Automated underwriting is gaining traction. Learn how algorithms are increasingly responsible for setting rates.

The traditional method of underwriting insurance involves humans (aided by statistics) assessing an applicant’s risk and suitability to receive a policy. Factors like age, gender, driving record, and personal/family health history come into play to determine if the green light is given and at what rate.

Today, purely digital solutions are on the rise, ousting human analysis in favor of artificial intelligence (AI) and algorithms. 

The method is certainly here to stay. According to a recent survey, 62% of carriers are committed to automated underwriting, and that figure is climbing due, in part, to pandemic pressures. Let’s unpack the pros and cons of this insurance evolution and highlight how it may affect people’s chances of getting coverage.

Defining automated underwriting

It’s a technological solution based on robotic process automation, AI, and algorithmic patterns designed to significantly assist with, or wholly assume, the traditional responsibilities of the underwriter. This computerized scanning and screening of applications has been helping providers use less manpower and time to approve or deny potential customers.

There is still a human element in some automated underwriting outcomes. Between approved and denied lies “refer/eligible.” This means the software doesn’t have all the data it needs to make the final call, so the application then goes to a human for final assessment. A lack of universal adoption of the tech also means that a real person is likely giving digital decisions at least a cursory look before the last call is made.

Automated underwriting for property insurance (auto, homeowners, and renters) is thriving through companies like Lemonade and Root, two “insure-tech” players that define the rapid responses made capable by automation.

The pros of automated underwriting

A fast “yes” to an application is a big plus for those needing coverage in a hurry and a boon for providers who gain more paying customers quicker. This is because AI moves exponentially faster than agents, analyzing greater amounts of data in a fraction of the time.

Gathering the necessary information to submit an application is also easier on the customer, who traditionally may need to source medical histories, bank statements, driver’s records, and more to pass them on to providers. This means more steps in the process before an underwriter can even get started. Many automated platforms, however, can swiftly gather that data digitally by mining it from the proper sources.

Automated underwriting, though not perfect, is essentially free from human error and bias that may affect an applicant. Automated underwriting algorithms simply see what they see and make application decisions based on the cold, hard data. This can be a fair way of approving or denying people, but it’s not without caveats.

The cons of automated underwriting

It’s ultimately the insurance underwriter’s role — organic or automatic — to assume the risk for potential events which involve the provider paying out (often, substantially). The bigger the risk an applicant poses, the higher their premiums must be to compensate.

The big downside here is also a benefit: automated underwriting can see deeper and more accurately into an applicant’s life than human evaluators can. The result: applicants are increasingly unable to hide or gloss over any factors that could raise their rates or see them denied.

Of course, hiding anything in an application is wrong. But the emphasis here is on details beyond the information actively asked for and volunteered. For example, personal devices like smartphones and watches can constantly report on their owners, right down to their heart rate. And factors like credit history and eating habits are all just data for the taking nowadays, and insurance companies may use it.

This level of detail informs more accurate underwriting, but it also raises privacy and fairness concerns.

How once-hidden data and provider agendas could influence automatic underwriting

The NAIC has reported on the grey area of insurers having access to and using epigenetic data to assess policies. Epigenetic information is found “above” or attached to the genome and registers biological impacts caused by behavioral factors including tobacco, alcohol, and illegal drug use. This means that our very genetic makeup could be next on automatic underwriting’s list of criteria.

And whether we choose to disclose it or not, everything from grocery lists, life decisions, and even DNA is increasingly out there in the sea of big data that AI sifts to make underwriting decisions. All this data was once invisible to human underwriters but is now becoming accessible to automated agents.

In addition, automated underwriting could create less-than-equitable outcomes. AI may be free of human bias, but it can create different versions of it. And the employment of automation is only as moral as its users and creators.

The insurance industry is currently under scrutiny on these grounds because some providers have been leveraging big data and AI to the detriment of protected classes, skewing the underwriting process to favor some more than others. This is a facet of underwriting that bears watching and may spur greater regulation. 

NICRIS Insurance focuses on finding clients the appropriate suite of products to protect them while providing insight on developing insurance topics. If you need some policy guidance or would like a free, personalized insurance review, just drop us a line.