The literature contains very few publications on the application of Process Mining methods for the analysis of event logs in Enterprise Collaboration Systems (ECS). This is not surprising because the analysis of digital support for collaborative work is extremely intricate due to various challenges relating to a lack of data access, poor data quality, unstructured processes and a lack of descriptive models. This article reports on the findings from an Action Design Research (ADR) project. The ADR team had access to a large instance of an operational ECS with more than 3000 users. The event log contains several million entries. Together with the platform’s operating team, intensive research was carried out over a period of six years on ways of analysing user activities on the platform. Several cycles were run to develop new methods and computational techniques to decipher the event logs and meaningfully describe the processes recorded in them. Thanks to the close collaboration between the researchers and the operators of the collaboration platform, it was possible to compare the real-world processes carried out in the platform with the processes discovered using a novel method for Social Process Mining (SPM). The result is a pattern analysis that discovers patterns in processes that have a high degree of correspondence with the real-world scenes of collaborative work. The research work has now reached a point where other software products are included (multi-system analysis) and a catalogue of collaborative work situations (scenes) has been developed to describe the process patterns that result from the Process Mining and graph-based analysis techniques.