Published on 06 May 2011
Validated on 10 Dec 2012
"With predictive analytics, we were basically able to close a hole in our pocket where money was leaking out steadily" - Bill Dibble, SVP of Claims Operations, Infinity Property & Casualty
Infinity Property & Casualty
Leadership Series, Smarter Insurance
While Bill Dibble doesn’t believe in “silver bullets,” he saw predictive analytics as something that could be used to help Infinity address a number of challenges and opportunities. While his area is Claims Operations, he articulated an expansive, enterprise-wide vision for the company and is leading the charge to make it a reality.
Good ideas usually start simple— and Dibble’s idea was. His insight was to “score” claims like lenders score credit, to provide a more systematic, efficient and accurate way to pinpoint fraud. But that was just the beginning. His breakthrough was to leverage the same underlying intelligence to create a smarter claims processing workflow. That made it possible to transform the way Infinity’s agents handle and route claims, resulting in a lesser reliance on external adjusters, lower adjustment costs and—because claims are handled faster—more satisfied customers.
Leadership is...Big picture thinking Bill Dibble pursued predictive analytics to address what he saw as a growing fraud problem as well as the company’s overall goal of achieving a world-class claims capability. However, in framing and then championing his vision, he took an enterprise view— looking at how it could be used across the company as a whole. “As soon as [Dibble] wrapped his mind around it, he became a visionary and an evangelist for predictive analytics—not just for claims—but also for pushing its value across the enterprise.” — Eric Eckert, IBM Business Analytics
Believing the intelligence. Infinity’s case shows that topdown support isn’t enough to make predictive analytics a success. Dibble’s line managers and regional claim managers were instinctively skeptical of something they didn’t necessarily understand. He sees it as a communications challenge. “For predictive analytics to be effective, consumers of the intelligence have to believe the results.” — Bill Dibble, SVP of Claims Operations, Infinity Property & Casualty
400% ROI with six months of implementation; Increase of $12 million in subrogation recoveries; As much as 95% reduction in time required to refer questionable claims for investigation; Increase in success rate in pursuing fraudulent claims from 50% to 88%; Ability to keep 25% of claims within the company’s first notice of loss area (up from 4%), enabling Infinity to sharply improve its Loss Adjustment Expenses (LAE) ratio
In his 25 years with Infinity Property & Casualty, all of it spent in the claims area, Bill Dibble hasn’t lost his enthusiasm for finding a better way to handle them. As he voices his ideas on how to improve claims processing, an evangelical quality comes through that has become familiar to his colleagues, both inside the company and across the industry. Call it a kind of professional idealism, one colleague says, but when Bill Dibble finds something new that works, he talks about it everywhere.
Dibble, Infinity’s SVP of Claims Operations, also has the ear of the Infinity’s executive leadership, who recognize the combined value of his experience and his perspective on the industry. That perspective comes from the frequent contact with his corps of adjusters out in the field, whose feedback from the front lines keeps him up to speed on trends—be they market, regulatory, or anything else—as they’re developing. By virtue of his experience and breadth, Dibble has emerged as a trusted go-to figure in the company, who isn’t yet done making his mark.
Seeking the edge
Back a few years ago, Infinity found itself—in a positive sense—at a strategic crossroads. In the wake of an initial public offering, strong growth and a series of successful mergers, the Birmingham, Alabama-based company had become one of the nation’s leading “nonstandard” insurance providers, specializing in covering higher risk drivers. Now, with their eyes on long-term revenue growth amid intensifying competition, the company’s leaders were poised to make a push into the standard insurance segment.
To succeed in a market dominated by big-name national providers, Infinity needed an edge. That’s the short story of how Infinity set its sights on developing a “world-class” claims capability, one whose speed, efficiency and accuracy would help keep existing customers and attract new ones. Infinity’s vision was no hollow commitment. With the support of management, the company created an internal task force to brainstorm ways of improving. But as subsequent events show, change doesn’t always come from where you expect it, and the path to transformative change isn’t always predictable.
In about the same timeframe, Dibble had been picking up a steady stream of intelligence from his adjusters in the field that, taken together, pointed to a rise in fraudulent activity beyond the “usual.” While gathering his options on how to respond, it was recognized that predictive analytics was a way of detecting emerging fraud schemes, and another contact from IT concurred. With the seed planted, Dibble proceeded to research and flesh out his ideas over the next few months. What grew out of it was a broader vision of how predictive analytics could be leveraged not only for fraud prevention, but also to improve other parts of the claims process— and beyond that, other parts of the enterprise.
Dibble predicated his vision on the idea that claims could be scored in the same way as credit applications. True, the company’s existing system of screening out questionable claims based on “red flags” was already doing this in a rudimentary way. But it was doing so in a way that required human judgment and a good deal of handling, as when cases were directed to others in the company for further review. Such touch points were not only costly to the company, but also tended to drag out the payment of legitimate claims, which threatened to undermine its customers’ satisfaction with the experience. What’s more, the static nature of the red-flag approach made it impossible for Infinity to adapt its scoring criteria to catch evolving or emerging fraud patterns.
Good timing, good message
Dibble knew he had a strong case to make before the executive team and that his timing was propitious, given the company’s increased focus on competitive differentiation. But he also realized that the best way to build support for predictive analytics was to deliver clear, definitive and measurable results—and that meant keeping it small and manageable at the outset. “The important thing was to create excitement and build momentum right from the get-go,” says Dibble. “Thrilling our people in the near-term is the best way to build faith in the broader potential [for predictive analytics] down the road.”
Dibble and his team chose subrogation as the initial test-bed for predictive analytics, since it was deemed to have the fastest and most direct path to payback. If an insurance company (like Infinity) pays a claim to its policyholder even though the accident is not his fault, subrogation is the process of getting the money back from the company (or people) that are liable for it. It’s an often costly, hightouch process, one relying heavily on human judgment, where the odds of success (getting restitution) aren’t certain. In short, it’s the perfect candidate for optimization based on predictive analytics—a point proven emphatically by Infinity’s results.
The company’s approach was to first identify the parameters that characterized cases that were successfully subrogated and then use that knowledge to create business rules that would score cases based on the likelihood of successful collection. This enabled Infinity to add structure and intelligence to what had been a judgment-heavy, labor-intensive part of the business. In the first month alone, the solution increased subrogation recovery by $1 million; after six months that total rose to $12 million. Now—in going back to Infinity’s executives to move ahead more broadly—Dibble had the numbers to back up his vision. “With predictive analytics, we were basically able to close a hole in our pocket where money was leaking out steadily,” explains Dibble. “We were also ready to show that this was just the beginning.”
In the next phase of its predictive analytics rollout, Infinity focused on fraud detection, again employing a score-based approach (driven by predictive business rules) to routing claims for investigation. It starts at the very beginning of the claims cycle—a process area known as first notice of loss—when claimants first report the accident. As data on the accident is gathered by representatives, Infinity’s rules engine builds a rating for each claim that corresponds to its fraud probability, with claims above a threshold automatically referred for review. By getting suspect claims into the hands of investigators within a day or two—a contrast to the one to two months it used to take—they can proceed on the case while the facts are still fresh. That’s one reason Infinity’s success rate in pursuing fraudulent claims has gone from 50 percent to 88 percent.
Intelligence as transformation catalyst
In fact, however, that only scratches the surface of how Infinity is using intelligence to streamline, speed up and otherwise optimize the claims process—all of which represents a realization of Dibble’s vision of a “world-class” claims capability. Traditionally, the first notice of loss area is the first of many customer touch points in the claims process, with the typical handoff made to adjusters in the field. Dibble’s bold idea was to extend Infinity’s analytics capability as a way of creating an optimized, intelligent workflow. The idea is that by asking “smarter” questions at the first notice of loss, Infinity can apply algorithms that sort claims into groups with different handling or adjustment requirements. This, in turn, enables claims to be routed for processing in the most efficient way.
Or—in some cases—it can prevent the need to route claims at all. In what is arguably the project’s coup, Dibble leveraged the solution’s intelligence to redesign the claims handling process. By using predictive analytics to identify the claims least likely to require investigation, Infinity made it possible for the first notice of loss area to handle a full 25 percent of the claims burden itself, claims that would, in the past, be routinely dispatched to adjusters in the field at considerable cost. In addition to giving claimants the “express” treatment—and the speed and convenience that goes along with it—Infinity was able to reduce its overall cost of adjusting claims, an improvement that shows up in its all important Loss Adjustment Expenses (LAE) ratio.
The fact that Infinity had a framework for intelligent processing in place was a key enabler of this transformation. But it wouldn’t have happened without the decision and commitment to make important changes at the personnel, process and organizational levels. In a one-year process, Infinity reoriented its first notice of loss employees to handle an entirely new set of tasks, creating what amounts to a lower severity adjustment force. Employees were retrained, job descriptions and goals were changed—and perhaps most importantly to Dibble—a new culture of accountability was put in place. “Our people at the front-lines taking claims are now making decisions that adjusters in the field would have made in the past,” Dibble points out. “To make it work, we needed to balance that extra empowerment with a greater degree of accountability—and employees who were comfortable with that new balance.”
Infinity Property and Casualty: The parameters of smarter claims processing
Instrumented: As claims information is being gathered, it is being automatically fed into a scoring engine that drives the processing workflow, including referrals to fraud investigators.
Interconnected: An integrated body of customer and claim information— augmented by analytics— provides Infinity’s agents with the means to make the right
Intelligent: Infinity’s ability to identify “express” claims as they are received speeds their resolution, improves customer satisfaction and lowers the cost of adjusting claims.
Going deeper to get smarter
And Dibble is not stopping there. The company is augmenting the intelligence of its claims processing by putting in place a text mining capability that will analyze the content of police reports, medical files and other documents related to accidents to help pinpoint narrative inconsistencies that may indicate underlying fraud. At the same time, Infinity is continuously updating its fraud models to further improve their accuracy. With Dibble as the catalyst, the use of predictive analytics is also spreading to other parts of Infinity’s enterprise, including direct marketing, where it has improved
response rates significantly.
With Infinity poised to expand its market presence, Dibble sees the use of predictive analytics to create a smarter and more efficient claims process as an important foundation of its future success. “Whether it’s fraud reduction, customer convenience or cost control, leveraging intelligence will be increasingly important to the way we differentiate ourselves in the future,” says Dibble. “We’ve shown our willingness to take some chances to make this happen and we have the results to show for it.”
Infinity’s intelligent claims processing solution is -
- IBM Cognos
- IBM SPSS Modeler
- IBM SPSS Risk Control Builder
- IBM SPSS Collaboration and Deployment Services
- IBM SPSS Professional Services
For more information
Please contact your IBM sales representative or IBM Business Partner.
Or visit us at:
View Leadership Series Web Portal for Infinity Property & Casualty
© Copyright IBM Corporation 2011 IBM Corporation 1 New Orchard Rd. Armonk, NY 10504 U.S.A. Produced in the United States of America. March 2011. All Rights Reserved. IBM, the IBM logo and ibm.com are trademarks of International Business Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol (® or ™), these symbols indicate U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the Web at “Copyright and trademark information” at www.ibm.com/legal/a Other company, product or service names may be trademarks or service marks of others. This case study illustrates how one IBM customer uses IBM products. There is no guarantee of comparable results. References in this publication to IBM products or services do not imply that IBM intends to make them available in all countries in which IBM operates.