This presentation introduces an innovative approach to solving complex safety engineering challenges—such as the retroactive separation of safety and non-safety software in legacy systems—using AI-assisted systems engineering. A central orchestrator acts as a digital project manager, decomposing complex tasks into verifiable sub-steps and providing specialized AI agents with tailored work products to prevent cognitive overload and hallucinations. By integrating external validation tools via the Model Context Protocol (MCP) and automatically generating evidence-driven artifacts (rationale), the system creates a transparent and traceable development chain. The AI serves in a primary assistance role by visualizing architectural decisions, thereby enabling targeted "Human Gates" for expert review and decision-making. This process ensures that human expertise is empowered by transparent data models while the AI significantly enhances the efficiency of producing standards-compliant safety artifacts.