The AI agent currently does not appear to have the capability to collect unanswered questions and Flag them. This creates a major gap, because there is no automated feedback loop to identify missing information, unresolved user queries, or areas where the KB needs improvement.
