It should be possible to play SSML in voice calls to change pronunciation, language, insert breaks and much more.
