The purpose of this article is two-fold: (i) to present RAPOSA, an open-domain automatic question answering system for portuguese that participated in QA track at CLEF 2006 for the first time, and (ii) to explain how RAPOSA is intended to be a key component of a larger information extraction framework that uses question-answering technology as the basis for automatic generation of biographies. We will make a first attempt to classify questions regarding their relevance for iterative biography generation and, according to such classification, we will then explain our motivation for participating in CLEF 2006. We will describe the architecture of RAPOSA, explaining the internal details of each of its composing modules. Next, we will present the results RAPOSA obtained on this year’s edition of the QA track, and we will identify and comment on the main causes of error. Finally, we will present directions for future work.