Several years ago, Allstate saw the growth potential in the burgeoning small business market and launched an initiative to reinvigorate its commercial insurance operation.
The company decided to target businesses with fewer than 50 employees in several vertical markets, such as accounting firms and florists. Focusing on select markets would enable Allstate agents to develop expertise qualifying them as “trusted advisors,” says Peter Logothetis, senior vice president and group CIO for the Northbrook, Ill.-based Allstate Business Insurance (ABI). “The name sends a strong message that Allstate agents are more customer-centric than product-centric,” he adds.
However, many of the company’s 10,000 exclusive agents and 2,000 independents had been concentrating on the consumer market, and they needed help getting up to speed on commercial lines. “We found ourselves getting a lot of calls for very basic information,” says John White, IT director for ABI. Wait times stretched to several minutes as call center representatives told agents which menu to pull down or button to click, or answered simple questions about underwriting guidelines for a certain class of business.
“We didn’t want the call center to serve as a help desk,” White explains. “It was intended to be an extension of the agency’s sales force, helping with clients who required specific expertise.”
Creating an Intelligent Virtual Assistant
Company management realized that it needed a knowledge-based help system that could provide agents with quick answers as they worked with clients. After exploring a number of options, including an internal system, Allstate decided to work with Earley Information Science (EIS) to create an intelligent virtual assistant that would make information more accessible and usable.
The first step was to gather information from channels, including the call center, agents, and the sales and marketing team. “EIS then helped us analyze the data to understand what content needed to be in the system,” White says. That included step-by-step instructions for common scenarios, such as providing proof of insurance for contractors or modifying coverage on a business owner’s policy.
The result of this intensive effort was the Allstate Business Insurance Expert, or ABIe, an online, avatar-based expert system that can talk users through a process. Because ABIe (pronounced Abby) is entirely Web-based, no hardware or software installation was involved. It’s housed in the same external data center as the policy administration system with which it’s integrated. After a small, successful trial, ABIe was rolled out to all of the company’s agents.
Users simply press F1 to access ABIe, which presents context-sensitive information relevant to the task at hand. “For example, if you are on the ‘change vehicle page’ in the auto policy section, it lists the top five things people search for on this topic,” White explains.
Users can request a specific piece of information by typing a few words, such as “underwriting guidelines for a delivery vehicle.” ABIe knows whether a user is an agent or underwriter and presents the appropriate information, tailoring the answer for the state in which the policy is being written.
Up-to-Date Content Is Critical
Content is continuously updated “so it doesn’t grow stale,” White notes. One person on staff is dedicated to curating and maintaining content—using data analytics to monitor searches, results and emerging trends—along with user ratings of content.
The focus on quality content is paying off. Overall, 90 percent of users get the answer they need after three or fewer questions.
ABIe now handles more than 25,000 inquiries each month—requests that previously would have gone to the call center. “Wait times at the call center have dropped to below 20 seconds at peak load,” White says.
The nature of calls has changed, too. “The call centers are now doing more high-level, value-added work, not just answering basic questions,” he adds.
Agents have enthusiastically embraced ABIe. By tapping its ability to provide immediate answers to specific questions and step-by-step instructions, agents can complete the quoting process without having to interrupt a customer interview to phone the call center.
“When agents can be self-sufficient, acting as trusted advisors without involving a third party, it’s a better experience for the customer,” Logothetis points out.