Why can't I enter a prompt first when "replacing content" with the new AI feature?

I love the idea of being able to populate screens/wires with sample content, but why can’t I tailor a prompt to get the correct content first? 

Currently, I have to select the frame, press “replace content” and then wait for it to change everything before clicking “make changes” to then provide it the context needed to replace the content with information specific to the app I design for.

Considering AI provides better outcomes with more context, why make me wait for the inevitably wrong content to populate before providing this option?