This seems a bit similar in nature to Waverly Labs Pilot, but I couldn't understand how exactly Travis is operated (even after browsing the landing page and watching the video). Does it has a microphone and a loudspeaker? Assuming I got the operation mode right, I also couldn't understand what kind of the technology is behind Travis (the one powering speech recognition, machine translation, and speech synthesis) and where it comes from.

What makes this so different from the powerful Microsoft and Google apps?