Cognitive automation moves beyond rule-based automation and thus imposes novel challenges on organizations when assessing the automation potential of use cases. Thus, we present an empirically grounded and conceptually operationalized model for assessing cognitive automation use cases, which consists of four assessment dimensions: data, cognition, relationship, and transparency requirements. We apply the model in a real-world organizational context in the course of an action research project at the customer service department of ManuFact AG, and present unique empirical insights as well as the impact the application of the model had on the organization. The model shall help practitioners to make more informed decisions on selecting use cases for cognitive automation and to plan respective endeavors. For research, the identified factors affecting the suitability of a use case for cognitive automation shall deepen our understanding of cognitive automation in particular, and AI as the driving force behind cognitive automation in general.