As more insurers experiment with artificial intelligence (AI) to automate repetitive processes and garner more accurate underwriting insights, experts speaking with Insurance Insider agreed that AI will accelerate an ongoing trend towards hyper specialization and will differentiate specialty players.
“You can become a preeminent underwriter without necessarily having all the legacy because there's many classes of business where new forms of risk data can give you blindingly good insight,” said Skyward Specialty CEO Andrew Robinson, one of four panelists on a recent Insurance Insider webinar on AI’s impacts on the specialty landscape.
“If you take management liability, which in the public D&O space is an incredibly soft market right now, the ability to distill disclosure information as a pretext for class action is immensely powerful.”
Robinson also spoke about Skyward’s application of AI: a partnership with Gradient AI to build a computational engine that allows better risk identification and selection by creating a narrative for each submission and classifying submissions based on risk.
“Whether it’s just being able to get a quick narrative on litigation when you're writing professional liability or getting a quick summarization from the medical boards on [physician] history, [which done] manually can take quite a bit of time, my view is this is not a cost takeout; it’s a growth accelerant.”
Robinson also referenced the application of large language models to sort through vast amounts of unstructured information like claims notes and to identify the claims that have the highest potential for reserve development.
Martha Notares, managing partner with fintech and InsurTech venture capital firm Brewer Lane Ventures, underscored that AI could also allow carriers to take advantage of micro-cycles within various lines.
“If you're using AI to leverage more data to understand the risk better, you are willing to lean in to risks that other people don't even want to touch,” Notares said.
“At the same time, you have an appetite for that risk because you understand it and therefore that cyclicality changes to be much more personalized to the risk,” she continued. “We see this already to some extent where people are able to place in a hard market because they explain the risk better or they have a long-term relationship.”
Robinson also predicted that the push for AI adoption will birth a new class of highly specialized vendors focused on the specialty insurance industry.
“With a lot of these tools so far, it's pretty easy to put together a slick demo, [but] it's pretty hard to put together something that tangibly impacts processes and is 100% correct across a range of use cases,” he said.
“I think you're going to see a wave of vendors that are highly focused on things like forms comparison or risk classification, or others that will be used by some incumbents as plugins alongside their own internal efforts around AI development.”
While the executives were overall optimistic about AI’s potential for the industry, Notares emphasized the importance of regulatory standards in navigating the biases inherent within AI.
Robinson pushed back on this, noting: “I have yet to see an example where the regulators don't go through a lot of pain before getting it right. It was 17 years in Florida before they got it right on homeowners’.”
Ben Sloop, COO of Amwins, characterized AI in insurance as an “evolution not a revolution”, highlighting progress made in automating repetitive tasks.
“From an insurance company standpoint, the firms that can run with the lowest cost base tend to have a sustainable competitive advantage. And to me, that feels like the long-term impact [of AI],” said Sloop.
“Are there some things about it that are ... faddish? Absolutely, but I think there is real substance over a five- to 10-year time horizon for incumbents that utilize it effectively.”