Kimberly Harris-Ferrante believes that in commercial lines underwriting automation initiatives, carriers should avoid trying to create "expert systems" that eliminate the underwriter. Instead, insurers should focus on providing "systems for experts"—technology that enables highly-skilled professionals to manage their workloads and time.

"Don't just think about low- to no-touch processing. Think about the ways you can automate behind the scenes, to help with productivity and consistency of decision making, and create an audit trail of how you made the decision so you can tweak the process over time," says Harris-Ferrante, vice president and distinguished analyst at Gartner.

"View automation in terms of activities or steps within the end-to-end process of writing business," says Donald Light, senior analyst, Celent. "By doing so, you could automate a handful or dozens of steps, including decisions, depending on the complexity of the risk."

Harris-Ferrante feels one of the most encouraging developments along these lines in the last 12 months is the concept of case management. "Case management is game-changing in complex underwriting," she says. "Case management melds process automation and decision automation, incorporating business process management-based systems to automate or guide decision making, predictive analytics, and collaboration, ideally baked into a workstation that minimizes 'swivel time.'"

This development has driven awareness that there is need for technology to provide underwriting support to combat "information overload."

"Most of the insurers that we talk to simply have too much information to digest," says Edward Gray, senior director of customer solutions at FirstBest Systems. "There are too many steps in the process, too many places to go look for data, and no time to get to them all."

Carriers need to look beyond processing speed in automation initiatives, maintains Steve Discher, executive vice president at Robert E. Nolan. "Underwriters need the ability to perform sophisticated analyses and take action by region, by agent, and so on, not just by product line," he says. "Building that type of environment is the next phase in automation—having the underwriting workstation tuned to profitability and the skills that go with it, bringing decision-making information together, and so on—not just speed of service."

AUTOMATION GOALS

In addition to bringing information together, underwriting systems are focusing on bringing people together in the underwriting process. Collaboration is a key objective, allowing underwriters to make better decisions by enabling connectivity to underwriting colleagues, agents, and coworkers in other areas such as claims and loss control.

"The first generation of underwriting systems focused on data and analysis, whereas the new wave is around case management and collaboration," says Harris-Ferrante. "You could have an underwriter who needs to go back and talk to the agent to ask them a question about the risk, or you may have a younger underwriter who needs to communicate with someone who is more skilled. The underwriting system becomes an electronic file folder for people to store all the content required to make an underwriting decision plus a collaboration platform for communication."

Collaboration was an objective of an underwriting automation initiative at California-based ICW Group. The carrier writes a number of commercial lines coverages, but targeted its workers' compensation business in a project to install FirstBest's UMS underwriting workstation and agent portal.

Previously, underwriting at ICW Group was anything but automated, with staff having to contend with redundant data entry into four different legacy admin systems: one to record the application and "reserve" the market for an agent; one to perform predictive modeling and help underwriters determine an appropriate price for the risk's exposures; one to generate a proposal; and one to administer business that was written. Installing the FirstBest front end, in a project ICW branded "Snap," allowed the insurer to eliminate the duplicate entry of data by providing a common interface for all systems.

"Our primary objective was to become more efficient and scalable," says Paul Zamora, senior vice present of workers' compensation underwriting. "None of the systems we had communicated with each other, which caused duplicate data entry and inefficiency. We were in a situation where about 40 percent of the quote requests we received we simply couldn't get to because we didn't have sufficient manpower."

In addition to providing an easier work environment and eliminating data entry redundancy, ICW also wanted to automate parts of the decision process and workflow and to provide a platform for online collaboration, both internally and with agents. "Snap is not just an underwriting tool—it's a communication and collaboration platform," Zamora says.

The FirstBest platform supports two methods of collaboration: notes and instant chat. Underwriters indicate they are available for chat when they log in to the workstation. Notes can be initiated by underwriters or agents either from within the Snap system or via email.

"Both methods are nice, but we do find agents actually prefer notes because they don't need to be logged into our portal in order to use notes to collaborate. Instead, they can just use the email system they already have available on their desk or in the field," Zamora says.

ICW underwriters also use the notes feature for internal collaboration. "Referrals are automatically routed by the system, and we can use 'private' notes that are not available to agents," Zamora explains.

That functionality allows ICW to connect staff among various offices to obtain input and needed underwriting authority. "In one branch, it's easy to collaborate by simply getting up and talking to each other, but when you're dealing with staff in different offices it's not that easy," Zamora says. "In the old days, underwriters would have to scan PDFs and send documents to me. Now when I get a referral I can see everything that's going on from within Snap."

The platform also is used by extended underwriting staff, such as loss control, to obtain account information. In the future, ICW plans to enable diary functionality to provide automated follow-ups on safety recommendations and other items.

In addition to enabling better collaboration, Snap has delivered ICW decision support in the underwriting of complex workers' compensation risks. "Traditionally we had been vulnerable because of inconsistency in the decision-making process," Zamora says. Different underwriters would have different views on not only what accounts were acceptable to write, but on what particular class codes and modifiers to apply to determining pricing that was competitive, yet profitable.

"Based on the training level of each underwriter, an agent might get a different response from different people," Zamora says. "We took the underwriting 'decision tree' for a large number of our class codes and went through them with our underwriting staff, determining what questions to ask and translating responses to those questions into a scorecard. Today, a junior underwriter can make a decision just as good—and just as fast—as a senior underwriter. Now, we are as strong as our strongest underwriter."

The process of establishing that decision tree initially included mapping out information for almost 200 class codes. "It was a cumbersome process to sit down and make sure when we are asking a question, we contemplated parameters around not just all class codes, but differences across all states in which we do business," Zamora says. "On the other hand, one of the benefits of the pain of putting Snap into place was having to revisit all our underwriting rules to extract and document that knowledge."

That process is ongoing, with ICW conducting monthly meetings to incorporate additional class codes into the decision tree or modify the rules around existing codes.

ICW measures the success of its automation initiative by several metrics, the first of which is the extent of system usage by agents. "We don't require agents to use the system, so the fact that the vast majority are using it means a lot," says Zamora. "We got to within 50 percent usage within a few months of rollout, and today over 70 percent of our applications coming through the door are coming in through Snap. Agents are using it to bind and submit business and to access quotes."

The second measure of impact is the increased amount of business that ICW has been able to process while maintaining profitability. Improved upload capability of NCCI, Workers' Compensation Insurance Rating Bureau of California (WCIRB), and ACORD application data has eliminated about 70 percent of the manual data entry previously required. Through a combination of reduced data entry, automated workflows, and decision support, ICW has grown its workers' compensation policy count by 80 percent despite a 14 percent reduction in staff requirements since deploying Snap in 2009.

STP

Although collaboration, workflow, and decision support are key objectives of today's underwriting automation projects, insurers haven't forgotten about the goal of STP for new business and renewals.

"We hear about the need for straight-through processing from at least 70 percent of the insurers we talk to," Gray says.

Travelers Insurance was able to increase the STP rate on business owners policies (BOP) in its TravelersExpress program from just 17 percent in 2007 to more than 80 percent today with an underwriting automation initiative. Travelers chose IBM's WebSphere ILOG jRules business rule management system (BRMS) within a service-oriented architecture (SOA) to execute a multivariate pricing model and apply decisioning rules around underwriting acceptability and pricing.

Travelers created its own multivariate predictive model, leveraging its decades of experience writing small commercial policies. The model uses application information and data populated from third-party sources, such as Dun and Bradstreet. Version 2.0 of the TravelersExpress system is scheduled for late 2011 or early 2012.

"Prior to the introduction of the predictive model, we had a traditional rating plan that relied on underwriters to manually apply credits based on their understanding of the risk," says John P. O'Connor, vice president and chief product/platform officer of small commercial lines at Travelers. "Today, we have established flexible, multivariate rating to support precision pricing and customer segmentation, consistent application of underwriting rules, and reduced manual touch points on new and renewal processing."

Additionally, underwriters can focus on more complex risks and rewarding tasks and, because manual underwriting workloads have diminished, agents receive a faster response from underwriters on applications that do not auto-issue. "When you wrap these capabilities together within a quote-to-issue solution, we have significantly improved agent satisfaction and ease of doing business with Travelers," O'Connor says. Making the TravelersExpress system easier to use has also translated into an increase in Travelers' sales force from 4,000 to approximately 10,000 agents.

Extracting underwriting rules from legacy systems and underwriting staff was a challenge, O'Connor says. "We had to find and validate our underwriting procedures, guidelines, and rules, then distill those into rules that could be automated and consistently executed," he explains. "However, doing so has brought us from underwriting as a craft to underwriting as a science." Travelers also created a new department within underwriting to manage rules, perform analytics, and recommend refinements.

ICW has leveraged auto-underwriting capabilities of Snap to enter the arena of "small comp"—a workers' compensation market the company could not afford to compete in previously. "Agents answer questions online, and the application will go down several paths—declination, proposal, or alternate proposals," Zamora says. Today, ICW writes about $20 million in small comp business.

"Automating small comp was a big challenge," Zamora says. "It was easy to determine all the information we needed to get from agents, but it was difficult to determine the right amount of information to get from agents within the five or ten minutes they were willing to spend entering data online."

STP has already become table stakes in the highly commoditized world of personal lines. "If you're a personal auto insurer, you can't survive without STP based on automated underwriting," Discher says. "If you look at large- or medium-sized personal lines carriers, most have fully implemented an automated underwriting process for easy lines like auto. Property is a bit trickier, particularly with large losses throughout the U.S. causing carriers to want to conduct more inspections, so carriers will use automated underwriting to issue policies but then initiate workflows to order inspections."

However, it has been harder to achieve STP in more complex commercial lines. "Companies have not found, or not made a sufficient investment to find, ways of scoring the totality of the risk in larger commercial insurance," Light says.

The problem is not a lack of data, but the law of diminishing returns. "You could have dozens or hundreds of variables, each with several options, combinations, and permutations," Light says. "You may not have enough loss history to assign an underwriting score with reliability. So while it's not impossible to automate decision making for complex risks, it takes a lot of time and might not be worth it for risks you see infrequently."

"Currently we're focusing our efforts on the small commercial risks," O'Connor says. "When we try to understand how rules come into play for larger pieces of business, it's a question of if you have critical mass to have rules for particular segments."

Even if a company can't achieve full STP, there is value to using an "automated underwriter" to at least provide the agent a quick "go/no-go" decision contingent on a more complete review by underwriting staff.

"Early indication of appetite is very valuable," Light says. "Identifying the 'deal killers' at the start of a submission, or using a quick-quote functionality to determine if a carrier is likely to be in the ballpark, is a tremendous benefit to agents."

PUSHBACK AND PROGRESS

No technology initiative is without its challenges, and the first obstacle to the deployment of underwriting automation is caused by how carriers have traditionally viewed the technology. Harris-Ferrante says that carriers' longstanding belief that underwriting automation was primarily a tool to replace underwriting staff and the traditional coupling of underwriting functionality within the policy administration system has limited their view of what they could do with the technology in commercial lines and slowed adoption of stand-alone underwriting solutions.

However, this has begun to change. "Initially, commercial carriers argued their business was too complicated to leverage automation," she says. "Today, they've woken up to the fact there are tools and technology that provide 'systems for experts' that can establish competitive advantage."

Change management is also an issue for companies deploying underwriting automation solutions, according to Discher.

"Organizations that have been used to doing underwriting in the traditional way have a difficult time pulling any responsibilities out of the underwriters' hands, even if data proves the results," he says. "There is definitely management involvement needed in the process, because some underwriters are simply not going to want to let go of any processes whatsoever."

ICW faced this challenge from both internal and external staff. "Even though we had to deal with four systems before, which was very cumbersome, and even though the new system was more streamlined, it took a while for people to get used to the new way of working," Zamora says. ICW addressed the issue by involving users in the change process and by continuing to demonstrate the benefits of the system once deployed.

This is a sound approach, Discher says. "Carriers need to engage the underwriting group to help validate the benefit of the system before, during, and after installation. They need to provide confidence to the underwriting ranks that they are getting to a better place with automation. They need to demonstrate that if the 'black box' can do the lion's share of processing for simple accounts and administrative tasks, underwriters will be able to apply their skills to more complex and rewarding tasks," he says.

"Have focus groups and sessions around the development and impact of automated underwriting," Discher adds. "Take a before and after review of blocks of business that were manually processed versus those that involved automation. Look at the results and put all the data on the table."

A danger to avoid is the possibility of increasing the underwriters' workload when new underwriting tools are deployed. Because automation systems use rules to guide decision-making, the underwriting process must be captured to create rules, and those rules must continually be refined through both auto-learning and through documentation of underwriting decisions taken on complex cases.

"Insurers are increasingly using new sources of information to assist in underwriting decisioning, resulting in slowing down underwriting decision time," Harris-Ferrante observes. "The use of modern underwriting systems can assist in underwriting productivity, as well as document and centralize the underwriting rules so there is improved consistency across decisions and a trail of how decisions were made."

Carriers may also contend with the issue of system overlap when exploring technology designed to deliver automation to the underwriting process.

"Today, most of the policy administration and management vendors do not offer best-in-class underwriting capabilities suitable for commercial lines, which is a gap that needs to be addressed," says Harris-Ferrante. "At the end of the day, insurers have no option but to look for multiple applications if they're interested in policy support and underwriting strength."

CONTINUED MOMENTUM

Although it may have a different look and feel among various lines of P&C, underwriting automation is the real deal.

"It [underwriting automaton] is here and now and it can create a tremendous impact," Discher says. "The technology is mature enough, the data is mature enough, and there are a lot of exciting things going on in the industry. The only question is how you are going to implement it."

Going forward, Harris-Ferrante expects to see more domain-specific customization of underwriting automation technology.

"Large, tier-one insurers will require less insurance-specific business processes and rules off-the-shelf in underwriting solutions, whereas mid-market insurers will need more prebuilt underwriting intellectual property delivered from the vendor. A few of the case management vendors are already taking this route, building out insurance frameworks or templates that can be leveraged during implementation to give insurers a starting point for building out the rules and workflow," she says.

For their part, insurers will increase their understanding of what underwriting automation should mean and how that meaning changes by line of business and case complexity.

"Underwriters are one of the most critical assets within the commercial insurance organization," Harris-Ferrante says. "Tools are needed to help make them more productive and allow them to leverage their skills in the most effective way. That is one way insurers can help meet underwriting profitability goals going forward."

Special Reports /

Weighted Down

Related Stories

Recommended For You