AI’s effect on the insurance underwriting process
Element AI Element AI
August 2 5 min

AI’s effect on the insurance underwriting process

Artificial intelligence is increasingly used by all kinds of businesses to handle large-scale data processing, with increasing levels of complexity. As AI becomes more sophisticated, it’s able to handle more and more challenging tasks and becoming increasingly useful to businesses and employees.

The insurance industry is no exception, and the underwriting process is one area in particular where AI can really shine. Its capabilities for helping to process and judge information makes it extremely useful for underwriters. Lloyds of London, one of the world’s oldest and most prominent underwriting firms, is among businesses rolling out more and more AI in its operations.

In this article, we’ll explore AI in underwriting, but first, let’s briefly run through what underwriting is.

Underwriting in a nutshell

Underwriting is at the very heart of insurance. It means guaranteeing a payment will be made to the customer in case of a financial loss, so that they’re “covered” if misfortune strikes. It also means taking on the customer’s financial risk, in exchange for a “premium” they pay the insurer.

The business model of underwriting depends on the underwriter receiving more in incoming payments from these premiums than they pay out if and when customers make claims. When deciding whether to underwrite a customer’s risk (whether it’s insuring their home, car or a ship laden with cargo) and calculating what the premiums or potential pay-outs might be, there are several key factors an underwriter must consider:

  • What is the level of risk involved? Is there a high or low chance a claim will be made?
  • What sum should the premium be? This isn’t only decided by the risk involved, but also the need to stay competitive with rivals’ rates.
  • How trustworthy is the customer? Is there a chance they’ll make a fraudulent claim?

How AI can help

Once, these numbers had to be crunched, research conducted, and odds weighed manually. But now we live in an age of ever-expanding, evermore connected and accessible data. That’s where AI can make itself useful – as an underwriter’s assistant.

AI can make underwriting recommendations using digital financial information, credit records, and more. It can access these data sources and use them to help the underwriter make a decision.

That means underwriters will be able to more effectively analyse a customer’s application for insurance coverage and check for signs that indicate risk and financial viability. They can also analyse a claim and assess its validity – for instance, to see if it meets the criteria for a pay-out, or flag it if it’s likely to be fraudulent.

AI can analyse huge amounts of applications or claims in bulk, categorize them quickly, compare them using data analytics, and find patterns and anomalies. This gives underwriters a far greater level of insight, across all their customers, and helps them to make more informed and better decisions.

Underwriting meets the Internet of Things

As our world grows increasingly more connected, thanks in part to the Internet of Things (IoT), there is also the potential for digitally assisted underwriting to become much more nuanced.

This isn’t a far-off prospect. In fact, an in-practice example already exists in the insurance industry: the black box in automobile insurance. It gets its name from the original black box flight data recorders used by airlines to discover the circumstances behind crashes.

Black box auto insurance is a program that offers you premiums based on your driving behaviour. Your behaviour – and its level of risk – is calculated according to telemetric data from a device that’s usually installed in your vehicle. Data captured includes GPS information, vehicle speed, location and travel distance.

The safer you drive, the lower your premiums. The riskier your behaviour on the road, the higher they go. Although some drivers would consider the black box a somewhat Orwellian intrusion into their driving lives, others may see it as being a far fairer alternative than being judged based on age, gender, location and other demographic details by which premiums are sometimes calculated. And a black box certainly provides a more accurate picture of your current driving behaviour than simply relying on your past record.

The future of the underwriting process

Black box insurance is only the beginning. With an increasing number of devices becoming “smart”, from wearables to medical devices to your toaster, providing more and more data for more and more insights, and increasing sophistication of technologies like deep learning and data mining, it seems that AI and data analytics will only become more prevalent within underwriting and insurance as a whole.

Of course, with every benefit, there is a potentially negative side to anything that deals with AI and the collection of data. While the use of black boxes will benefit insurance companies, as well as those individuals who don’t stray too far outside the sphere of safe habits, this type of monitoring of individual personal behaviour can be used for shadier means if employed maliciously.

For a great example of the less-than-ehtical uses of data collected through personal black boxes, listen to our latest podcast episode (S2E2) that discusses a real-life instance regarding Fitbits distributed to employees of a particular establishment.

As ever, any game-changing new technology bears benefits if it’s adopted properly and ethically, and the risk of being outpaced by competitors if it’s neglected. When it comes to harnessing AI in their underwriting process, insurers should see it as a matter of being prepared for the future, optimizing their business in the present, and ensuring the proper and ethical use of the data being collected.