I wonder if anyone knows if AVSpeechUtterance supports SSML?
From their documentation it seems like they don't.