We've been playing with the idea of trying to answer inane questions as if they were actual serious business problems that require solving.

It took some doing to get ChatGPT to go beyond just making up some nonsense that sounded good, but we documented all the prompts and responses that actually got used in the process, and there's a reasonable sort of logic to how and where it landed.