Process support environments (PSEs) are widely used for modelling, enacting and analyzing human intensive processes. The benefits of a PSE become apparent when processes to be supported are long lived and distributed and contain heterogeneous components. Generally, such processes are subject to dynamic evolution, i.e. they have to be changed during their execution. Unfortunately, virtually none of the existing PSEs consider dynamic evolution of processes. This article explains the concepts and techniques underlying a set of components developed in the ESPRIT Project Process Instance Evolution (PIE) that support the dynamic evolution of processes. These concepts and techniques are demonstrated using a real-world scenario from the automotive industry.