One of the first things we are taught in school is that we “learn from our mistakes”. But, this only works if we identify issues (mistakes, failures, etc.) and understand “why” they occurred so we can prevent them from recurring.
First challenge is to build a scalable capability to identify, capture, contextualize, and categorize issues so they can be “mined” for ongoing learning and build upon a federated set of knowledge repositories.
Issues are “types” of problems, which need to be contextualized so they can be analyzed individually, cross-referenced with other issues to detect synergies, dependencies, and ensure they are properly decomposed to one or more points of causality. The above analysis requires numerous techniques to assess complexity, run comparisons, conduct correlations, and determine root cause.
This type of analysis is time consuming and requires numerous passes through the demand repository (laterally and vertically). In addition, forward and backward chaining through current & past issues is needed to build learning patterns and continually optimize discovery and understanding.
As the demand repositories (all 3) grow and mature, so too will be the challenge of conducting forensic analysis and ensuring optimized cycle times for addressing problems in a timely manner.