Action research is widespread in many of the background disciplines that underpin the e-Government field andis beginning to take root as a legitimate e-Government research method. Canonical Action Research (CAR)is the most widely used form of action research; however it relies on premises that can be problematic in thee-Government context. This article details some of those underlying assumptions, and shows the difficulties that result when applied to a relatively typical e-Government case study: the implementation of an advanced email system in Danish public administration. The empirical experience calls many of the standard premises into question, and these are categorised. The authors summarize the resulting experience as nine challenges for action researchers working in the e-Government field, and investigate some possible responses.