Revolutions start in the most unexpected places and with the most unlikely heroes. Who would imagine that the conventional wisdom of the Industrial Age would be challenged by copier repair technicians -- "tech reps" -- at Xerox? Or that field research by anthropologists would support a new set of management principles for competing in the Knowledge Era?

The story begins in the 1980s. We were looking for ways to boost the productivity of the Xerox field service staff. Before deciding how to proceed, we launched a study. An anthropologist from the Xerox Palo Alto Research Center (PARC), a member of the work-practices team, traveled with a group of tech reps to observe how they actually did their jobs -- not how they described what they did, or what their managers assumed they did. That research challenged the way Xerox thought about the nature of work, the role of the individual, and the relationship between the individual and the company. It was the first shot in a revolution.

Here's what the anthropologist saw: Tech reps often made it a point to spend time not with customers but with each other. They'd gather in common areas, like the local parts warehouse, hang around the coffee pot, and swap stories from the field.

Think how a garden-variety reengineer would interpret this finding: Here's "low-hanging fruit" -- easy pickings for immediate productivity gains. Simply reroute the tech reps, cut out the conversation, eliminate the dead time -- and pocket the savings.

The anthropologist saw the exact opposite. The time at the warehouse was anything but dead. The tech reps weren't slacking off; they were doing some of their most valuable work. Field service, it turns out, is no job for lone wolves. It's a social activity. Like most work, it involves a community of professionals. The tech reps weren't just repairing machines; they were also coproducing insights about how to repair machines better.

These technicians were knowledge workers in the truest sense. And it was through conversations at the warehouse -- conversations that weren't a step in any formal "business process" or a box in any official "org chart" -- that knowledge transfer happened.

So Xerox turned conventional wisdom on its head. Rather than eliminate the informal conversations in pursuit of corporate efficiency, we decided to expand them in the name of learning and innovation. Using the Denver area as a pilot project, PARC distributed two-way radio headsets to the tech reps. The radio frequency over which the tech reps communicated became a "knowledge channel" through which they asked each other questions, identified problems, and shared new solutions as they devised them.

But the headsets had limitations. For one thing, no one captured the knowledge the tech reps created. The field staff might communicate in real time to diagnose an unfamiliar problem and generate a solution, but the insights often evaporated once they finished the job.

So we took the tech rep experiment to the next level. In France, working with Rank Xerox, PARC recently unveiled Eureka, an electronic "knowledge refinery" that organizes and categorizes a database of tips generated by the field staff. Technically, Eureka is a relational database of hypertext documents. In practice, it's an electronic version of war stories told around the coffee pot -- with the added benefits of an institutional memory, expert validation, and a search engine.

Eureka operates as a free-flowing knowledge democracy, much like the natural, informal collaborations among tech reps. It relies on voluntary information exchanges. Any tech rep, regardless of rank, can submit a tip, but they are neither required to nor are they explicitly rewarded. In Eureka, the coin of the realm is social capital: the incentive to be a good colleague, to contribute and receive knowledge as a member of the community.