Abstract
Automotive software development doesn’t suffer from too much engineering discipline, it suffers from disciplined engineering that has been implemented poorly and fragmented across disconnected tools and processes. This keynote challenges a common industry perception that equates ASPICE with process, compliance, and bureaucracy, arguing instead that the real barrier to speed is fragmentation, not the standard itself. The talk reframes the conversation away from ASPICE (and any single framework) toward the broader concept of “Engineering Excellence,” defined as the ability to consistently transform intent into high-quality products through disciplined, connected, and continuously improving engineering; a capability that standards like ASPICE, ISO 26262, Agile, DevOps, and Lean can inform, but none of which singularly embody.
Central to the talk is the idea of an “Engineering Operating System” (EOS): a unified, interconnected system in which requirements, architecture, verification, defect analysis, and institutional knowledge reinforce one another rather than existing as isolated silos. The presentation argues that while tools and methodologies change constantly (waterfall to Agile, on-prem to cloud, manual integration to CI/CD, and now AI-driven development), the underlying engineering principles, what are we building, why, how do we know it works, what happens when it fails, and what did we learn, remain timeless.
The keynote then pivots to artificial intelligence, asserting that the industry is asking the wrong question by focusing on how AI can accelerate software development. The better question is whether organizations have an engineering system robust enough for AI to operate within, because AI amplifies whatever system it’s given, for better or worse. If requirements are ambiguous or architecture uncontrolled, AI will scale those problems rather than solve them. Rather than replacing engineering discipline, AI is positioned as a new layer within the EOS, capable of identifying ambiguity, predicting defects, assessing change impact, generating tests, and preserving institutional knowledge, while safety-critical assurance mechanisms like ISO 26262 traceability and cybersecurity remain essential and non-negotiable.
The talk concludes with three challenges to the industry: measure engineering by outcomes rather than compliance; build connected systems instead of isolated processes; and ask how AI can amplify engineering excellence rather than whether it can replace human expertise. The closing message is that the future of automotive software isn’t less engineering, it’s better engineering, built on a system where standards, human expertise, and AI work together rather than in competition.