Enhancing Digital Product Passports for the Circular Economy with Generative AI

The transition from a linear “take-make-dispose” model to a Circular Economy (CE) is central to the European Green Deal. Achieving CE goals requires transparent, reliable access to product information across all life cycle phases, so stakeholders can make informed decisions. The European Union’s Digital Product Passport (DPP) addresses current gaps by standardizing life cycle data exchange. Yet, efficient information retrieval from DPPs remains challenging due to high implementation effort, limited semantic richness and differing interpretations across systems. This paper proposes a Generative AI-based approach that enhances an existing Asset Administration Shell (AAS)-based DPP with graph-structured Retrieval-Augmented Generation (GraphRAG) on a Neo4j graph database and the Model Context Protocol (MCP). Two interaction paths for end users are implemented: (i) a pipeline that converts natural-language questions to Cypher queries and (ii) an MCP-based variant that leverages schema-aware tools to validate queries against complex AAS schemas. A use case from a dynamic production environment is employed to demonstrate and evaluate the approach. Results provide initial proof that integrating an AAS-based Neo4j graph database with MCP improves transparency and clarity, reduces retrieval time, and supports product-related decision-making during the use phase of a product. The key contribution of this study is a novel approach that leverages Generative AI solutions to not only enhance data accessibility for the product’s non-technical users, but also advances CE objectives through more efficient and effective DPP utilization.