Normal Technique for Iterative Development of Prompt Using Multiple Cases

In the Furniture Fact sheet example in “ChatGPT Prompt Engineering for Developers”, Dr. Ng suggests first trying 1 case and then once mature, testing with multiple cases. When doing this multiple case testing, do prompt developers typically put in 1 case at a time and evaluate each one before trying another? Or is there a technique to have ChatGPT eval multiple cases from a file?