I like how agents just answer the question, giving a simple 'Yes' or 'No' if that is appropriate for the question

How do I train humans to do the same?