Text Mining tackles the task of searching useful knowledge (patterns) in a natural language document. Given the cost of a (full) morpho-syntactic analysis of a textual database, specially when the linguistic rules are not respected, most text mining techniques process without using the linguistic structure of those documents. In this Information Extraction framework, Grammatical Inference techniques can be used to extract the structure of a text (or of some of its sublanguage). This will allow an informed research of useful information in the textual data bases. In this paper, we present the contribution of the Grammatical Inference in the Text Mining field by reporting an Information Extraction process we applied to a seminar announcement corpus.