It is more about the workflow, than anything else. I would agree that it is not necessarily intuitive.
If you only want to have a single VLAN you can just do vlan <vlan name> <vlan number>. If you have multiple VLANs attached to a VLAN name, you must designate it as a pool up front on the local controller, which is pretty much where it is interpreted. There are things that you can only do with a single VLAN name, such as a server derivation rule, that you cannot do with a pool, and this helps the controller to limit commands to only things that it supports.
In the user guide, please search for the section "Creating, Updating and Deleting VLAN Pools" which will give you the accurate workflow for both the GUI, as well as the commandline.