Leveraging knowledge graphs in facilities management: A scoping review

Background: Knowledge graphs (KGs) and Semantic web technologies (SWT) are increasingly explored in facilities management (FM) to address persistent problems of interoperability, fragmented data, and limited automation across operational systems.

Methods: A scoping review was conducted in accordance with PRISMA-ScR 2020 to map KG applications in FM and identify emerging research directions. Web of Science and Scopus were queried for English-language, peer-reviewed studies applying ontologies, SWT, linked data, or KGs at building-level FM. Studies focused on non-FM contexts, earlier lifecycle phases without operation and maintenance relevance, or artificial intelligence and machine learning approaches without explicit SWT/KG integration were excluded. Following de-duplication and qualitative screening, 48 studies were thematically analysed.

Results: Four dominant application areas were identified: (1) ontology-driven semantic interoperability (n = 36), (2) systems integration across FM data sources (n = 24), (3) data and information retrieval from graph-based stores (n = 14), and (4) automated compliance and validation through knowledge-based reasoning (n = 6). Most studies employed modular, domain-specific ontologies deployed on graph databases to link static building information modelling and asset data with near real-time streams. Emerging directions include the use of KGs as semantic integration layers in digital twin (DT) architectures, and KG-grounded retrieval-augmented generation to enhance trustworthy and explainable access to FM information.

Conclusion: This review consolidates dispersed research on KG applications in FM and outlines a structured research agenda spanning modular ontology development, integration with DT and compliance workflows, as well as longer-term opportunities for scalable and auditable generative AI applications. Findings reflect peer-reviewed journal literature and may under-represent emerging industry practice.