After working on multiple AI projects, one thing I have realised is that we should be investing a good amount of effort into the transparency of the process, from input to output.
They call it Explainable AI, but ultimately, it means you make it feel less magical and smarter.
It is easier said than done, but any AI project you are working on, investing efforts in this is going to give you greater rewards down the line.
Meanwhile, I can't explain why I did this.