Semantic and LLM-Enhanced Data Integration for FAIR-Compliant B2B and B2G Ecosystems

In today’s data-driven environment, organizations increasingly depend on high-quality, interoperable data. This research presents a scalable, FAIR-compliant (Findable, Accessible, Interoperable, Reusable) data integration and governance framework for business-to-business (B2B) and business-to-government (B2G) contexts, addressing both metadata and data interoperability. The solution combines ETL pipelines, Semantic Web technologies, and Large Language Models (LLMs). Metadata interoperability is achieved via RDF schemas, SKOS vocabularies, and linked data principles; data-level integration is supported through schema mapping, format conversion, and LLM-assisted transformation. A dynamic ETL process uses semantic schema mapping and LLM-driven vocabulary suggestions to reduce manual alignment with domain ontologies. The framework is validated through a formal evaluation strategy combining real-world B2B and B2G case studies with comparative analysis against existing solutions. The system’s applicability spans across critical B2B and B2G scenarios in domains such as finance, healthcare, public administration and logistics, illustrating its capacity to enable trustworthy data sharing and governance across institutional and organizational boundaries.