Guidelines for manual process discovery for semi-automated processes

Abstract Semi-automated operational environments, where human operators and automated systems coexist, pose significant challenges for process discovery approaches, which typically target either fully manual or fully automated workflows. This paper presents a comprehensive guideline for manual process discovery specifically tailored to semi-automated contexts. Building on established methods (document analysis, direct observation, structured interviews, and collaborative workshops), we adapt and extend each method to capture both human-driven activities and machine-generated events. The resulting artifact comprises four modular guideline packages, offering step-by-step procedures, guiding questions, and optional tooling to uncover control flow, data flow, resource assignments, and system interactions. Use cases in an Industrial Internet of Things (IIoT) setting illustrate how these guidelines can be applied to complex, data-intensive scenarios, though the approach remains domain-agnostic and can be transferred to other semi-automated environments. The development followed the Design Science Research (DSR) cycle, incorporating iterative refinement and validation via expert workshops, surveys, and real-world use cases. Our evaluation confirms that the proposed guidelines enhance the completeness, accuracy, and efficiency of manual process discovery in hybrid human–machine landscapes.