In this work we introduce the knowledge acquisition procedure supported in the KITE.it process management framework. We then illustrate how Process Mining techniques are supported in this framework proposing a running example featured in a Data Loss scenario. We then provide some lesson learned that can be generalized in similar contexts. In particular, we show how applying an appropriate Semantic Lifting to the log may help to discover behavioral patterns of the process that is actually being executed. Our conclusions spotlight that it is viable to verify whether some non-functional properties hold during the process execution. Moreover, we describe the impact that Semantics Lifting has on support and confidence of the inferred probabilities of observing these behavioral patterns.