Can You Negotiate the Twists & Turns?

When asked to conjure up an image that represents the motions of an insurance company, most people probably wouldn't think of anything speedy and nimble—say, a ski racer or fleet jungle animals such as jaguars or gazelles. Insurance companies, and the technology infrastructure that has typically served them, are more likely to inspire images of the more lumbering species of the African savannah. But insurance companies are facing intensifying pressures—competitive, economic and regulatory—that demand greater agility from their supporting systems.

But those systems themselves are crying out for reinvention. Since wholesale replacement of rigid, monolithic legacy systems is not a viable option for most insurers, evolving approaches, such as maturing Web services technology and "on-demand" computing, offer hope—and hype.

In the past, insurance companies' IT departments were characterized by a certain inertia, working off a firm and predictable set of business requirements. That has changed dramatically in recent times, according to David Annis, CIO, The Hartford (Hartford, $167 billion in assets). "In the old days, we knew exactly what the system was intended to be, how it was intended to be used, who was going to use it, and what it needed to do. And we built rigidly to those specifications," Annis says. "If you did that today, it's likely the environment may have changed substantially as you get things implemented." What's needed today "is to engineer technology environments where you might not know all the requirements up front, but which will enable you to respond flexibly to meet requirements that emerge over time."

Catching The Competition

Among the challenges driving those requirements are expectations of online delivery of information and functionality more in line with insurers' financial services siblings in banking and brokerage, which have simultaneously become their competitors, according to John Stone, financial services practice director, Software AG (Reston, VA). "Those business pressures are combining to cause insurance to come out with new products, and to change products relatively quickly in response to changing market demands, during a time that's very volatile financially. At the same time, they need to do so very, very efficiently," he says.

Moreover, as more insurers have become public companies, their technology investment calculations have taken on a different character, making that efficiency challenge all the more acute. There isn't a lot of cash around to create new infrastructures in order to satisfy the demands of the day, and nobody is going to rewrite a system at a cost of a hundred million dollars when the old system is doing the job. "From a business standpoint, then, the trick is to get more out of the same old infrastructure without drastically altering it, and without taking on the financial burden or risk of doing that," says Stone.

It doesn't make matters easier that many of today's systems were created up to two decades ago or longer, before anyone foresaw the repeal of the Glass-Steagall act, which restricted banks and brokerage firms from direct competition with insurers. What is being asked, Stone suggests, is a little like taking a 1970s Cadillac and trying to turn it into a 2003 Porsche.

"One of the fundamental infrastructure challenges that they face is nothing less than getting all of their systems to talk to each other and to do so in real time," Stone says.

XML is rapidly becoming the lingua franca for communications among disparate systems, and Web services presents the latest incarnation of services-based architecture, which enables agnosticism among communicating parties, Stone asserts. "Where the system is, how it works, what the technologies are, all of those things in an XML environment become transparent to me—and that is a huge enabler," he says. "It's also not super-expensive, which enables companies to implement new systems more quickly, and also enables them to finally become independent of all the differentgenerations of technology that unfortunately make up their infrastructure."

In the post-dot-com bust period, it almost goes without saying that enthusiasm for technology promise should be tempered with caution. "We've all been here before, where we thought there was an enabling technology that was going to radically change the way technology was delivered and create business value," says Michael Bernaski, who leads Accenture's (Chicago) North American insurance practice.

That's why Accenture's view of XML and Web services is that they're excellent tools if properly applied. "We're very early on in the learning curve, and very early on in terms of the types of software and skills that are available to implement these technologies," Bernaski observes.

He sees such approaches within an atmosphere of greater seriousness and pragmatism, with a "back-to-basics" focus on core distribution products and policy and claims systems across the industry. However much this is owing to current economic and market conditions, it is also due, Bernaski asserts, to the effect of having more tech-savvy people in the executive suite, as the business rookies of the technology revolution mature. "They're too pragmatic to believe that EAI enterprise application integration is going to universally make technology more flexible. They might lean toward something like a product factory focus for policy admin as being a key to flexibility, but when it comes to blanket technology solutions, like Web services, XML or EAI, they're not buying the silver bullet anymore."

The Hartford's Annis might agree, but emphasizes the continuity between past and present offerings. "Even before you get to some of the specific technologies required for Web services, we started building services-oriented architectures a couple years ago, where all these big old applications started to get broken down into a set of more discrete services that you can now use in different ways." he says. "We built the application integration architecture that lets us take the old applications apart, and put them back together in much more flexible ways."

The challenge, Annis says, is to continue to reinvent infrastructure while functioning in the every-day acts of doing business and servicing customers. The industry has already passed a critical round in moving toward such a state, he argues, "bringing in a bunch of integration tools that let you put these system services together in more flexible ways," and achieving milestones such as "creating an enterprise directory service, some new security services, bringing in middleware for application integration, bringing in portal technology, and content management delivery tools," he says. "Those are all the things which have allowed us to take services which traditionally had been delivered to employees, and deliver them to our agency distribution partners through an agency portal, or to our customers through a call-center platform."

An important next step, Annis asserts, will be moving to more of a utility data center environment, enabling "getting more bang-for-the-buck out of the computing platforms that everything runs on." In the old world, every business application had its own hardware and software platform, servers, storage, etc. In the utility model, "it becomes more of an environment where all the applications are built to a set of platform standards, and now we have a certain amount of computing, network and storage resource. We're able to dynamically reconfigure things to respond to the needs of the business," Annis adds.

IBM (Armonk, NY) is touting such an approach in a big way, according to Mike Adler, global leader, insurance consulting practice within IBM's Business Consulting Services (BCS). "We're working with our clients to help them become more of an on-demand business, where they can create more variable cost propositions to deal with some of the cost pressures they face," he says.

The idea is simple enough. Insurance companies can reduce redundancies and under-utilization of resources by converting fixed expenditures on those resources to variable costs by getting out of the business of providing non-core functions to themselves."

Among the inefficiencies that an on-demand approach can address is that of the underutilization of servers, which at insurance companies run on average 20 to 30 percent of capacity, according to Bill Pieroni, general manager, global insurance industry, IBM. Through a "grid computing" model servers can be "strung together across the enterprise, so you bring utilization above the 80 percent mark and dramatically consolidate servers, reducing redundancy and excess capacity."

This is only part of the benefits an on-demand approach provides in making companies more adaptable to shifting exigencies, ranging from product launches to man-made catastrophes, according to Pieroni. "On-demand computing can provide a very broad set of capabilities that include grid, but also include things like claims processing," he says. Thus it serves to help companies navigate boom and bust cycles, "but more importantly, it serves to create a utility-like computing model for the majority of the costs associated with operating a business. And reducing your fixed cost and 'variablizing' it really creates consistency in your earnings, transfers items off your balance sheet, reduces your cost of capital, and reduces the volatility of your earnings."

A Few Years Off

The promise of on-demand is regarded with some skepticism by Accenture's Bernaski, who uses the power-grid metaphor to make his point. When it comes to connecting to resources, "you've got a three-pronged outlet," he says. "You could plug a two- or three-pronged plug in there, but that's the limit of flexibility that you ultimately have. Any device that gets plugged into it has to either have a two- or three-pronged standardized adapter as part of it." Thus, Bernaski adds, while the notion of on-demand has been around a long time, "companies can't do things differently today anticipating on-demand's benefits; we don't believe that it's something that will materialize change in the market over the three- to five-year horizon."

However, Pieroni points to uncontroversial aspects of on-demand that can yield tremendous benefits—albeit benefits that many executives apparently lack the vision and determination to leverage. For example, when it comes to outsourcing, "deferring the potential to save several hundred million dollars annually on IT because of the unwillingness of CEOs and senior leaders to engage at the appropriate level of detail in transforming the organizations is really not an excuse, and yet that is precisely the issue," Pieroni avers.

---------------------------------------------------

Nationwide's Utility Infrastructure

Nationwide Insurance ($24.5 billion in assets, Columbus, OH) is developing a service-oriented technology infrastructure based on principles of loose coupling, according to Srinivas Koushik, enterprise chief technology officer.

Having standardized platforms and reduced variability in its production environment, Nationwide is currently working toward the capacity to apportion work over disparate systems.

Nationwide had planned a project to extract and analyze massive amounts of actuarial data to drive business, but was held up because of the complexity and the expense of needed processing capacity, according to Koushik.

But it's not as if such capacity is lacking: The firm has roughly 40,000 computer-equipped employees, much of whose processing capacity is idle. The challenge, according to Koushik, is to "see how can we steal some of the cycles from the desktops," he says, adding that a pilot is planned for later in the year.

A grid computing approach is considered a very promising future option for projects at Nationwide. "Like most companies, we've got about 35 percent going toward strategic or new development projects and 65 percent spent on maintenance of the hardware, software and the applications that are running on top of it," Koushik says. The on-demand computing principle has the potential "to switch that ratio around, or at least rebalance it so that we're at least at a 50/50 state," he adds.

-----------------------------------------

State of Readiness

It's not just bad news that can lead to increased demand for business agility. Probably the biggest challenge faced by American Family Mutual Life Insurance Co. (Madison, WI, $10 billion in assets) is its own success. "Last year several of our lines of business grew very quickly," says CIO Byrne Chapman. "If you add to that things like the Terrorism Act, HIPAA and the need to reduce unit costs and the expense ratio, that puts a lot of pressure on us to change quickly"—a tough demand for an IT department with about 1,000 core employees and a budget of around $230 million.

The first step toward maximum flexibility is the placement of internal IS consultants with each line of business, Chapman says. That collaboration involves creating quarterly rather than annual plans, and also prioritizing projects quarterly, he adds. "It gives us the ability to bring on or shrink resources, giving us the maneuverability to do what the business needs done, when it needs it."

American Family has also taken something of an on-demand approach to computing resources, guided by a dedicated capacity planning group within IS. "We always look to buy scalable solutions and just-in-time buys based on what's going on in our capacity planning reports, and we spend a lot of time measuring what's going on in our infrastructure," Chapman explains. "We have two large data centers and a large network, which gives us the ultimate in flexibility for where we apply computing resources."

Anthony O'Donnell has covered technology in the insurance industry since 2000, when he joined the editorial staff of Insurance & Technology. As an editor and reporter for I&T and the InformationWeek Financial Services of TechWeb he has written on all areas of information ... View Full Bio