I would love to use non-standard English words as part of my Skill (like being able to use an acronym as a name etc.) so I personally think it would be cool if there was a way to use advanced pronunciation guides or something to tell Alexa how to say something/hear something. Even if it is just for intents and skill names that would be nice. Where you can use an onscreen (when you are entering in the field) expanded keyboard and you can use all those crazy accent marks that you see in pronunciation guides in dictionaries (
http://public.oed.com/how-to-use-the-oed/key-to-pronunciation/ ). It would also be cool if it would understand the ASCII versions(
http://www.antimoon.com/how/pronunc-ascii.htm ) of those when you send them in the string to the speech output. I know having Amazon add every obscure word and name would be insane but just having an advanced literal that can be used for predefined stuff (like invoke names, skill names, and speech outputs) would be cool. Or a way that the developer's device can be taught how the invoke name is said like a skill specific version of the voice training.
I can't imagine future Echos not having IPA or similar support. There's a ton of phonetics work that needs to be done if it is to become a truly international appliance. I'm stunned that this isn't in the 1.0 version. I'll bet Google's 1.0 appliance will have sophisticated phonetics built in.
Well, it's clear there are a lot more capabilities there than they've made accessible via the ASK. Obviously [i]their[/i] skills aren't limited to PlainTextCards since they can put HTML in them. And there are obvious undisclosed markups so they can insert pauses where they want. Not to mention streaming audio. They just haven't decided to make them available yet. Probably due to concerns about misuse. I can think of a few ways that they might be used to fake out a customer.