From BIM to Agentic AI: predictive management systems and phygital digital twins for built heritage

Information flow management is the backbone of digital transformation in the construction sector: the Data Sharing Environment (ACDat) enables integration and full interoperability, extending the potential of Building Information Modeling (BIM) and promoting sustainability in the management of the building life cycle. At the same time, the emergence of Agentic AI introduces a cognitive shift: autonomous, adaptive, and proactive agents take on design, cognitive, and institutional roles, helping redefine and even anticipate decision-making processes, organizational models, and market dynamics. In this context, the article presents the e-Building platform currently being tested, which combines ACDat, Agentic AI, and digital twins within a phygital logic. The platform automates document classification and building information handover through natural language and information constantly updated by the digital twin, with the result of evolving ACDat from a management infrastructure to a true cognitive ecosystem. The e-Building pilot (TRL 3–4) demonstrates that integrating Agentic AI with ACDat transforms predictive management from a technical tool to a paradigm of anticipatory governance