This paper presents the results of the project REVER (Reverse Regesta), which investigates new paths for the study of papal documents by combining the methodological framework of papal diplomatics with artificial intelligence. Although regesta have long played a central role in the organisation and study of papal sources, the production of new collections remains a highly demanding scholarly task. REVER addresses this challenge by developing workflows, resources and tools that support the creation and structuring of regesta within a digital research environment. The project is grounded in a critical analysis and valorisation of the editorial tradition of the large documentary corpora produced by the medieval papal curia. Drawing on established historical source collections, the project produced two main outcomes: annotated datasets and the development of Regexta, a digital tool for the study of pontifical documentation. A curated training corpus was created by pairing medieval papal documents with their corresponding regesta and digitised and annotated through customised implementations of eScriptorium and INCEpTION, combining manual and automated strategies while preserving key diplomatic features. Building on this corpus, Regexta was developed as a multilingual, multi-stage tool for the summarisation of pontifical documents. By generating new, structured, and semantically consistent regesta in accordance with established diplomatic standards applied to medieval documents, the tool enables the large-scale organisation, exploration, and analysis of papal documentary corpora, including digital and born-digital materials.