Car maintenance is complicated -- and even expert technicians need help sometimes.
Auburn Hills, Mich.-based DaimlerChrysler Corp. has found a way to help its mechanics help themselves. Its Service Technical Assistance Resource (STAR) Center provides a portal for technicians to answer their own questions via a case-based reasoning (CBR) search tool. Technicians input the vehicle problem in natural language to start a troubleshooting session, which narrows the problem with follow-up questions. The more difficult problems are handled by call center agents, who also use the application.
Launched in 1993, the STAR Center was initially a group of employees comprised of corporate lab workers, district managers and technical advisers who provided specialized advice in areas such as the body chassis, drivability, electrical and power train systems. With the group's help, DaimlerChrysler developed an in-house knowledge base to help the STAR call center agents train themselves on the intricacies of the wide variety of domestic cars.
Soon enough, the company started feeling more pressure from dealer technicians to extend STAR's hours. So in 2001, the company elected to make the tool available to dealerships through its portal. At the same time, executives realized the need for a more sophisticated application that could provide more detailed information than the call center agents were working with.
That set off a year-long search for a new tool -- and DaimlerChrysler
"Our IT department was using a Web portal for the dealers," said Larry White, technical engineering supervisor and CBR coordinator for STAR Center. "They had an extensive set of standards and requirements that any business process owner had to follow in order to deploy a Web application."
DaimlerChrysler created a decision-making team that included members from both the general Chrysler IT department and the IT department that supports the contact center, as well as purchasing and post-sales support. The team also brought in an independent consultant to help with the purchase.
"We used his expertise in developing a feature function matrix that was key to helping us understand what was important," White said.
DaimlerChrysler required each vendor it evaluated to develop a pilot application by building a small knowledge base taken from roughly 200 call records and search examples. In addition to feedback from the pilot project, the team considered vendor viability, quality of professional services available and application development, White said.
Ultimately, DaimlerChrysler selected a tool based on Kaidara's Advisor.
"We wanted something that didn't require a lot of understanding of the software," White said. "We wanted users to be able to enter data into the application and share it with other agents across the company and dealer network." The implementation took one year to complete, but White feels it would probably take much less time for smaller companies that don't have to comply with the requirements of a corporate IT department.
DaimlerChrysler was able to transfer 10,000 actual cases, or roughly five years of data, in five months. Predictably, the company had some call center agents who worried the tool might replace them and were reluctant to use it. A little extra attention and a group of early adopters encouraged those fearful users and helped solve the problem.
The company also tested the tool with focus groups and regional offices before launching it. With 260 applications within the DaimlerChrysler dealer portal, one of the company's policies for including an application is that users do not require a training class to use it.
The company has managed to foster adoption through brochures and announcements and by highlighting STAR during training for new technicians. Additionally, technicians holding for a call center agent are encouraged to use the self-service application.
The system also provides a method for users to provide feedback on how effective the system was in helping to fix vehicle problems, said Priscilla Hedin, STAR Center manager.
"Case-based reasoning -- that's the difference," White said. "The most important thing is the application has given us a tool to share information across departments and the company. We're actually deploying this [tool] in some manufacturing environments for internal use."