We can say there are several key point to make the voice assistant?
Make the standard hardware while it is not opposite to the user habit. Directly speaking, the new products must avoid the mobile and PAD, or it must be covered.
Rich back-end content. For the voice assistant, hardware, music even call are all the content. The content need to fill the long tail content from the beginning to gradually . Extending its meaning separate the Alexa and Echo, and make it back-end content as much as possible.
Establish technical advantages. Here, the technique includes both pure algorithms such as deep learning and domains that are closely linked to physics such as acoustics. When the content did not increase to a certain extent, front-end techniques (acoustic and speech recognition) are more semantic than semantic. As they affect the speed and accuracy.