Education is one of the most interesting applications of virtual worlds, as their flexibility can be exploited in order to create heterogeneous groups from all over the world who can collaborate synchronously in different virtual spaces. Additionally, they can create opportunities to offer educative contents with the advantages of online courses, but also the feel of “presence” that only virtual worlds can provide. However, the interaction in social virtual worlds usually takes place in text mode, given that usually only textual and chat facilities are provided to communicate with avatars and bots. This makes it difficult for people without enough technical knowledge, or with visual or motor disabilities, to interact with such systems. To solve this problem, the authors propose a methodology for building bots as intelligent embodied conversational agents that can communicate with the users through oral as well as visual modalities.