Is this a solution looking for a problem? Just because you can call it AI-driven doesn't mean you should call it that or do it.
AI is not the key question; the key questions are what are your limits on panel degradation and how are you going to have a physical measurement to evaluate that?
