An obvious integration point would be Alexa (#1) and Google Home. I've been working on a voice control library in Node called Sonus that gives you customizable offline hotword detection and streaming recognition.

It's super lightweight and runs really well on the Pi. It's currently used in a few large projects like my smart-mirror and Picco.

All you'd have to add in terms of hardware is a USB microphone/webcam like the PlayStation Eye. I'd be happy to lend a hand integrating if anyone is interested.