P&G Scales AI Inspection: What Fastener Plants Must Prove

Siemens and Procter & Gamble announced an expansion of AI-based visual inspection across P&G manufacturing operations on September 16.

The system processes inspection results on Siemens Industrial Edge, collects quality data, and can trigger alerts or remove defective products. Siemens reports product-specific scrap reductions between 10 and 20 percent.

No fastener result appears in the release. A fastener plant cannot transfer P&G figures to heading, thread, drive recess, coating or sorting operations.

For fastener production, the test starts with a defect taxonomy: missing threads, burrs, plating voids, mixed parts and head damage. Track false accepts, false rejects, detection rate by defect class, cycles from camera change to production, and lot-level links between images, inspection results and certificates. Results must hold across part variants, surface finishes, shift conditions and line speeds.