Abstract
Automotive OEMs and Tier 1s face relentless pressure to deliver embedded software faster and at lower cost, without compromising the rigor demanded by Automotive SPICE, ISO 26262, and ISO/SAE 21434. This talk examines how AI and, increasingly, agentic AI are reshaping embedded systems and software development across every phase of the V-model, from requirements engineering and architecture through implementation, verification, and validation.
We will explore concrete, emerging applications: AI-assisted requirements analysis and traceability, automated work-product review and consistency checking, test-case generation, and defect prediction — capabilities that turn quality assurance from a reactive, documentation-heavy activity into a continuous, predictive one. Rather than positioning AI as a shortcut around process, we frame agentic AI as a quality multiplier that strengthens ASPICE compliance and functional-safety evidence while compressing cycle time and cost.
Attendees will leave with a clear view of the current state of the art, the practical opportunities and limitations (including trustworthiness, explainability, and assessor acceptance), and a pragmatic perspective on where agentic AI can deliver measurable improvements in development efficiency and product quality today.