The Eureka project started as an artificial intelligence project we designed and built an expert system for Xerox field service technicians that would help them diagnose and solve machine problems, and which would run on their laptops. Our field observations revealed what technicians really needed: not a system for the known problems the problems already in the book help with problems for which no standard solution seemed to apply. To solve these problems, technicians often invent solutions and, equally important, they readily share these inventions with other technicians in their group. We realized that the technician work community could become our expert system! But the key to making this community knowledge sharing system work was not so much the technology as it was the social process that supported the knowledge sharing. The greatest motivator turned out to be fame or, put another way, reputation. Every solution we called them tips would have the authors name on it. And the crucial factor in establishing trust was having all the tips that were submitted to the community knowledge base be vetted by expert technicians by the communitys most trusted members, who would also be the authors peers rather than some distant group of people working for management at the field service organizations headquarters. In this way, the system would literally be owned by the work community itself. Eureka made its debut in 1994, and in the dozen years of its operation it has saved Xerox over $100M in service costs.
Whalen, J.; Bobrow, D. G. Communal knowledge sharing: the Eureka story. Chapter in Making work visible: ethnographically grounded case studies of work practice, edited by Margaret H. Szymanski and Jack Whalen. Cambridge, UK: Cambridge University Press; 2011; 257-284.