For seven, they could be talking about building a service to handle a new document, based on unit components it contains. As an example, I would look at it the way we use WM IS to do segment processing. Well for a new document, if it has elements compatible with existing services on a system but not assembled in a static handler, an auto build of a new processor can happen to work on the data. That would allow for a tight processor that doesn’t have to worry about trying to loop through unused elements and you only have to make the initial segment handlers for updating processing tasks.
That then would be a Document-Event Driven Architecture. I am not sure how easily we can take advantage of this from a Field-Event Driven Architecture. Maybe something along the lines of parsing a simple string to make the processor, like “check inventory of item x and reorder” having three units. But that maybe be better built and faster processing from a statically existing service. Just some thoughts. Good day.