See the image…. the results returned seem to be way off the mark on "How did Paul die?" I assume compared to having other prioritized resources? What else could be causing this? Wouldn't it make more sense for the AI to do a more complete survey of my library to present results that truly appeared to answer the question?
Look at the difference between my smart search, and the screen from a mp seminars smart search of the same question. Theirs is much more accurate in trying to provide resources to answer the question…
My results:
Morris Proctor seminar results: