The volume of information available on the Web is constantly growing. Due to this situation, users looking for documents relevant to their interests need to identify them among all the available ones. Intelligent agents have become a solution to assist users in this task since they can retrieve, filter and organize information on behalf of their users. In this paper we present two experiences in the development of interface agents assisting users in Web-based tasks: PersonalSearcher, a personalized Web searcher, and NewsAgent, a personalized digital newspaper generator. The main challenge we faced to personalize the tasks carried out by these agents was learning and modeling specific and dynamic user interests. Our proposed approach consists of incrementally building a hierarchy of users’ relevant topics and adapting it as agents interact with users over time.