Building an AI Coach that speaks "Human" – What are your game analysis pain points?

Hi all, 

I’m working on a project to bridge the gap between "engine evaluations" and "human understanding." I'm building an AI Coach that doesn't just give you a number, but discusses the game with you like a real person while providing deep data insights.

I want this tool to be useful for everyone from beginners to masters. To get it right, I need to know:

* What frustrates you most during your post-game analysis?
* What is the one thing you wish an AI could explain but it currently can't?

Whether you are 400 or 2400 ELO, I’d value your input. Looking forward to hearing your thoughts!

By the way, this is for a startup competition. Any advices are welcome. Thanks^^

yes, kind of. there are multiple way to accomplish this. I am basically trying to explain how stockfish will do in natural language with LLM’s help

not only the best move. I might also explain the second best. and the highest winning rate move(if recorded in database)