We present an initial investigation from a semi-experimental setting, in which
an HMD-based AR-system has been used for real-time collaboration in a task-oriented scenario (design of a museum exhibition). Analysis points out the specific conditions of interacting in an AR environment and focuses on one particular practical problem for the participants in coordinating their interaction: how to establish joint attention towards the same object or referent. Analysis allows insights into how the pair of users begins to
familarize with the environment, the limitations and opportunities of the setting and how they establish new routines for e.g. solving the ʻjoint attentionʼ-problem.