If the controller is typical with the big fat dropping resistor in the LVPS section, it will generate at least a few watts of heat right there.
The FETs will also create significant waste heat from both switching and resistive losses. Let's just give a round number of perhaps 5W of heat from each phase bridge, so call it 15W of heat. (actual losses depend on FETs used (RDSon, how they're driven by the controller, etc), system voltage, actual usage at any moment, current flow, etc)
So you could have 20W or more of heat generated within the controller during use.
To give some comparison, a soldering iron only needs around 15w to be effective at melting the metal solder.
(when in contact with it, to be sure, but if you were to enclose the iron in a small box the box and anything in it would heat up fairly rapidly)
So putting the controller in an enclosure of it's own will trap all that heat within the controller, which will heat the parts within it more than they are intended to be, and age them more quickly than if it were out in the airflow as it is designed to be (there is a reason the casing is a heatsink; but the heatsink can't get rid of heat if there is nowhere for it to go).
If it's in an enclosure with something else, the something else will also be heated; this will help the controller for a while (until everything reaches equilibrium), but it may not be helpful to the something else, depending on how much heat is generated, what that equilibrium temperature is, and what the something else is and how well it can take that heat.
If the controller enclosure is itself a heatsink that does have outside air exposure, and the controller is bolted to it such that it passes as much of it's heat to it as rapidly as possible, it will help mitigate the above problems.
If your controller doesn't generate that much heat, or isn't used very hard or very much, and temperatures are low enough, it might not be an issue.