I've been wanting something along these lines for a number of years for use in emergency services communications. typically there are radios and speakers stuffed wherever they can fit.
It would be much nicer to feed each radio into a common system and have the result able to be mixed to multiple sets of outputs.
It sounds as if this comes extremely close to what I was thinking of. The only additional feature I would want is the ability to 'position' an audio source. If you have multiple radios you need to listen to, you could position them at different 'positions' in your audio space. This could be done as the audio is fed into the system, but it's much better if you can do it as you feed the audio out to the zones, that way you can have things positioned differently for a person using headphones than you do for speakers in an operations center (where you can actually position the audio 360 degrees around the room)
This sort of capability has a lot of other uses. Think of people who may be on troubleshooting calls where they want to have one phone line open to sysadmins fixing the problem and another open to managers talking about the impact of the problem. Today this is almost impossible to do, but with VoIP and proper positioning of the sources, you could potentially not only differentiate the different calls by position, but also the individual speakers, allowing for much better identification of who is saying what.