The term “process intelligence” covers all aspects of smart digitization and automation of discrete processes in medicine and biotechnology, i.e., administrative, technical and scientific processes that consist of individual steps, which are usually described in the form of flowcharts, source code, protocols, standard operating procedures, or clinical guidelines.
From a technical point of view, “process intelligence” includes the digitization of processes, i.e., the manual translation of process descriptions from an analog, often paper-based form into a machine-usable, logical and semantically consistent form based on standard formal notations. Such processes are implemented, executed, controlled, documented and subsequently evaluated in software – either directly in a programming language as an executable program (e.g., for controlling a device) or indirectly through a process management system.
From the analytical point of view, process intelligence includes the development of automatable methods for the evaluation of process data and their implementation in software. Such methods are used to retrospectively document errors, track errors, and derive their dependencies, causes, and costs. Furthermore, process mining can automatically compare real processes with their original definition and filter out differences and deviations to identify changes to processes and derive improvements. On the cognitive side, we use the term “process intelligence” to describe approaches that make it possible to automatically translate process descriptions from natural language texts into technical process descriptions (natural language processing), to plan processes automatically (AI planning) and to develop time predictions about process sequences under consideration of complex boundary conditions (scheduling). In doing so, we draw on state of the art in science and technology of machine learning and artificial intelligence.