This paper presents a scalable approach to the identification, extraction, and modeling of administrative procedures within the Italian public sector. Despite ongoing reform efforts, Italy lacks an official, comprehensive catalogue of such procedures—an absence that hinders administrative simplification and digital transformation. To address this gap, we propose (i) the development of AdmPModeler, a pipeline that uses Large Language Models (LLMs) and prompt engineering techniques to convert legal texts into structured, human-readable process models, and (ii) a nationwide web crawling effort targeting the public administration websites, where procedural documentation is legally mandated to be published. This work lays the foundation for a national repository of administrative procedures—that support transparency, accountability, but also enable advanced policy experimentation, simplification initiatives, and automated compliance verification.