I find many AI features unhelpful - if I need text to image I can do it in Photoshop or Express, I don't need my layout program to do it. What I would love to see AI do is be able to analyze documents and automatically fix accessibility issues, such as reading order and export tagging of headers etc.