A goal is to combine process modeling with question answering and discovery.
With a qualitative model of some process, one can follow the process by monitoring data flows; if one is at some location
in the model (as a graph, being on a particular node which describes the current situation), the model provides predictions of what the process can go next. If the data fails to described an expected next situation, one experiences an expectation failure
. Expectation failures are the stuff of discovery.