I find that people are a little hard on your co-worker Joe (i named it Joe, make things easier)... what he said is not exactly right, but the main idea is here.
Some battery pack and inverter can talk to each others through the CAN protocol, is it mandatory ..? No ... is it better ..? I guess for some cases it's better. Will using the AG4 battery pack a problem with a non AG4 inverter ? I can't see why, the included BMS decide for himself to protect his cells, so the pack is safe.
You got to imagine it like ...
1 - First case : CAN protocol : You got 2 people, one is the Chief in his Office and the other is a guy that is in the cells room, he check voltage, SOC, temp .. and report those info to the Chief .. which will decide how much current or whatever voltage got to be sent, the Chief decide .. !
2 - Second case : No communication : You got 2 people, one is InverterChief in his Office and the other is CellsChief in the Cells room, they both decide, even if InverterChief decide to send 58V, CellsChief will convert it to the needed 48V... even if InverterChief let the door open and let access to 100A, CellsChief can decide only to take 50A. They do not know what the other decide, but ... the real responsible for the battery pack is CellsChief, even if InverterChief think he is ?
Like in a real company... those 2 way can be succesfull .. !
Now imagine a complex system, with water pumps for cooling, with vents to be open for whatever needed... the CAN protocol is the way to go, cause everybody on the CAN network is aware on what's happening, battery pack report a hot temp => pump+10% speed, temp is freezing => Water heater on.... you got the idea.
So .. maybe Joe worked on very big systems (i said maybe) and for him .. communication is mandatory which seems fair. Will your system catch fire if they do not communicate ... no, it will not, cause CellsChief is responsible for those cells NOT to catch fire.
My examples may not be super accurate, but .. it's the idea.