A software agent helps people filter, sort, and prioritize messages that were already addressed to them, and it also helps them find useful messages they would not otherwise have received. In some cases, it responds automatically to certain messages, and in other cases it suggests likely actions for human users to take. In this paper we focus our attention on such type of software agents, and introduce NetFinder, an intelligent agent that supports users for information gathering in general and news filtering from on-line newsgroups and personal emails in particular. In the following we briefly introduce our agent system, the basic idea of its design, the learning algorithm it adapted, and the structure of the system.