The paper reports on a multi-layered corpus of Italian, annotated at the syntactic and lexicosemantic levels, whose development is supported by a dedicated software augmented with an intelligent interface. The issue of evaluating this type of resource is also addressed.