- First you create an overview diagram where you add a set of decisions, perhaps all the decisions that need to appear in a set of business processes, along with the most critical input data and knowledge sources for those decisions.
- Each of these decisions is going to be decomposed and doing all this on one diagram can get complex. So, instead, you create a diagram for each such decision (or perhaps for a small but closely related set). Having created a new diagram you can use the search in the palette to find the decisions, input data and knowledge sources you already created and simply drag them onto the diagram to associate them. As you add them to the diagram existing connections are automatically added.
- You can now drill into these decisions, adding more detail, without cluttering up the overview diagram. You can repeat this process to build diagrams for specific parts of the decision model (perhaps a part of the model owned by a particular group), to support a particular review meeting or for any other modeling purpose.
- Every time you do this you associate the same underlying objects with the new diagrams. And when you open these underlying objects to edit them you will see all the requirements defined for the object on every diagram- a complete context view.
Here is a short video example: