AI and machine learning for quality management
Where quality management in the automotive industry came from, how it has evolved, and the six changes I expect AI to bring to it.
In the automotive industry, how quality is managed decides whether a vehicle programme lands on time, at cost and without surprises. This piece looks at quality management in three parts: where it came from, how it has evolved, and where I think it is going.
1. Traditional quality management
Traditional quality management rests on three things: Advanced Product Quality Planning, project management methodologies, and performance management.
Advanced Product Quality Planning (APQP) remains the cornerstone. It gives every stakeholder a structured way to communicate throughout the product lifecycle, and it is what keeps development timely, cost-effective and quality-driven.
Project management methodologies. Historically the industry struggled to deliver innovative vehicles promptly, because the traditional approach leaned heavily on physical prototypes and exhaustive physical testing. Those processes were crucial for product quality, but they were slow, and they got in the way of the agility an evolving market demands. In project management, quality is one of the three pillars, alongside cost and time.
Performance management. The third pillar is the scorecard: faulty parts per million, scrap rates, warranty cost. Those metrics are familiar for a reason — they are how high standards are maintained over time.
2. The evolution
Old methods do not produce innovation on their own, so quality management has had to change in three ways: a fresh approach to product development, the integration of advanced information management technologies, and performance monitoring that happens simultaneously rather than afterwards.
A new approach to project management. The long timelines and the reliance on physical prototypes have been transformed by communication technology. Teams can collaborate across sites in something close to real time, which accelerates development and makes room for innovation.
Advanced information management. Tools like three-dimensional conferencing give real-time access to data, efficient communication, and a direct line to customer feedback for quality monitoring.
Simultaneous performance management. Quality and supplier-quality issues now get a real-time response, continuous monitoring, and lessons learned that are shared between every party in the process rather than filed at the end of it.
3. The future
Three pillars are emerging: autonomy, open-source and adaptable solutions, and the integration of artificial intelligence and machine learning. Six things follow from them.
Open-source product and technology development. Collaborative product development on open platforms lets knowledge be shared and development run efficiently. Open-source principles give manufacturers the flexibility to select the best components and technologies from a wide range of suppliers.
Global, integrated, adaptable, standardised and modular products. With AI's help, future vehicles will use standardised software and hardware components that are interchangeable across brands. Suppliers and technology companies will lead the development of common subsystems, components, hardware and software.
Transparent, shared supplier performance metrics. Supplier performance will be tracked continuously by AI-based systems in the cloud, and those metrics will be available to every manufacturer. That transparency means resources, knowledge and experience can be shared across industries to minimise quality issues.
AI inside quality management and supplier quality management. AI will not replace quality engineers. It will take part in the Production Part Approval Process and Part Submission Warrant, assessing PPAP elements, signing off PSWs autonomously and collecting supplier quality data for manufacturers. It will extend into design assistance, material option evaluation, compliance with country requirements, risk assessment, FMEA and supplier assessment.
AI for root cause analysis and issue allocation. Machine learning will be used extensively for problem-solving. AI will address quality issues, allocate them, and communicate autonomously between customer vehicles, service, suppliers and production facilities. Emergency response, corrective actions and progress monitoring will be managed with everyone kept informed.
Early, autonomous, preventive quality systems. Future vehicles will carry diagnostics far more sophisticated than today's. Autonomous diagnostic systems will tell manufacturers about potential issues promptly, which makes resolution faster, more detectable and more transparent. Preventive actions will cover current production and future products, continuously monitored and improved with AI's help.
Where that leaves us
The road ahead for quality management promises innovation and transformative change. The questions I keep turning over are these: what do you make of the three future pillars — autonomy, open-source collaboration and the integration of AI — and how do you see them changing the automotive industry and beyond? If you work in quality, in programmes or on the supplier side, I would like to hear your perspective, your challenges and your aspirations for the field.
First published on LinkedIn in January 2024, when I was a LinkedIn Top Voice in Quality Management. Lightly edited for the site.