Social learning in general can be understood as any learning resulting from the observation of the behaviour of other animals. Corvids present a well-suited model for investigations of social learning because they exhibit the socio-ecological traits that promote advanced social intelligence in other species and therefore are likely to attend to the actions of their conspecifics. Among corvids, New Caledonian (NC) crows (Corvus moneduloides) are known for proficient stick and hook tool use abilities and reports of cultural variation of their tool manufacture techniques in the wild suggest that they may socially learn aspects of this tool manufacture. So far all of the experiments investigating social learning of tool use in NC crows have used sticks but no other types of tools. In our study we investigated the acquisition of a novel non-stick tool use skill in captive NC crows by presenting them with a conspecific solving a task by dropping a stone inside a vertical tube. Here we show that NC crows are capable of socially learning how to use novel tools. In addition, our study provides evidence that NC crows are able to causally intervene (learn a cause-effect relationship by using only observed correlations and to act to benefit from this cause-effect relationship) in solving a problem task, a skill that was recently suggested they do not possess. Furthermore, it reveals that direct demonstrations do not facilitate social learning in crows, which validates the use of non-direct protocols in social learning studies for this species. We hope that our work and the new experimental design will be a starting point for more studies investigating the ability of NC crows to socially learn and solve non-stick or no tool use problem tasks. Such a new approach would shed more light on how flexible and proficient NC crows are in their social learning and tool use skills.