Getting it Right: Underwriting Automation of the Future

Of the many interview opportunities I had at ACORD LOMA, one of the most forward-looking was with Salvator Branca, SVP at AIG Casualty on the topic of underwriting automation.

Although he didn’t offer a timeline, Branca offered a vision of insurance products bundled with other physical products, say an insurance policy bundled with an auto or home purchase for example, and sold by companies not traditionally associated with the insurance business. Those companies would theoretically be using data from outside providers, such as Google, which is collecting massive amounts of data on users. It could even be Google, he mused.

“We are going to wake up someday and find that Google is going to do this. Why wouldn’t they?” Branca said. “They own the mailing list for the new generation; they don’t want a policy, they don’t even look at paper anymore. The kids want to see something there that basically says: I’m covered. And if they want to dig deeper, they can. But that’s a sea change for the insurance industry. We’re still talking about e-signature. I mean, come on! We’ve got to get past that. We need to prove to these kids that they’ve got coverage, that everything they want to know, we can answer. We know where they are, we know who they are; that’s what they expect.”

All of which raises the age-old-insurance question: Who owns the relationship? And the answer in the future could be the entity that owns that person’s data.

“If you have a relationship with Google and they know everything about you, you are going to gravitate toward them for everything – as long as they don’t screw up,” Branca said.

While automated underwriting and insurance bundling is a likely future for the industry, there are several attendant concerns, he said.

“This is the future. But I do worry that we are going to hit the wall. We need more subject matter experts. We are in this transition period and we need to hold on to the subject matter experts that really understand our business,” he said. And that understanding must be part of the underwriting rules and turned into a reliable scoring engine.

“We talk about artificial intelligence, now it’s machine learning, it’s basically taking what’s in an underwriters mind – and not just the perfunctory forms, but the intuition. ‘It’s a university, is it on a hill? Does it have wind? Does it flood?’ It needs to take all of that and put it into the machine,” he says.

Thoroughness and sensitivity to the customers’ needs continue to be critical.

“How do you tie it into that one location with specific questions that are geared to that location? So, demographic information without violating privacy, because we can never forget that we are in business to protect that insured. And this is something to worry about. Our main goal is to make sure that person has the right coverage,” Branca said.

“Everything we do has to be fair and it needs to be right,” Branca said. “What happens if we make a mistake and say, ‘we have your name and you’re a bad credit risk,’ when in fact you pay your bills on time all the time. You may not even know it, but when your name comes up on a cred model? You’re at the bottom, you’re out and no one is even thinking about you, that fast. That’s the kind of mentality we are going to have to recognize. We have to do it right, because we owe it to that person. And it should be the same whether it’s Sal’s Carpet Cleaning or a large national account. We have to make sure we are doing this correctly and we need to move more people to do this.”

Chris McMahon is a senior editor for Insurance Networking News.

Readers are encouraged to respond to Chris by using the “Add Your Comments” box below. Healso can be reached at chris.mcmahon@sourcemedia.com.

This blog was exclusively written for Insurance Networking News. It may not be reposted or reused without permission from Insurance Networking News.

For reprint and licensing requests for this article, click here.
Digital distribution
MORE FROM DIGITAL INSURANCE