Bob Balahan pours scorn on Enterprise Location, and claims you don't need fancy software to identify experts – you just have to ask around.
Asking around is a great algorithm. But it just has one or two small distortions.
I was a Senior Member of Technical Staff at Texas Instruments, and I was also a member of the committee that voted new members onto the Technical Staff. The Technical Staff was supposed to be made up of employees who were both technical experts and mentors/role models for more junior staff. However, some members of the committee had a pretty fixed idea as to which departments/jobs contained expertise. There was one guy I was keen to see appointed onto the Technical Staff, because he had outstanding technical expertise, but he was in a sales role because he also had outstanding business skills. And he was in a country with a small office, where a few good people had to wear many hats. Many people were reluctant to recognize his technical expertise, were prejudiced against him because of the job he was doing.
The social process advocated by Bob has a tendency to overvalue people with certain symbolic tokens of expertise – job titles or white coats or PhDs – and to overlook people who lack these tokens or refuse to carry them.
Furthermore, perceived expertise is itself a social process. Someone who disagrees with prevailing opinion may be downvalued as an expert.
I certainly don't want to lose asking around, and any KM software I use is going to have to enhance/complement it, rather than replace it. But if KM software gives me a significantly different result, then that might tell me something important – both about the subject of interest, and about the social knowledge processes themselves.