The point I was trying to make was merely that creating an 'aim-able' directional antenna is practical even in cheap consumer hardware, e.g. beamforming in home wireless APs. I have no idea how tightly those sort of cheap consumer implementations are capable of aiming though. I agree that 5% efficiency sounds pretty bad.
Strangely enough, that whitepaper presents beamforming as a (poor) alternative to Cota, which implies that Cota uses some other method 'aim' its power. No details are provided on what that method is though. Also, they list Technovator as an example of RF beamforming, but from what I can find Technovator's solution uses near field resonant coupling, so not RF beamforming at all.
Unfortunately that "whitepaper" seems more like a promo piece than an actual technical document.