I have tested the enhanced TTS in a flow to listen to options,.
The standard and wavenet voice options (tested a couple of each) playback dynamic data fine from what I can tell, however the neural voice options are not playing back currency value correctly, at least for en-GB.
If I play £164.20, it is being pronounced as "one hundred and sixty four point two pounds" instead of "One hundred and sixty four pounds and twenty pence"
This is forcing TTS to present using expression e.g.
ToAudioTTS(Append("£",ToString(Task.CreditLimit.amount)))
I have also tried with a decimal variable, same effect.
I assume this is not by deisgn otherwise dynamic data playback for currency items is going to be a challenge.
PLEASE, if there is a UI to listen to the various options let me know as painful to have to update flow and publish every time to listen to alternate voices!