[hunter2_]

I am the last person to suggest throwing AI at a random problem, but this actually sounds like a good match for LLM training/prompting...