I was hoping to get structured data. For example, parsing an voice will give results like {"title"... "line_items": [...], "date":...}