Interesting Problem
An interesting point raised by SS. I read a paper by a social scientist about using 'grey data': the sorts of noise in the system that doesn't lend itself to conventional analysis.
Some time ago I was working with an engine manufacturer to develop a CBT on a new engine. The problem was that their main workforce had aged and was all coming up to retirement. So, how best to capture hundreds of years of knowledge before it walked out of the door.
The proposal was a conventional CBT based on the manufacturers documentation running alongside a 'knowledge base'. Every time someone found something not quite as the book says or new (the engine was new and additional snags were emerging as operational hours were building) these were submitted to the knowledge base. A couple of the 'real' experts would be responsible for reviewing the input. Their job was not to provide answers as such but to aggregate the data into knowledge, cross reference to the CBT and to commission changes or additions to the CBT as experience built.
Aircraft technical issues are much the same. Snags written up and then signed off 'NFF' could be grey data. You see the same with airport arrival briefings. Management of the arrival process is 'expertise' but we don't do much t capture the dynamic nature of that expertise and make sure it is shared around.