The way biological neurons self-organize into a neural network could correctly be described as swarm behavior. In fact there is quite a bit of research into this and it's amazing how during development you can move a neuron from one part of the brain into another area, and it will self-correct and take up the proper role in the new region. There are complex emergent properties here, which your Bose-Einstein condensate lacks. Which is also why it has not practical use beyond theoretical research.

I subscribe to the cellular automaton interpretation of quantum mechanics, that our most fundamental laws work like the game of life. A basic part of their program is to increase entropy. This results in homogenous and lack of complexity. These is where your counter point falls apart. Your Bose-Einstein condensate is less complex. Unless we could change the laws of our universe, we have no way to program its behavior and is therefore not useful. So if you must, I will quality my original assertion that it will be a programmable swarm-like behavior. Although whether it be from computer code or DNA, this is a characteristic than any traditional swarm behavior possesses.