Biggest pain point right now is typing manually all the alt-text values for each image, especially when the images all vary.