We engaged Viston to reduce visual defects on our brake assembly line and prevent unplanned downtime on CNC cells.

Business challenge

2.2% defect rate on a high-volume line and six unplanned stoppages per quarter causing missed OTIF targets.

Our approach and solution

Viston deployed edge computer vision on NVIDIA Jetson, trained YOLOv8 models on our labeled defects, and integrated PLC signals.

For maintenance, they ingested vibration/temperature streams, built anomaly detection with MLflow-managed models, and set predictive alerts in our CMMS.

AI applications and benefits delivered

Real-time defect detection, drift monitoring, predictive maintenance with early warning.

Benefits: fewer scrap/rework, higher OEE, fewer line stoppages.

Cost of implementation

EUR 110,000 including edge hardware, data labeling, MLOps, and OT security.

Time to implement

16 weeks for two lines and three CNC cells.

Tools and technologies used

AWS SageMaker, Jetson Xavier NX, OpenCV, YOLOv8, Kafka, Grafana/Prometheus, MLflow, OPC UA connectors, HashiCorp Vault, S3 with VPC endpoints.

Quantitative outcomes

Defect rate reduced 47% (2.2% → 1.16%).

Unplanned downtime reduced 32%.

Throughput increased 11%.

Payback in 9 months.

Key performance indicators (KPIs) tracked

OEE, FP/FN on vision, MTBF, MTTR, scrap rate, rework hours, alert precision/recall.

Pre- and post-implementation metrics

OEE: 73% → 80%.

Scrap: 3.1% → 1.9%.

Unplanned stops/quarter: 6 → 4.

Vision precision/recall: 0.94/0.91 after calibration.

Stakeholder quotes or testimonials

“The inline vision system removed subjectivity and made quality scalable. Operators trust it because false alarms are rare.” — Plant Manager

Regulatory or compliance considerations

GDPR for any operator-facing cameras, ISO 27001 controls, TISAX for automotive information security, network segmentation for OT.