Newsletter Sign-up

Starting in September, the Armed Forces Insurance Exchange, or AFI, an insurer licensed to write allied coverages, auto liability, fire, inland marine and general liability, began marching in a new underwriting direction.

At that time, Leavenworth, Kan.-based AFI, which has provided coverage to military professionals throughout the United States and overseas since 1887, launched a new automated underwriting system.

"The technology identifies the policies that require underwriter attention, so the majority of applications and renewals will be processed automatically," Dutton says. "It's quick, powerful, and we expect it will have a major impact on future losses."

Dutton and her team are going to be keeping their fingers crossed. It won't be their first shot at improving their company's underwriting capacity. AFI had previously tried using a mainframe-based system, but eventually abandoned it because it was too cumbersome.

This time, Dutton explains, things will be different.

The system, known by the acronym SURE, applies company-specific underwriting rules consistently and uniformly against incoming policy information. Supplied by Insurance Data Processing Inc., in Wyncote, Pa., the system allows AFI underwriters to focus on more strategic activities, the kind that require human intervention.

Dutton says that when an underwriter inputs a rule using SURE, it is then matched to the proper risk characterization. For example, if an application for a homeowner's policy indicates a swimming pool on the premises, the software application will automatically and directly tie that fact to underwriting. Best of all, Dutton notes, you don't need a programmer to add or change underwriting rules in the SURE system--it can all be done by the underwriters themselves.

When it comes to personal and commercial property/casualty underwriting, rules-based underwriting technology is starting to have an effect on an insurer's bottom line, proponents of the technology say.

Apart from new, rules-based systems offered by vendors like IDP, software applications that can truly automate the underwriting process are getting some real traction, say the experts.

Dax Craig, CEO of Valen Technologies, a Denver-based provider of predictive modeling technology to the insurance industry, sees three major underwriting technology trends in the property/casualty industry.

Carriers in commercial and personal lines are looking more carefully at external data. Craig says carriers used to look at standard underwriting data such as loss history or business characteristics. But they tended to ignore external factors, which are the basis for the emerging predictive modeling technology. Adding new factors, it turns out, makes the applications much more accurate.

Next, Craig says that carriers, mainly in the middle market, much like AFI, are beginning to apply business rules to the underwriting process in some form of technology-driven rules-engine method.

The third major technology trend, Craig says, is that carriers are using predictive modeling more often. Carriers work with a rules engine and external data that previously had not been part of the underwriting data mix. "Underwriters can use it to more accurately assess the risk of applicants by identifying undiscovered patterns in their application data and prior experience," he says.

CHEAPER COMPUTING POWER

Craig says the latest underwriting technology trends primarily are fueled by the boom in raw computing power. With processing speeds on the increase and prices decreasing, real-time decision tasks have become cheap. Second, he adds, up until now, there just haven't been truly automated systems out there.

"The actuarial models were pretty manual, used for portfolio analysis," he says. "But when you have a real-time engine decision, that makes it easier. And that type of technology hasn't been available until recently."

Craig compares the technology underwriting scenario of today to the 1980s, with the emergence of enterprise resource planning software from giants like SAP and Oracle. Back then, if you needed a manufacturing system, you went to a consultant, and they built one. But soon, vendors like SAP came along and automated that process off the shelf. "The same thing is happening now for underwriting," he says.

Gail McGiffin, a partner in Accenture's Financial Services Solutions Group, basically lists the same three major technology trends as Valen's Craig--data aggregation, predictive modeling and workflow automation--adding that all should result in some substantial strides for insurers.

McGiffin also says there is a growing appreciation for the need to have high-quality data earlier in the underwriting process, data not confined to information received from agents on the application form.

"Carriers need to leverage other sources beyond basic policy application data," McGiffin says. "That's one good source, but improved and advanced underwriting technology needs to infuse other data sources into the process. Data aggregation will improve decision making."

McGiffin says the changes in underwriting technology are due to a combination of factors, including improvements in the extraction, preparation and aggregation of data. Advances in mapping technologies are also having an impact.

"It's not just looking on the postevent data for catastrophe management, but also using data earlier in the decisioning process," she says. "There is a growing comfort level with the use of external data for the purposes of underwriting, which historically hasn't been the case."

It used to be that carriers could only look at dozens of variables. Now they are able to look at hundreds, even thousands. "That way, you have greater granularity in rules and predictive models over time," she says.

AFI's Dutton, for one, is convinced that the company's new underwriting system can accomplish that goal and more. The firm has a large book of property coverage along coast lines, so if a major hurricane is storming into the Gulf of Mexico, SURE will allow AFI's underwriters to quickly insert a rule that puts a moratorium on a specific county within the hurricane's potential path.

The system is customizable to meet AFI's underwriting rules, Dutton notes. She adds that the technology will process information from AFI's policy administration system and generate a report that will tell AFI which policies require action, such as a motor-vehicle report, an increase in limits, a referral to an underwriter or a cancellation.

"An expert underwriting system gives the underwriter more time to spend with the risk," she says. "Using a tool like SURE, the entire underwriting process is geared to help the user create his or her own rules, risk characteristics and reason codes. When they sit down to input data into the system, there is so much clarity there."

For midsize carriers such as AFI, this new underwriting technology also is a way to level the playing field with larger carriers, which have used predictive modeling but due to their size and dependence on mainframe systems, have a much tougher time adopting new technologies.

"We have to compete with the State Farms of the world," Dutton says. "If you watch the insurance industry every year, it gets tougher and tougher. That makes it more important to be able to use technology to our advantage, and offer a competitive rate."

Dutton says that because many larger carriers have homegrown underwriting programs that sit on a mainframe platform, moving to a system like SURE is a major challenge for them.

"We believe this new technology will give us an advantage," she says. "For us, being a small organization, it's important to be nimble."

In the words of Phil Sweetman, member of PA Consulting's Management Group, the technology appears to be for real. "I've noticed that the major software suppliers have something about predictive modeling software on their Web sites," he says. "As short as six months ago, it was more vaporware than reality."

TOM STARNER, a Philadelphia-based writer, contributes frequently on technology issues.

September 15, 2005

Copyright 2005© LRP Publications