From Documents to a Living Product Model: How Systems Engineering Artefacts Become the Foundation for AI-Augmented, Continuous Engineering in Safety-Critical Products
Complex products such as tractors, aircraft, and medical robots are increasingly defined by software, adaptive functions, and AI components, introducing non-deterministic behaviour that traditional, document-centric engineering workflows can no longer control. Systems Engineering provides the right mental model — requirements, functions, architecture, and verification — but this knowledge remains fragmented across tools and specifications, inaccessible to both engineers and AI agents.
The decisive enabler for AI-augmented engineering is the deliberate transformation of Systems Engineering information into a structured, ontology-based product knowledge graph connecting every artefact from use case to code. Specialised AI agents — for architecture challenge, requirement consistency, test derivation, and continuously regenerated Design-FMEA — can then operate on a shared, traceable product model rather than isolated documents.
Using TRIAS, an AI-FMEA agent deployed by the authors in industrial projects, the paper illustrates how continuous, knowledge-graph-driven risk analysis challenges architectural decisions and exposes robustness gaps throughout development. Context engineering and SE validation gates ensure AI contributions remain auditable and aligned with ISO 25119, ARP4761, and ISO 14971.
Well-structured SE artefacts are the semantic foundation that makes AI agents trustworthy — and documentation becomes an automatic output of a well-maintained product model, not a manual burden.
Alexander Krumm holds an M.Sc. in MBSE and Aircraft Systems Engineering from Technische Universität Hamburg. He began his career at Continental as System Lead Engineer for Embedded Systems in Autonomous Driving Vehicles, before moving to CLAAS where he served as System Architect for Global Electric and Electronic Systems. He subsequently joined CARIAD (Volkswagen Group) as Senior System Architect, leading cluster-specific drive system and body system architecture. Since June 2025, he is CEO and Co-Founder of VPATH AI GmbH, developing AI-driven engineering intelligence tools for safety-critical and model-based industries.
Dr. Frank Schouren is an independent consultant and founder of Schouren Consulting, with over 20 years of experience leading transformation programmes in industrial R&D organisations. His expertise spans Systems Engineering, quality management, AI agent integration in product development, and (A)SPICE. He has held programme management roles at major German automotive OEMs and tier-1 suppliers, and consults on risk analysis methods including FMEA. Dr. Schouren co-develops and operationalises AI-based engineering solutions — including TRIAS — at industrial customers alongside VPATH AI GmbH.