P&G Takes Its AI Scrap Killer Global

Procter & Gamble (P&G) is rolling out an artificial intelligence (AI) quality inspection system worldwide, expanding a tool that has proven effective on individual production lines to become standard equipment across the company's manufacturing operations. Built in collaboration with Siemens, the system inspects products in real time at full production speed, resulting in a 10-20% reduction in waste, depending on the product. New installations can be deployed up to five to 10 times faster than traditional vision systems.

The platform, called the Visual Inspection Cockpit, addresses a persistent challenge for conventional machine vision systems. Textured consumer goods like diapers and wipes can shift, stretch, and wrinkle during production, requiring rule-based cameras to be reprogrammed whenever there are changes in material, package design, or line settings.

P&G's deep learning models can handle these variations, while Siemens' Industrial Edge hardware runs the inference next to the equipment, preventing defective products from entering the packaging and assembly process.

Paul Thomas, P&G's director of machine vision and applied AI, explained that the solution was engineered to tackle the myriad of industry challenges that traditional vision systems couldn't address. Plant engineers can now configure, train, and update inspection models themselves without relying on a data science team on-site. This standardization allows the AI model and shared edge hardware to be installed, trained on new products, and switched on within a fraction of the time required for bespoke vision systems, enabling P&G to replicate the technology across factories rather than funding separate integrations at each location.

The company's AI-driven approach aligns with its broader Supply Chain 3.0 program, which has scaled up to $1.5 billion in cost of goods sold savings and has seen the successful piloting of a fully automated four-hour night shift in Berlin. P&G aims to realize productivity gains of 15-60% per automated shift. The company's focus on scrap reduction, which does not depend on sales volume or price increases, is crucial for maintaining profitability amid economic challenges such as tariffs, commodity costs, and cautious consumer behavior.

P&G's AI scrap reduction figures are internal, but the standardized approach—utilizing a single model and hardware platform across multiple plants—distinguishes it from isolated factory experiments. The company aims to complete its Supply Chain 3.0 initiative by 2030, with Schulten emphasizing the importance of accelerating the rollout of these technologies across its global operations.