Monday, July 9, 2007

Expertise Management

Arjun Thomas, of Gridlock, put together a list of processes for a basic expertise management system in Expertise Management:

  • Identify the experts

  • Describe the expertise of the expert

  • Provide an expert matching mechanism

  • Enable users to communicate with the experts

  • Provide a feedback mechanism

  • Manage the process

The problem I run into is getting the description of the expertise. If you rely on the users to supply their own expertise, you end up mostly with missing information and exaggerations.

Automated systems, like Microsoft's Knowledge Network, seem to work better for the oddball items. The automation has trouble distinguishing among the types of expertise within the organization. It is much easier for me to find some one with information on "arms dealing" than it is to find the best person to talk about "securities litigation."

My ideal expertise system would start with an automated system, that layers in the persons marketing bio, and a manual control on top to enhance and clarify the expertise. I just have not seen a system that does this.

1 comment:

  1. I think you may be hoping for the imposssible., In my experience, there is a tremendous amount of expertise needed in order to develop the proper taxonomy (to classify the expertise), synonym database, context engine and search classification that you need in order to get close to that first automation step. And evenhaving done that you need people looking at the output to properly identify experts...

    To be simplistic by example - If I am classed as an expert in "stocks", then how do you automate knowing whether I am a type of chef, a farmer, any miriad of financial analyst-types, an investor or an accountant?

    Folks that you should look at who have gone as far is I think you can with commercial (rather than academic) applications for this type of thing are Revere ( and expertview (



Note: Only a member of this blog may post a comment.