AFAIK Wayland is "simply" a compositor protocol.
appman is an application that uses that protocol - so the notion of comparing the two for me is not really clear.
If you read carefully in the link you posted:
Having made the decision to have multiple process for the user interface display, the next technical problem to solve is getting the outputs of each process onto the display. Historically the approach has been to use X and write your own window manager.
Also, the text does not explain "how to do with wayland" it says wayland is the better way to go for the reasons it lists.
This text says nothing on implementation, and, appman is basically the implementation of the text you posted.

What are the cost/benefits of using appman as opposed to QtWayland to create a multi-process UI?
Again, using appman is not exclusive to using wayland.
Appman is using wayland.