Until recently, it was assumed that co-representation of others’ actions, an essential part in joint action, is biologically tuned. However, research demonstrated that we also simulate actions of non-biological interaction partners under certain conditions. In the present study, we investigated whether perceived intentionality or perspective taking is the underlying mechanisms of this phenomenon. Participants saw a short video fragment of a non-biological agent (i.e. a triangle) as main character. The movements of this agent were either described as intentional or as unintentional. Furthermore, participants were instructed to either take the perspective of this non-biological agent or not. Results show that perspective taking and perceived intentionality both lead to action co-representation of non-biological actions. Possible explanations for these findings are discussed.