Our Mobility Center of Excellence (CoE) engineers developed a Multi-Agent AI Copilot platform that automates the transformation of OEM specifications into ASPICE-compliant System (SYS.1) and Software (SWE.1) requirements.
The solution leverages a dual-agent architecture consisting of a System Architect Agent and a Software Engineering Agent to automate the transition from unstructured customer specifications to structured engineering requirements. By separating system-level and software-level engineering activities, the platform improves the quality of requirements, establishes technical guardrails, and enables end-to-end traceability throughout the development lifecycle.
Business Objectives
- Accelerate specification-to-requirements transformation and reduce engineering lead times from weeks to hours
- Automate generation of ASPICE-compliant SYS.1 and SWE.1 requirements from unstructured OEM specifications
- Establish automated bidirectional traceability between source specifications, system requirements, and software requirements
- Prevent requirement inconsistencies through engineering guardrails aligned with vehicle constraints and safety requirements
- Reduce manual documentation and traceability effort for engineering teams
- Improve requirements quality through automated validation and conflict detection
- Support ASPICE and ISO 26262-aligned engineering processes
- Enable scalable and repeatable requirements engineering workflows across vehicle programs
Solution
The solution follows a V-Model-aligned approach, where specialized AI agents progressively transform engineering intent into standardized system and software requirements. Built-in validation mechanisms continuously verify alignment between system-level constraints and software-level requirements, ensuring consistency, traceability, and compliance throughout the requirements engineering process.
Key Highlights
1. Dual-Agent Requirements Engineering Architecture
Developed a specialized AI framework consisting of a System Architect Agent and a Software Engineering Agent, enabling structured transformation of OEM specifications into ASPICE-compliant system and software requirements.
2.Automated Specification-to-Requirements Transformation
Enabled intelligent processing of unstructured customer specifications and automated generation of SYS.1 and SWE.1 requirements, significantly reducing manual engineering effort and accelerating requirements development.
3.V-Model Aligned Engineering Workflow
Structured the requirements engineering process to mirror the Automotive V-Model, ensuring system-level requirements are established before software requirements are derived, improving consistency and engineering accuracy.
4.Technical Guardrails and Validation
Implemented built-in validation mechanisms that verify software requirements against approved system constraints, safety objectives, and engineering boundaries, helping prevent inconsistencies and requirement conflicts.
5.ASPICE-Compliant Requirements Engineering
Built a standardized requirements engineering workflow aligned with ASPICE practices, enabling repeatable, scalable, and compliance-ready requirement generation.
6.End-to-End Traceability Automation
Established automated bidirectional traceability between OEM specifications, system requirements, and software requirements, simplifying compliance verification, impact analysis, and audit readiness.
7.Engineering Productivity Enhancement
Reduced manual documentation, requirement mapping, and traceability activities, allowing engineering teams to focus on validation, refinement, and decision-making.
8.Scalable AI-Powered Platform
Developed a reusable and configurable AI-powered platform capable of supporting multiple vehicle programs and future software-defined vehicle initiatives.
Outcomes
- 90 % reduction in specification-to-software requirements transformation time
- Delivered automated bidirectional traceability between OEM specifications, SYS.1 requirements, and SWE.1 requirements
- Improved requirements quality through automated validation and conflict detection
- Reduced manual traceability and documentation effort across engineering teams
- Identified engineering ambiguities and requirement inconsistencies earlier in the development lifecycle
- Strengthened ASPICE compliance through standardized requirements generation and traceability management
- Enabled engineers to focus on review, validation, and approval activities rather than manual documentation tasks
- Established a scalable AI-powered foundation for future automotive software development programs
Technologies Used
- AI Platform: Multi-Agent AI Copilot Framework
- Large Language Models: Azure OpenAI Service (GPT-4o)
- Requirements Engineering Standards: ASPICE SYS.1 & SWE.1
- Functional Safety Alignment: ISO 26262
- Architecture: Dual-Agent V-Model Requirements Engineering Workflow
- Core Capabilities: Requirement Generation, Traceability Management, Validation, Conflict Detection
- Deployment Environment: Secure Enterprise Cloud Infrastructure
- Integration: Automotive Requirements Management and Engineering Toolchains
