To be fair, video quality is usually great in developed regions where ISP's do not intentionally slow down internet speed via bandwidth throttling.
One exciting prospect is the advent of AR/VR, where instead of streaming 30 1080x720 frames a second, we can stream 60 virtual character positions a second (a la multiplayer FPSs like CS:GO) that operate as stand ins for real bodies / faces. I think games such as VRChat will have a real impact on the enterprise teleconferencing space once the proper market is targeted: replace the childish characters with more "indifferent" avatars, add in some "business casual" background environments, etc.