| Global regulatory compliance for hardware and software is increasingly untenable due to overlapping jurisdictional mandates, such as the European Union (EU)’s New Legislative Framework, Artificial Intelligence (AI) Act, and Cyber Resilience Act (CRA). Manual alignment of legal requirements to product specifications via traceability matrices and EU Declarations of Conformity (DoC) is labor-intensive, error-prone, and unscalable. We present a context-aware Retrieval-Augmented Generation system designed to automate this workflow. By indexing regulatory corpora and utilizing domain-specific prompting, the system generates traceability matrices and draft DoCs grounded in verifiable legal citations. In a case study involving an AI medical device, the platform reduced compliance effort by 63% and increased document quality by 20% against expert benchmarks. Notably, the system identifies cross-regulation dependencies, such as CRA requirements for products outside the primary scope of the AI Act, to ensure comprehensive certification strategies. This framework offers a scalable solution for intersecting legal mandates across global jurisdictions. |
*** Title, author list and abstract as submitted during Camera-Ready version delivery. Small changes that may have occurred during processing by Springer may not appear in this window.