When modelling across lots of different processes there can be large amount of duplication where tasks, events etc appear across many of them where those artifacts are actually refer to exactly the same thing across all those processes.
What would be useful is if we could have a difference between copy and duplicate, where duplicate would be an exact clone of that task, event etc. where only one exists across all processes using it.
This begins to clean up the models but is also provides some great impact analysis capabilities because you can begin to find all processes that carry out particular tasks etc.
It’s a really good question and possibly you are right - My thinking was you would have a separate palette of existing reusable artifacts you could drag in (alongside the existing shape palette).
How would you see this type of thing being achieved if the modeler doesn’t play a part in it?