It will enable VOSS to provide quick and advanced warning of potential problems with their products. This will increase control over the quality of shipped goods, preventing defective products from being shipped. The result will be reduced complaint costs, improved company image and increased customer satisfaction.
It designs and manufactures, among other things, brake and fuel hoses, multi-connectors and valves for major automotive brands. Their products serve the safety of thousands of people driving cars manufactured by their customers.
The data is aggregated into a single system on which Machine Learning algorithms such as Artificial Neural Networks or Gradient Boosted Decision Trees are taught. The algorithms are able to predict with precision down to individual manufactured parts the probability with which a part is defective (the defect was not detected by the workers at the production stage) and will consequently be sent to the customer, resulting in a product complaint. Based on the probability returned, it is possible to signal which parts should be reviewed again by a human before they are sent to customers.
Additionally, we use explainable algorithms that indicate the factors that have the greatest influence on the algorithm’s decision. As a result, VOSS is able to pay more attention to the most problematic aspects of production, reducing the number of product defects that arise.