Interesting discovery about expert systems from KM and me:
From Working Knowledge, Prusak & Davenport, p. 137:
One recent research study found that only a third of the expert systems developed in the 1980s were still in use by 1992. The systems were abandoned less for technical reasons than for organizational ones - their sponsor moved on, there were difficult politics in extracting knowledge from experts (or representing users are less than expert), or the systems were viewed as too expensive to update and maintain.
The claim has always been that expert systems weren't adopted because it was too difficult to "extract" knowledge from the experts. And that the systems were only able to recommend actions based on the context in which the expert(s) provided the knowledge. Certainly, this is reflected in the "too expensive to update and maintain" comment. But just like any other initiative, politics and sponsorship are important as well.