Hamilton, AIG add intelligence to underwriting with Attune
Automated underwriting driven by big data has been popular in personal lines auto for several years. But the technologies are now making insurance underwriting more efficient and driving a better customer experience in other business lines as well.
Hamilton USA and AIG developed a unit called Attune last year that uses advanced data and analytics to underwrite small commercial insurance. Joining Hamilton and AIG on Attune’s development is Two Sigma Investments, which lent expertise in data science to the project.
The companies say they collaborated on the effort because the population of business owners and sole proprietors is skewing toward digital natives who expect the same experience from their commercial insurers that they’ve received from auto carriers. Insurers need to update their processes to meet that demand.
“The small commercial insurance market is moving,” explains Ed Pulkstenis, EVP of underwriting at Hamilton USA. “Some studies say that more than 50% of small commercial business owners will be millennials within the next few years. We think that Attune is a platform that will meet them where they’re going.”
Attune works by leveraging third-party data and analytics models to generate a quote with limited information — sometimes only the business’s name and address. Agents are still a part of the service flow, but without having to ask as many questions to generate a quote, the experience for the end customer is smoother and not as onerous as it has been in the past.
“For small-commercial business in particular, there’s been this tension between effective underwriting and ease of doing business,” explains Ed Pulkstenis, EVP of underwriting at Hamilton USA. “The traditional view is that to do better underwriting, you have to ask more questions, validate more information, and do more back and forth. Attune says that is a false choice.”
Today, most of the small commercial business is written through intermediaries, and so Attune was created for use by retail agents. However, the companies plan to extend access and functionality to wholesalers, and potentially add online self-service capabilities down the road.
To succeed, however, Pulkstenis says Attune needed to be a stand-alone entity and approach the small commercial market from a fresh perspective. Attune operates as a separate operating unit from the companies, not as a technology platform within the traditional line of business. So, the partnering companies end up sharing the wealth from the policies instead of being in competition. That illustrates how digital innovation changes a lot of attitudes around the nature of the insurance business, Pulkstenis explains.
“The legacy pull can be really strong. It’s hard to break away from old biases, old philosophies and traditional ways of doing things,” he says. “It’s a new venture. By taking AIG’s capacity, and using that with Hamilton’s expertise on a new platform, we get the best of both worlds.”
For now, Attune is being used to write only business owner policies. Over the next six months, however, AIG and Hamilton will be transitioning business from legacy systems into Attune, Pulkstenis says. The plan is to roll out an umbrella or excess and surplus product during the fourth quarter of 2017, followed by workers compensation, auto, and potentially supplementary products.
“AIG has committed a material amount of business to the platform,” Pulkstenis says. “They see this as an opportunity to decommission old systems to reduce their costs, and to be a more viable participant in this new market that’s emerging for small commercial."
Dealing with data
Pulkstenis says that Attune comprises is a cocktail of proprietary, esoteric, traditional, and foundational technologies. While cagey about the specific data sources Attune considers, Pulkstenis does offer some insights into the thought process. A lot of traditional underwriting questions simply were not that valuable, he says, and success comes down to understanding a relatively few number of items.
“Companies almost always ask small-commercial customers how many years they’ve been in business, though what’s of interest is management quality. And there’s a lot of external information that allows us to assess management,” he explains. Rather than asking the customer a question, the system queries a number of databases that ultimately offer better information.
Much of that data, Pulkstenis says, is tangential, but insightful, and related to the company’s exposures and past claims: “We then take that data and design business intelligence around it to bring it back to the underwriting theme we want to answer.”
Though third-party data isn’t always perfect, Pulkstenis explains, the higher volume makes up for it.
“You may have 10 [fields to populate with] data, but three could come in blank, and two more are, for one reason or another, unreliable. But by having a wide variety of answers, you can triangulate toward the right information.”
While Attune is applying increasing amounts of automation to the underwriting process, Pulkstenis says, for now, some manual components remain.
“As we build a book of business, we are going to be able to find patterns and take advantage of machine learning and more sophisticated artificial intelligence,” Pulkstenis says. “Attune says we don’t have to choose between good underwriting and ease of doing business, or between having a legacy platform or great capacity. We don’t have to choose between contemporary technology expertise vs. traditional insurance expertise. The magic happens when we bring those things together.”