How commercial insurers and underwriters are using AI
Insurers and underwriters are using artificial intelligence (AI) for more effective pricing and risk selection in the commercial insurance space, attendees at Insurance Bureau of Canada’s (IBC) Commercial Insurance Symposium heard last week.
In particular, insurance professionals are using AI to deep-dive into telematics, especially in the marine, aviation, and commercial auto space, says Shaughn McClusky, global financial services and insurance industry specialist with Amazon Web Services (AWS). They’re also using AI to probe sensor data for more personalized underwriting.
“Telematics information and the data access [can] enable you to be more effective in your pricing and risk selection by understanding more granularly what the associated risk is for that commercial policyholder,” McClusky says during the symposium, which was part of IBC’s InSight Summit.
“[In] the sensory space right now, you can really start to predict and project, based on sensor information, [which] people in these organizations that you’re looking to insure or underwrite effectively…are the good ones — in terms of, they have more proactive and more predictable maintenance activities to ensure that their businesses, from an ongoing entity perspective — [and] are doing the right things versus those who aren’t,” McClusky adds.
Data gleaned from this sensor information can help differentiate pricing.
“So, we’re seeing customers really starting to use data and price more personalized…[from] a usage-based perspective,” McClusky says. “You can say, ‘In the commercial auto space, if somebody’s not driving their truck, should they really be paying the same premium for not driving that truck when it’s on the road as much as somebody who drives it every day 24/7, like an Uber?’
“So certainly, customers are starting to investigate opportunities for that.”
Related: Meet your new underwriting assistant: GenAI
Getting into better utilization of data and automation in the underwriting space “is really starting to be top of mind for the industry,” McClusky says. He points to a statistic from Accenture that automating underwriting processes could allow for efficiency gains of up to US$160 billion by 2027, as quoted by Munich Re.
“I myself have had conversations — most recently in the last couple weeks — with three insurers in the Canadian space, two insurers in the U.S. space, talking about, ‘How do I improve my submission intake?’” McClusky says.
For example, an insurer may have a stack of emails with 100 requests for quotes. “From those 100, what are the ones that are aligned to my risk appetite? What are the ones that I have a likelihood to win? What are the ones I really want to spend my time on?” McClusky asks.
Using AI, insurers can transcribe risk exposure from an email, consume Excel files of inventory lists, and consume brokers’ disclaimer statements and submission information. “And how do I simplify the understanding of that so I can be more proficient and respond to my quotes faster before my competitor does?” McClusky says.
“We’re really seeing a lot of interest and submission intake triage through intelligent document processing, whether it be an actual document or whether it be an email,” he says. “Basically, shuffling that big stack of 100 and getting the top 35 — ones that you know you’re going to win.”
AWS customers are even using AI to “look at their workflows — little pockets of use cases where they can drive some operational efficiencies,” McClusky says. “So, looking at workflow opportunities where they can bring automation in place and allow their skilled resources to do more complex heavy-lifting capabilities versus the ‘go find this information somewhere, consolidate it, summarize it and then start to use it.’”
