Assembly processes are a particular class of production processes that have gained increasing importance as to their impact on flexibility and efficiency of the overall process. Assembly processes are characterized by high variability concerning the type and logical flow of tasks to be performed, the devices and tools used, the materials and information to be processed. Due to the complexity of such processes a relatively high ratio of tasks is still performed by human workers. Keeping track of materials, devices and tools used and at the same time providing digital assistance for increasingly complex assembly tasks requires adequate modeling techniques as a prerequisite for respective systems design. In this paper we discuss challenges for modeling assembly processes based on a real-world assembly line from our learning factory lab. Finally, we propose an architecture for a process-driven assistive assembly system which shows the interplay between process based monitoring, task assistance and workplace sensors.