Because the sensors are not listening their radio transceiver (RX), only transmitting (TX), by default, to save battery.

Yes, but it's a protocol oversight to have zero confirmation whether a control command was received (and applied).

It's a mesh network, so a response from the central controller could take a long time. The user probably doesn’t want to wait that long, you don't know how long it'll be, and it would burn a lot of batter life.

I used to work on zigbee network embedded firmware, i wouldn't put that in my home if you paid me.

It's all valid trade-offs until you require a confirmation of such a control command.

Don't let curiosity kill the cat, what's so bad about it? :) Is it Zigbee itself or the firmware quality? IoT without the S?