This is the final installment in a series about natural language and knowledge management. Previous posts can be found here.
So what do we have so far?
- Automated Natural Language indexes need supplemental information to work well.
- Knowledge workers can do a better job of indexing than machines.
- Many knowledge workers are not only willing but eager to get their knowledge into the hands of an audience who will use it.
Given these facts, what work is left for the knowledge management folks to do? Why doesn't knowledge management just flow right from the desire of the knowledge workers?
There is still a lot of work for the knowledge management folks to do, but success does not lie in providing a technology solution that has minimal impact on the work practice of the knowledge workers. The organization has to be ready to accept a change in work flow. Knowledge workers have to be given an opportunity (and sometimes, incentive) to annotate their work. Knowledge consumers have to include searching for information artifacts into their work flow.
My colleague Irene described the situation as follows: "if you want something different to happen, you'll have to do something differently." But what is the different thing we should be doing?
Knowledge management efforts usually focus on technological solutions that will have minimal impact on the workflow of the knowledge workers. They focus on finding ways to mine information from legacy systems while ignoring the plentiful resources of the workers who are already there. It should come as no surprise that knowledge management efforts are so often frustrated; they are working hard to reduce the difference in how people are working as much as possible. They aren't doing things differently, so they don't make any difference.
So what should knowledge management efforts focus on? They should focus on finding creative ways to get the knowledge workers involved in their own knowledge management. How can you find a way to make it worthwhile for them to create their own metadata? Do you do it like the Kennedy Space Center, where the knowledge workers do an end-run around KM? Do you do it like the Japanese company, who gave their engineers a tool that created reports that convinced project managers that a feature really would take six months to develop. As a side effect of this report, the designs had useful metadata that increased their (re-)usability. Solving this problem - figuring out how the knowledge workers can do something different so they can have different results - should be the main focus of KM groups. Technology should be introduced only as a way to achieve this goal.