Statutory article retrieval (SAR) targets retrieval of legislative provisions relevant to a natural language question. The lexical gap between everyday queries and specialized legal language, as well as the structural dependencies of statute law, makes it a challenging task. Here, we introduce JuriFindIT, the first SAR dataset for the Italian legal domain and the first to explicitly encode cross-article references extracted from national legal code. The dataset covers four macro-areas—civil law, criminal law, anti-money laundering and counter-terrorism, and privacy—and includes 895 expert-authored questions and 169,301 generated ones, linked to more than 23,000 statutory articles. We provide retrieval models fine-tuned on JuriFindIT, proposing a pipeline that integrates dense encoders with an heterogeneous legislative graph, achieving consistent improvements over prior SAR approaches.

Grounded legal question answering requires retrieving the relevant statutory provisions, generating legally sound answers, and supporting them with explicit evidence. We introduce JustifIT-QA, the first Italian statutory QA dataset with expert-written answers and fine-grained evidence spans, and present DARv2, a dense retriever that improves coverage of relevant legal texts. Building on this retrieval layer, we study a judge-based feedback loop in which an LLM critiques and refines generated answers, and we evaluate LLMs as judges for legal answer correctness. Results show that feedback improves expert-rated answer quality when relevant context is retrieved, while evidence extraction benefits from stronger models but remains challenging. We release the dataset, code, and retriever to support future research on grounded and explainable legal QA in Italian.