Losing the Ball and Chain

The problem facing insurers that want to upgrade policy administration systems is it takes money to retire the old systems and buy and install the new ones, Jenkins says. Because of the prohibitive cost, he notes, "Carriers that merge with [both having] legacy systems may not ever have one homogeneous system." He points out 10 years ago, the number-one priority of CIOs was replacing legacy systems, but many simply turned to Web-enabling those legacy systems. "Now, there's a serious movement to replace legacy systems. The market is up over the last several years, so the money is becoming available," he says.

Not many policy administration systems were sold in the last year or two, according to Jenkins. "Obviously, that's changing now," he says. "Companies realize their old systems are business inhibitors. They can't get the data out to make decisions. Technology issues are involved, and COBOL programmers are no longer around. Legacy systems are not scalable or robust enough to absorb new information or more volume."

Carriers, especially in cases where two companies with legacy systems merge, must pay twice as much and use twice as many people to support both systems, he asserts. Still, the high cost of conversion keeps many insurers in the legacy arena. "[Legacy systems] do work," says Jenkins, "but there's a lot of drawbacks to that. Old systems don't lend themselves to straight-through processing. And it's very difficult to bolt on things such as predictive models because old systems aren't flexible enough."

Jenkins acknowledges the presence of middleware solutions in the market that can tie disparate systems together, "but again it's extending the life of the legacy systems. You may be throwing good money after bad if there's a good reason to get rid of legacy systems." IT departments want to be able to reduce complexity, he says, but "some of these systems date back to the 1970s. The people who understand them have retired. There's no current documentation. These systems have been customized and bastardized to the extent the old documentation doesn't fit anymore, so it's a mess."

How does a carrier know when to pull the trigger on replacing legacy systems? "You have to consider what the business needs are that would drive it," Jenkins comments. "Are the systems inhibiting [the company] from being competitive? You want to bring things to market quicker. If you have straight-through processing, for example, you can issue policies almost instantaneously." On older systems, he indicates, quoting can take between three and five days. Other policy processing can take 15 to 20 days, he continues, "and that's not providing customer service. The agent or broker is going to go with the company that can turn the business around faster."

As to how long it takes to make the transition from legacy systems to more advanced technologies, much depends on a company's lines of business and territories, Jenkins says. The transition at Penn National involved only nine states and was not that complex, but it still took eight months to a year for one line of business, he reports, adding, however, that a company can update some lines concurrently. A company also may be able to reuse some of the tools from a previous transition, so other lines can be done more quickly. In addition, he suggests, "you don't need as many developers, because of the power of the tool, and you can get the users doing some of the work for you internally."

The current regulatory environment in the insurance industry also is having "a lot of impact" on policy administration systems residing on legacy systems, Jenkins says. Sarbanes-Oxley, HIPAA, "and even the Spitzer stuff" increase insurers' need to get to data quickly from an auditing standpoint, he states. "We want to trace who did what; the reporting requirement is embellished. There is a need to establish checkpoints. Older systems don't allow that transparency of data; it's difficult to get the information out of COBOL code, and it's very costly to do this as well."

The additional reporting requirements, he says, "inhibit the move away from legacy systems, because they take up IT budgets. The regulatory environment is slowing us down. It's slowing down our ability to deliver better customer service. Between state and federal regulation, we have a two-headed dragon."

According to The Hartford's Plotkin, when it comes to policy administration systems, "there is there is no silver bullet, no single solution that everyone is moving toward. We have invested years and millions of lines of code and a lot of money in our legacy systems. The struggle is rationalization of what we already have into new solutions."

The key to such integration, he maintains, is further development of standards in property/casualty industry. "We don't have fully enforced standards," he observes. "It's voluntary. We're pushing for more and more standard components in these tools, such as ACORD. Standards allow us to leverage own architecture and ensure our data is highly usable.

"We've invested heavily in SOA and componentization," he continues. "We're currently investing in a portal front end–one view for everyone who looks at Hartford. So, we need to be sure whatever administration system we buy integrates with that." He adds Hartford has established business rules that also must play nicely with an administration system.

"Most [policy administration] packages come with their own rules," he says. "We need to work with them to make them work for us. We don't want an admin system vendor to tell us what to do with our systems. We're looking at an administrative solution for commercial lines that will fit into our overall environment. We're looking for synergy."

As for the current regulatory climate, Plotkin comments the key to dealing with the changes is the ability to archive and manage data. "In today's world, I'm saving everything," he notes. "How do I do that? I need to make sure the vendors are working with us. Smaller vendors tend to go to virtual or smaller insurers; as they scale up, they're not prepared to handle more complex systems.

"The regulatory climate is a very complex animal, and we've been living and breathing it for a long time," he says. In order to stay abreast of requirements, he advises, insurers must retain appropriate data and volumes of data and have it accessible to regulatory bodies. "I also need to store data I may not have stored in the past," he says. "This also could require insurers to store data in different ways than in the past in order to make sure the data can be delivered within a specified time frame."

Nickler, at consulting firm BearingPoint, echoes the CIOs' views in saying the biggest challenge in policy administration technology will be "a need to consolidate legacy platforms. Most companies today haven't done that, so they have a number of legacy environments," he says. "Most big companies have more than 10." He agrees while money to go for consolidation hasn't been available in the recent past, companies now want to speed up that process.

"Time to market is an issue," says Nickler. "Companies need flexibility. Carriers are looking to migrate to new platforms right now. Elapsed time to provide new platforms is much longer than one year, however. That can be afforded only by a company that has a consolidated environment."

Those companies that have consolidated legacy systems, Nickler believes, are in a better position to focus on new product generation. Such consolidation, he says, can reduce the complexity of the interfaces. On the other hand, he notes, "the cost to introduce a new system for life insurance personal lines really is expensive; it easily could be $50 million plus."

Nickler points out carriers typically want a quick return on their investments, something they are not likely to see with lengthy systems update projects. "It takes a long time until the client feels something has changed, and that's the problem," he explains. "Clients are becoming more demanding, and changes made internally will not always be evident to the client."

He stresses the best approach for companies that don't want to do a complete overhaul is to make changes in discrete components. That way, he says, "the client can see the changes sooner with new products. Each business has to decide which services are most important, then discuss with IT how to provide improvements."

Commenting on the effects of regulatory issues, Nickler contends, "You need to put investment into internal controls to define processes and demonstrate the security of the data. You also must guard against fraud. Most of the effort is ensuring your own quality, and that's something we should have been doing anyway."

Over the next two years, Nickler expects to see more "service-oriented architectures (SOAs) where you can take out components." He expresses doubts policy administration systems that can be used across international borders will be the standard, "just because the regulatory environment and the structure of the business is so different." He adds the most important issue to be addressed in the next two years is definition of standards. "There are some ongoing initiatives trying to standardize the exchange of data," he says. "The insurance world will be more and more a network. All parties need to use same data, but we have no standards. SOA gives us much more possibility for standardization."

"In two years, there will be a lot more off-the-shelf policy administration solutions available to carriers than there have been," predicts Jenkins. Carriers then must decide whether to buy such packages or build their own systems. "We're looking at it now," he says. "We're looking at components and picking the best of breed; they're all open source, so you can mix and match. Open source is the wave of the future. I don't think solution providers that aren't open source will survive. It allows you to plug into any distribution channel." In addition, he points out, "The use of ACORD XML is key so we all speak the same language."

Jenkins says he also expects to see "more straight-through processing–one and done." He also foresees more componentized functions and "less touches" (human interventions) in the process with straight-through processing.

For his part, Plotkin says he expects some consolidation among vendors and solutions. "More and more people selling components as opposed to whole solutions," he observes. "We're seeing more best of breed as opposed to large scale." He also envisions systems moving into the front office and being used to aid product development. "We'll have more modeling capabilities. That will allow us to develop and implement tools much faster than we do today. For multiline carriers, it will be very important."

There will be more self-service and Web-enablement for insureds and for agents, Mr. Plotkin predicts, allowing them to do more than they can today. "There will be more and more security built into these processes," he adds. "What drives it is it's something our customers are asking for. It's regulatory, too. We have an obligation to maintain and secure the data. This is the new normal state of affairs."