The aviation design process, structured around the V-model, systematically guides the development of complex systems like aircraft through phases such as requirements definition, architecture design, implementation, integration, and verification. Automation is increasingly vital for meeting demanding certification standards, with advances in modeling, simulation, code generation, and test automation improving both efficiency and reliability. Artificial intelligence (AI) and machine learning drive further enhancements in design optimization, compliance verification, and error detection. Specifically, Retrieval-Augmented Generation (RAG) models show promise by enabling efficient knowledge extraction from technical documents, regulatory guidelines, and test reports. Despite these innovations, the adoption of automation and AI is tightly constrained by rigorous aerospace certification standards such as DO-178C and ARP 4754A, which impose strict requirements on software implementation and verification. Additional standards like DO-330 introduce further complexity, particularly regarding the qualification of process automation tools, especially those based on non-deterministic systems such as Generative AI. This paper investigates the impact and role of Generative AI-driven tools, including RAG models, in automating systems engineering tasks. It provides a phase-by-phase analysis following the V-model, spanning requirements, implementation, integration, and testing, and examines the challenges of aligning Generative AI tools with existing qualification and certification standards.
Klein et al. (2026) studied this question.