The digital transition of built heritage has converged heavily on HBIM, whose hierarchical data models now strain under the demands of cross-domain decision-making in digital building management. This volume advances an alternative framework, Buildings as Networks, which conceptualises buildings as graph-based knowledge systems. Graph representations connect heterogeneous building data more naturally than tabular relational models, forming broader interoperable superlogical structures. They also provide the semantic scaffolding that AI systems need to ground probabilistic behaviour in domain knowledge. In light of the epistemological revolution brought by AI, the book argues that rethinking heritage data architecture is a precondition for machine intelligence in the built environment.