers and reinsurers are very aware of. 

According to Jon Dean, head of retirement strategy at Altus, AI and machine learning will be used to forecast trends; tracking problems in workflow back to root cause and improve process efficiency – moving AI from a customer focus to an operations focus.

Mr Dean says: “Combined with big data, this will increasingly be used to track and forecast trends. Firms are just starting to do this with vulnerability, identifying potential for customer harm based on large samples, and also for claims analytics, especially fraud detection, and underwriting.

“IBM Watson can already heavily automate the underwriting of risk with potential to minimise human interventions.

“This was already changing with pioneer firms but will accelerate with incremental capabilities of leading cloud services.”

Relationship with providers

Advisers could also see some changes in the way providers interact with them.

The use of AI and big data has already enabled insurers to streamline some application processes. 

However, such changes are easier to implement for direct applications, says Mr McKenna.

“In the advised market you would need to see the development of a new type of portal service which would interact directly with each individual insurer’s underwriting engines. 

“Knowledge being gained by insurers about how to use AI is considered very valuable intellectual property, so I would not expect to see them willing to share this with third-party underwriting engines that make each insurer use a version of their own software.”

In the longer term, AI has enormous potential for mass personalisation of life cover, although Mr McKenna says this could make it far harder to support adviser comparison services as it could present a challenge where each company is providing dozens of different options to clients.