Historical documents to semantic knowledge models: an AI workflow for industrial heritage

The ongoing digital revolution has profoundly impacted industry and society driving the urgency to reconsider and innovate current industrial heritage recovery and valorization activities. Indeed, the industrial heritage field faces increasing challenges related to the management and interpretation of historical data, much of which is unstructured, dispersed, and difficult to integrate into modern conservation practices, requiring high expertise and manual work for structuring unorganized information into digital knowledge bases and information models. This research explores how the tangible and immaterial information can be processed and integrated into an ontology-based system using an AI Assistant. The aim is to simplify the structuring of historical information through a process of instance generation, allowing the transformation of archival content into formalized and semantically enriched entities —such as machines, production spaces, historical events, and actors—based on a specific information ontology for industrial heritage. This study addresses a critical gap by introducing AI-driven methodologies for semi-automation in heritage practices offering new opportunities for industrial heritage documentation and interpretation.