From idea to AI solution - Case study (17:40 - 17:50)
How can AI control the quality of the QA process? The story of the development of an innovative AI system based on video analysis, from the initial idea to a testable prototype, presented by representatives of the supplier (Cognexa) and the customer (Hollen).
How do you safeguard quality when people, time, and monotonous work are the deciding factors? A Slovak company from the field of quality control in the automotive and manufacturing industries answers with a pilot deployment of artificial intelligence that monitors adherence to work procedures and brings new data as well as reassurance. The story shows why robots don’t always help and how AI can add value as early as the prototype phase. The company has been engaged in quality control, rework, and technical support across Slovakia, the Czech Republic, and part of Hungary for 23 years. From a single location it has grown to 13 branches and approximately 460 workstations, with its services used not only by major manufacturers such as Volkswagen, Kia, and Škoda Auto, but also entire supplier chains. In the "Cinderella" work they separate good parts from defective ones and can intervene within an hour when a problem appears on the line. Over the last decade nearly a billion components have passed through their hands, and the claim rate has remained below 3.6 ppm, which speaks to precision. An analysis of claims and 8D reports, however, showed that more than half of the defects are caused by failure to follow the work procedure. During long and monotonous inspections it is tempting to shorten steps, especially if a thousand parts in a row look flawless. Not even bonuses or penalties at four in the morning will fundamentally override this human factor.Cinderella in manufacturing and numbers that compel
Pavol Václav
