Smarter manufacturing
In a landscape where artificial intelligence is becoming increasingly common, Sacmi chose to follow its own path. The result is Classy-AI, an application developed entirely in-house, designed to tackle a fundamental question in manufacturing: “What are the defects in my production?”
Unlike standard market solutions that rely on commercial AI libraries to solve isolated inspection tasks, Classy-AI was conceived to reflect Sacmi’s expertise in artificial intelligence. The goal of Classy-AI was to provide a smarter, holistic solution that delivers value beyond defect detection, explained Sacmi. When Classy-AI was launched four years ago, its primary users were quality control personnel, focused on removing non-conforming parts to ensure compliant batches. Today, the focus is on automated feedback in cap and preform production lines. Thanks to advanced sensors and intelligent systems, machines are no longer passive players, they learn, adjust, and optimise processes in real time.
Classy-AI automatically names defect and identifies drifts, without the need for data scientists or technical interpretation. Actionable insights are immediately accessible, and results in no wasted time or complexity. This streamlined approach vision system powered by Classy-AI is installed on Sacmi’s CCM and IPS machines.
After extensive internal validation of Classy-AI, the solution has been installed at customer sites in Europe, where it’s integrated with new anomaly detection algorithms for more easy defect identification and process drift monitoring. Following a one-year beta testing phase, customer feedback results in these key parameters:
- Significant reduction in false rejects;
- faster root-cause identification;
- improved machine setup time and production stability;
- reduced reliance on specialised personnel for data interpretation.