Automating Systems Engineering Workflows in System Composer Using Generative AI

Automotive programs increasingly adopt model-based systems engineering (MBSE) to manage system complexity, yet early architecture modeling is still largely manual. Requirements often arrive as natural language in documents or spreadsheets, while architecture models, interfaces, and tool-ready scripts are produced later by specialists. This gap can lead to slow conversion from requirements to analyzable architectures, inconsistent interfaces, unclear system-of-interest boundaries, and rework as requirements evolve.

Explore a repeatable, automated workflow that uses a generative AI (GenAI) assistant to transform unstructured natural-language requirements into System Composer™ modeling artifacts. Rather than producing only informal diagrams, the approach generates MATLAB^® scripts that create structure and behavior views while maintaining boundary interactions for the system of interest.

The solution integrates TCS Auto AI Brain, an automotive-focused agentic AI framework, with System Composer MCP. Dedicated AI agents extract architectural and behavioral information from requirements and convert it into repeatable MBSE artifacts in System Composer. The workflow supports functional architecture generation, behavioral modeling, and traceability to requirements, while embedded modeling rules help ensure consistent, runnable model creation.

Follow an adaptive headlamp example in the automotive body domain to see how the workflow supports architecture development from requirements through modeling and verification activities. By keeping GenAI execution within a locally deployed TCS Auto AI Brain environment, the approach supports intellectual-property protection and cybersecurity requirements. Expected benefits include faster architecture bootstrapping, improved interface consistency, clearer boundary definition, and greater resilience to changing requirements.