Flexible process support by automatic aggregation of implicit and explicit user behavior

A company's ability to flexibly adapt to changing business requirements is one key factor to remain competitive. The required flexibility in people driven processes is usually achieved through ad-hoc workflows which are naturally highly unstructured. While most process support systems have either high process support or high flexibility but seldom both, the approach presented in this paper eliminates this trade-off by mapping unstructured messages onto processes as well as patterns for self-adjusting and context-sensitive process recommendations based on the analysis of user behavior, crowd processes, and continuous application of process detection. Special focus within the paper lies on the ability of a user to interact with the system during run-time in case of a mismatching of an event and the corresponding self-adapting features. The approach is evaluated in the context of the European research project Commius.