Insurers who replace underwriters with computers are rejecting a lot of small business applications for reasons that are unclear to the brokers, a commercial broker suggests.
“I think that insurers will start to realize that the black boxes are actually kicking out a lot of business that maybe does have a few issues, but an underwriter in the olden days used to just underwrite to those issues,” said Michael Loeters, senior vice president of commercial Insurance & Risk Management at Prolink, in a recent interview.
Loeters was referring to some insurers who are using a computer program to vet applications and the computer decides what is rejected.
When a human underwriter receives an application with some high risks, that underwriter often deals with it not by rejecting it outright, but by amending terms, adjusting the deductibles or putting some limitations on coverage, said Loeters.
“I am not sure the early machine learning and artificial intelligence technologies that are being implemented by insurers are going to have the capabilities to do that, so they are going to kick out a lot of good business,” said Loeters.
“What we are dealing with as brokers is that insurers are using technology to try and simplify and streamline the qualification of submissions, and we are getting declines that are not based on a person looking at the submission. The submission went through a black box and the black box decided, based on whatever data it is using, that it wasn’t a good risk. As brokers, we have never had to deal with that before.”
AI is where computers mimic human cognition and activities – including pattern recognition and learning from experience, reports Celent. Machine learning allows a computer to understanding something from experience, so it adapts its behaviour as it learns more data.
“I totally get and understand why insurers are doing that – obviously it has a huge impact on costs for them,” said Loeters. “That’s going to be an interesting dynamics, because now you are going to have insurers fighting over a much smaller piece of the pie and I think those insurers who do invest in the underwriting expertise are going to have a windfall of opportunity because they are going to take these and go, ‘yes, I can write this risk but I will need to make some adjustments to the terms, the coverage, the pricing et cetera to make it happen.’”
Recently Loeters submitted an application for cyber coverage to an insurer – who he did not identify – which was declined.
“The decline was somewhat ambiguous, so I called the underwriter to try to understand why they declined it, because from a cyber risk perspective, I thought this insured was probably as good as it gets, and she admitted to me that she never saw the submission,” said Loeters.
That insurer puts the application into a third-party software platform that it uses to give it a report.
“If the report has a score of less than X, than it’s an automatic decline,” said Loeters. “And this report said they had poor governance and poor controls, and I said ‘Well, how can you say they have poor governance and poor controls?’ and she admitted to me that she never even looked at the submission.”
In that instance, the application from Loeters’ client outlined all of its governance and controls.
“Every question on the application was answered correctly, they are audited by three different organizations for network security and privacy, and they are certified three different ways from a network security and privacy perspective,” he recounted.
But the underwriter told him the software gave the insurer “different information, so it was an automatic decline and she was unwilling to open the file.”
Only a small number of insurers are underwriting this way, “but it’s where all insurers want to go,” said Loeters.
“They are all currently investing in systems like that, or once they have their back end systems up to snuff it will be one of the very first things they do,” he added.