Abstract. Using AgentSpeak-DL, we propose an agent model to autonomously query the web of data to gather additional knowledge related to the current set of beliefs. The knowledge acquisition mechanism uses additional assertions specified with the beliefs as inputs to construct the queries. The assertions come from a description logic approach to specify the agent belief base. Information from the queries is used to establish an ontology relating agent’s beliefs to knowledge from the web of data. Main contribution of this research is the specification of a mechanism to enhance the agent’s knowledge with semantic web techniques.