Run simulationsRun question sets on your semantic model to see where it answers well, where it misses, and what to fix next.