Support APLA documents & directives for reprompts
The new APLA Documents are great for providing rich audio responses at runtime to users.
However, we are still stuck using the reprompt object in responses, and this can only contain PlainText or SSML.
For a user who has just experienced an immersive experience but then hears a low-quality reprompt (especially when the audio element is used), this could seem disjoint.
It would be great if we could use APLA documents in multimodal reprompts too :)