Lessons Learned about Lessons Learned

Thursday, September 25th, 2008
20080922-KMWORLDArtMurray.jpg

Tom Beckman and Art Murray, of Applied Knowledge Sciences, are presenting on the topic of the development and implementation of lessons learned systems at the U.S. Air Force.

Why Lessons-learned systems are so important to the U.S. Air Force.

Why implement a Lessons-Learned system?
Too much time, money and resources are wasted by repeated errors and duplication of effort.
Organizations are facing unrelenting pressure to do more with less...

I can tell these slides will fly by...

Problems with mainstream Lessons Learned systems

Lessons Learned 1.0:
After Action Reviews --> system
problem: user has to take initiative to search
Keyword searches  do not always lead to lessons needed.
Single-lop learning at best: band-aids, local fixes, etc.

Premise of Lessons Learned 2.0 Research & Planning:
Guided process of capturing rich knowledge
Much of Lessons Learned 1.0 is document centric. Behind them we have communication about those docs (e-mails, news letters, etc.) this is where knowledge emerges...
Goal: Identify knowledge "nuggets" and get those in front of people. Then expand this with stories to complete the lesson.

KM Life Cycle -
Capture/Share/Apply
Lessons Learned 1.0 typically focuses on "Capture" stage, possibly "Share"

How do you apply it?

Application points of Lessons Learned (Kent Greenes)
        Plan work --> Perform Work --> Analyze work --> Repeat

This amounts to learn before doing, learn while doing, and learn after doing. This feeds repositories but the knowledge is often sparse. The key is to connect the knowledge nuggets with knowledge expertise via communities of practice, SMEs, practitioners, peers, etc.

Three tools that support the Lessons Learned system:
Case Based Reasoning
- better precision & recall
- based on feature match proximity and
-weighted feature importance
Using generalization and abstraction to match requests for lessons learned with the library of existing lessons.
Intelligent Agents
- Autonomous software entities that act as a user's assistant to perform discrete tasks.(Alerting agent)
- Perform tasks based on knowledge of the user, tasks, domain, and environment.
Performance Support Services
- Provides services that enable lessons learned transfer
- Services provided at point of use and need

Applying lessons-learned about designing lessons learned system:

Presentation & discussion

Discussion/Comments (0):


Discussion for this entry is now closed.