The IST-funded European project OmniPaper is investigating techniques to obtain a novel online news experience. These include XML- and Artificial Intelligence related technologies. The OmniPaper architecture starts from distributed news archives, all within different operating environments, database formats and indexing mechanisms. SOAP (Simple Object Access Protocol) is used to create a uniform access method to these archives. Rich indexing and meta-data structures, such as Topic Maps, make intelligent search possible. A cross-archive intelligent index (or ‘knowledge layer’) contains concepts, relationships between them and occurrences (articles) in different languages. The actual OmniPaper prototype gives the user a graphical relational view on concepts relevant to the current query. The relation view is a representation of a small piece of the web of concepts. Using this highly interactive view, the user can refocus his query using different word senses of the query keywords. This allows the user to redefine/refine his query if one of the query keywords has different possible meanings (e.g. bank as a financial institution or as a piece of furniture). The relational view allows the user to navigate between different concepts relevant to the query.