When Mike Koscielny joined AAA of Michigan four years ago, the Dearborn-based auto club was entering expansion mode. It was acquiring AAA clubs in Illinois, Minnesota, Iowa and Wisconsin-and planned to grow in other nearby states.As a result, senior managers were considering how to accommodate the increase in business, says Koscielny, director of regional underwriting operations. Joking over lunch at Brainstorm's Business Process Management Conference in Chicago in April, he tells Insurance Networking News: "We thought we might have to add on to our building just to house the extra employees we'd need to meet our growth targets."

Today, the Auto Club Group (AAA of Michigan) operates in nine states - now including Nebraska, North Dakota, Indiana and Ohio - and, the organization has not had to construct new office space for more employees.

In fact, the Auto Club moved into these additional states and processed 35% more applications with three times as many independent agents, while hiring only six more underwriters.

The increase in business without an equivalent increase in staff was accomplished primarily because of technology, says Koscielny. Specifically, the insurer implemented a business rules engine called Blaze Advisor, from Minneapolis-based Fair Isaac Corp., which automatically issues 99% of its automobile and 92% of its homeowners policies.

"We've gone from manually reviewing 100% of (auto) applications that come through the door to having the system issue 99% of them," says Koscielny. Although 20% of new business applications are reviewed by underwriters after the system issues quotes, that's only to verify specific information, he notes. A full 99% of auto insurance applications are issued by the system within a day.

For homeowners insurance, the results are similarly impressive. Approximately 8% of the Auto Club Group's homeowners applications are sent to underwriters for review-and that higher number is due to the fact that the company requires an underwriter to review applications for high-value homes, says Koscielny. "It has nothing to do with the technology-or regulatory requirements. It's just our policy."


The initial thrust of the Auto Club's automated underwriting initiative actually focused on building a user-friendly and efficient Web-based sales and service portal for its independent agents.

"We were operating with three (policy administration) legacy systems and no front end," says Koscielny. "It was all green-screen technology, and we had both employee agents and independent agents writing business with us. But, we had no real easy way to get that done."

Those agents were processing business directly in the insurer's systems, with no edits or checks, Koscielny adds, which meant underwriters had to review every application that came in-just to make sure all the information was correct and the policies were rated properly.

Knowing their market share was expanding, management quickly realized the need to standardize and automate the organizations' disparate and antiquated underwriting processes. "It was imperative that we get rules (technology) in place so we could reduce the number of applications underwriters had to look at," says Koscielny.

Working with Bermuda-based Accenture, the insurer built a standardized Web-based, front-end sales and service portal, while also investigating rules technology vendors with the help of Accenture and Gartner Inc., Stamford, Conn.

"We immediately gave up on the idea of building (the rules engine) ourselves," Koscielny says. "We wanted much more flexibility, and we wanted a technology that would allow us to write rules not just for auto and homeowners insurance, but also for other areas of the enterprise."

As a result, claims and "the club" areas of the business were also involved in the vendor selection process. Claims was interested in using the technology for fraud detection, and the club-which provides membership benefits, such as towing and travel services-saw potential to automate its processes as well.


Meanwhile, Koscielny and his team were defining the underwriting rules and documenting them in an Excel spreadsheet repository. This is a "living document," he explains, which the business owns and maintains and is written in a language both business and technology staff understand.

"We made a couple decisions very early on," says Koscielny. "Defining what technology we wanted and what our requirements were." One of those requirements was flexibility in the rules management process, he says.

"The insurance industry is heavily regulated, and typically, in any company, getting IT resources to make changes is difficult. We didn't want to get hung up in the whole stream of (IT) priorities when rules need to be changed."

The Auto Club narrowed the field of vendors down from eight to two, and, finally, to Fair Isaac in December 2002.

In March, the insurer introduced its first rules-based functionality: auto quote in Illinois. A month later, it rolled out auto new business in Illinois. After that, it deployed another piece of functionality every 60 to 90 days, including homeowners quote and new business, and auto policy change and renewal.

"As we moved into new states, we'd use the automated system," says Koscielny. Currently, the Auto Club has 1,300 rules in the system, which processes business in Illinois, Indiana, Minnesota, Wisconsin, Iowa and Ohio.

The company will add Michigan this year and North Dakota and Nebraska in 2007, while it also standardizes on a Huon policy administration system from U.K.-based Innovation Group.

Auto policy renewal, which the Auto Club introduced in 2004, was its first venture in using the system for a batch process, which was another required feature when the insurer was selecting a rules technology vendor, says Koscielny.

"We needed a tool that could handle large volumes of business on a nightly batch basis-as well as real-time business," he says. "Claims runs all their fraud detection through the system at night, and we (underwriting) needed it for the renewal process."

Automated policy renewal involves running tens of thousands of policies through the rules engine overnight. "That was a challenge for us," says Koscielny, who emphasizes the importance of comprehensive regression testing when adding new functionality to a rules engine. Because automated renewal places such a heavy load on the system, the Auto Club spent most of 2004 just enhancing that process.

"Before we even flipped the switch (for automated policy renewal), we ran the processing (through regression testing) to make sure what came out the other end was what we expected it to be," he says. "Everything we've built is built on top of everything else-and systems can break. The last place you want to do regression testing is in production."

Automating the underwriting process to enable cost-efficient expansion into new markets was the primary reason for the Auto Club Group to implement a rules engine, but that's not the only benefit of using the technology.


Recently, the insurer introduced a loss-free discount to members who have had no claims, and a premium-increase waiver when members file their first claim. These are competitive offerings the insurer was unable to provide before.

"In the past, when we've talked about introducing (an accident waiver), the technology folks told us it would take thousands of hours of programming and cost hundreds of thousands of dollars," says Koscielny. "With Blaze Advisor, we had it up and running in 60 days."

In fact, the Auto Club's actuaries are now able to develop far more sophisticated pricing models, while its underwriters have been transformed from application processors to account managers.

"Our underwriters look at the risk in its entirety now," says Koscielny. "The decision is more difficult than 'Is this in the right tier?' or 'Is this eligible?'" That routine decision is made by the system.

Now, the underwriter is asking: Is this the right risk? And what are the other qualifications of this risk? Does this include a car, a house and the boat?

What's more, underwriters are working more closely with agents than they did before, says Koscielny. "That's a much more rewarding job than calculating the pricing tier for every piece of business," he says.

Register or login for access to this item and much more

All Digital Insurance content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access