As a user I would like a means of identifying what issues Kai is likely to generate accepted solutions for versus which are difficult problems and have a low likelihood of an accepted fix.

This is planned to be achieved by a minimum of updating rules syntax to provide a manual means for a rules author to capture a confidence level.

In addition we would like to dynamically update this confidence level by capturing usage information from other developers in the org as they complete migrations.  The intent is the solution server will capture this usage data and be able to provide guidance of what issues are likely to be resolved and accepted and which are more difficult and less likely to be resolved from a LLM.

Engineering's design document: 

https://docs.google.com/document/d/1rSsBrHf_RDyZYmX0OV8MWelB55rQVv-KnyISXK-KFlM/edit?tab=t.0

Tracked in upstream at:

https://github.com/konveyor/kai/issues/761