The Vision AI feature is really useful, and necessary in the market today with all the competition that exists in this space. However, Vision still continues to use American English when making recommendations, and is frustratingly inconsistent in the recommendations it makes between notes; having the same issue across two notes and not flagging each occurrence of the same error.
Vision AI recommendations must be thoughtful, reliable and consistent in its methodology.
I know that Vision AI is a little different compared to the other AI models that exist in the market, but it still has to work and be a feature that can be trusted and relied on otherwise it has no place of being.