This paper discusses how research with practitioners can help reconcile the top‐down requirements of headquarters with the bottom‐up local needs in the context of global information systems. Based on a 12‐month canonical action research project that took place at the Chinese branches of a French multinational corporation, our research revealed and addressed workarounds that the Chinese users of a company‐wide global enterprise resource planning system had put in place that were not expected nor desired by company headquarters. From the local users' point of view, they were necessary to deal with Chinese legislation and cultural practices, but from the French headquarters' point of view, they meant that many of the potential gains of global standards were lost. Activity theory was used as a focal theory to analyse each of these workarounds and business process management as an instrumental theory to design solutions to the workarounds. We describe in detail how we used canonical action research to successfully deal with exemplars of each of the three types of workaround identified (data adjustments, process adjustments and parallel‐system adjustments). Unusually, the research relates to post‐implementation change rather than to that looking at change occurring before and during implementation. We argue that canonical action research and the particular combination of activity theory and business process management are appropriate for dealing with workarounds and this has not been demonstrated previously. Further, our research – deemed successful by managers, users and researchers alike – took place in China where previous literature suggests only limited success with such global systems.