Siemens and Procter & Gamble are expanding an AI-based visual inspection system across P&G manufacturing operations worldwide. Developed jointly by the companies, the Visual Inspection Cockpit, or VIC, inspects products in real time at full production speed and has reduced scrap rates by 10–20%, depending on the product, tells Design News.
High-speed consumer goods manufacturing presents difficult inspection problems. Delicate and textured materials can shift, stretch, wrinkle, or overlap as they move through production. Traditional machine vision systems often require significant reconfiguration when materials, packaging, or operating conditions change. VIC uses industrial AI to adapt more effectively to these variations.
The system combines P&G’s deep-learning models with Siemens’ Industrial Edge platform, industrial PCs equipped with Nvidia GPUs, and supporting AI hardware and software. It analyzes live camera images and can inspect thousands of products per minute. The technology is designed to detect challenges such as low-contrast defects, overlapping components, and problems in highly decorated products and packaging.
Plant engineers can configure, train, and update inspection models through the Visual Inspection Engineering Tool without relying on dedicated data scientists. Inspection results are processed close to production equipment and integrated with manufacturing operations. The system can issue alerts or automatically reject individual defective products. Collected quality data also helps teams track process stability and identify opportunities for improvement.
A major advantage is scalability. VIC is delivered as a reusable application on Siemens’ Industrial Edge platform rather than as a custom vision system for each production line. According to the article, this approach can make commissioning five to 10 times faster than traditional bespoke systems.
With established infrastructure, integration patterns, and DevOps processes, P&G can replicate the technology across plants and inspection scenarios with less deployment overhead. The project demonstrates how AI-based machine vision is moving from individual factory experiments toward standardized, globally scalable quality-control systems.
