Action Design Research (ADR) is a well-known research method within Design Science Research (DSR). An essential characteristic of the ADR method is the need for researcher-practitioner collaboration (RPC). While there is abundant research on RPC regarding information systems projects in general concerning explanatory and normative knowledge, there is very limited prescriptive knowledge on how to execute RPC in ADR projects. Successful collaboration in ADR projects is imperative since the development of socio-technical IT artefacts requires frequent interaction in organisational contexts. However, RPC can be hard to manage due to competing interests. Therefore, the purpose of this paper is to present prescriptive knowledge for how to manage RPC. We have analysed a collaborative ADR project consisting of several researchers and practitioners. Based on a grounded theory approach, we have developed theoretical models based on challenges identified in an ADR project. The models provide prescriptive knowledge regarding: shape the IT artefact based on organisational intervention, exploit the mutual dependency between developing design principles and IT artefacts, and contextualise and generalise learning. Each model involves logical relationships between: conditions for the challenges, actions taken to address the challenges and consequences of the actions taken. The guidelines were deducted from the models and consist of recommendations that could be considered in future ADR projects.