More open/extensible APIs. I want to be able to register my own actions with them with ease. Right now Alexa seems to be the most flexible, but even then you go

More open/extensible APIs. I want to be able to register my own actions with them with ease. Right now Alexa seems to be the most flexible, but even then you gotta do some serious hacking to get it to work. Another thing I was working on at Microsoft back in the day was contextual deep voice search mixed with touch. Essentially the UI pattern was tap and hold on an app, then speak. The voice command would be sent directly to the app itself so you could have it be recognized by it. Example: hold your finger on the pictures app and say, "show me Mike". It'd then bring up pictures of Mike.