The scenario runs a set of your questions through the AI assistant and scores how relevant the information it retrieved was — a number for each question.