This seems really smart. It would be fun if the author could post some metrics here. I'd be really curious if you guys ran a split test of audio vs text feedback to see how the numbers change. It seems obvious this will lower the barrier for many people to leave feedback. If you have hard numbers on this that are convincing, they should definitely be front-and-center on your website.
One potential concern is that spoken feedback on the whole will be less precise and clear than written feedback. In general though I've found feedback from users is generally pretty poor on the signal to noise ratio anyway, giving only broad strokes of where problems are, so this might not matter. As a fun bonus with audio feedback you can pick up their tone of voice to gauge just how pissed off and frustrated they really are!