With Expert Network, members of an organization can quickly learn which colleagues have the information
they need
You probably know a knowledge broker -- one of those invaluable people in any organization who, although he or she may not have the answer you need, can nearly always steer you to
the person who has it. In a rough sense, that is what
Expert Network -- a Lotus Notes-based application soon to be rolled out on a pilot basis at IBM's Thomas J. Watson Research Center -- is about, according to Kate Ehrlich, who led its development at Lotus Institute. However, while the human knowledge broker's connections are usually limited to a specific area or specialty, Expert Network may serve as a knowledge map for an entire organization. In fact, plans call for the application to be scaled up to all of Research, according to Cathy Lasser, Research director of information technology.
FINDING THE RIGHT PERSON
Scientists at several Research labs were asked earlier this year what they needed most for their work. Topping the list was "a way to find other people who are interested in the things that I'm interested in." It was not surprising, then, that when Bob Easton, manager of collaborative computing at Watson, heard Ehrlich describe Expert Network in a talk last March, he sensed it might be an answer to a lot of researchers' dreams. "When Kate said she was looking for a pilot site for the program, I told her we were very interested," says Easton. "Things progressed quickly from there."
Ehrlich held several focus group sessions at Watson. In one such session, participants simulated the way the application would work. They posted personal profiles on a wall and then walked around the room trying to match questions they had written down to information in the profiles. One woman was seeking someone who had worked with the company she was currently doing a project for. From reading the profiles, she was able to find a person who was already in the room.
BUILDING AN EXPERT NETWORK
The application is built from two types of information: explicit and tacit. The first is drawn from existing personnel profiles created by employees themselves. These profiles convey the main focus of people's work, whom they work with, what they have done with what they are working on, any papers or talks, professional interests outside of work and so on.
Other information on people's interests is derived from tracking the databases they contribute to and the Web pages they visit. "This is tacit information," says Ehrlich. "It's the type of information that is usually hidden and wouldn't show up in official records."
A key issue was what information researchers were willing to share, and with whom. As it turned out, most of the people that Ehrlich spoke to were fairly willing to share information about their work, and even about whom they worked with. "What they were not willing to share," she adds, "were things like home phone numbers, except with members of their immediate group. That alerted us to the fact that privacy is not just an on-and-off type of concern."
TO SHARE OR NOT TO SHARE
Acceptance of the system may depend more on cultural factors than on the technology itself, says Ehrlich, who has a background in social and behavioral research as well as in cognitive science. In this regard, sharing may seem an unnatural process in a society that stresses self-interest. For example, how do you get people to use a system in which one group may do the work and another may get the credit? "One way is to make contributing to the system part of one's job responsibilities," says Ehrlich, adding that reciprocity and a reward-based system might also be needed to encourage people to contribute.
One potential inhibitor to widespread use, the focus groups revealed, was fear of contacting the wrong person. In response, a module called "people in common" was built into the system. "Anytime you find a name, this feature will tell you all the people you know in common," says Ehrlich.
"You can then go outside the system and chat with this mutual contact." The list of people in common is derived from tracking email traffic.
For people on the receiving end of a search, there was some concern about being the target of too many inquiries. "This led us to think of ways that technology might help us to solve this problem," says Ehrlich. "We found that it's easy to add a field to the profile that asks 'What is your willingness to be contacted?' and to use that as a filter. Another approach might be to add something to a profile that directs inquiries to a list of other people. That would allow others in the organization to become more prominent as authorities in their field."
The benefit to the organization may well be an added incentive to participate fully in such a knowledge management system. "We can well claim to have the finest research people anywhere, but unless we can locate and marshal our knowledge easily and effectively, we won't be as competitive as we can be," says Lasser.
In Six Degrees of Separation, playwright John Guare hypothesizes that a chain of
only six acquaintances separates you from anyone else on earth. The only problem in
finding your way to the one you want to find is tracing the links between the intervening people.
"That's an apt analogy to what Expert Network can do," says Ehrlich. "You can find your way to
the greatest riches a research organization has -- the knowledge inside someone's head --
when you may start out without even knowing who that person might be."
Geoffrey D. Austrian, a retired senior editor of IBM's Think magazine, is a freelance writer in Newton, Massachusetts.
