||The explosive growth of information sources (databases, knowledge bases, and applications providing structured and/or semistructured data) has created a big need for new and efficient generation of tools and techniques that support automated and intelligent data integration. The goal of a data integration system is to provide users with a unified view to transparently and efficiently access, relate, and combine data stored in multiple, autonomous, and possibly heterogeneous information sources. In this paper, we show how software agents, can be used as an automated method for querying multiple information sources. They act either autonomously based on their intelligence or are controlled by human principals. Information agents can be created to represent a particular application domain and information sources can be mapped as views over this application domain. We present an algorithm of how user queries over a specific application domain can be reformulated and executed with the help of the information mediation agents.