What a pleasure it is to hear one of our clients announce that they have gone from a 7% to a 2% rejection rate in just 1 month!

 Interview with Hugues Poiget, CEO of Scortex, a French company founded in 2016 and a member of the Trigo Group, a leader in operational quality. Scortex is committed to automating, transforming, and making visual inspections in industrial production more efficient with its quality control solutions using AI.

Can you describe Scortex in a few words and explain its mission in the industry?

Hugues Poiget : At Scortex, we help manufacturers reduce the cost of inspecting 100% of production by developing Artificial Intelligences for them and designing with them the vision systems that allow for inspecting the appearance of the parts they manufacture. Our mission does not stop at sorting parts; our vision is to generate the data needed for production optimization through Spark, our quality control automation system. Tomorrow, we can imagine that the adjustment of the production machine parameters will be made in real time based on the quality measured by our systems at the end of the line.

What technology is behind Spark, and why is it an asset for manufacturers?

Hugues Poiget : At Scortex, we teach the machine to sort the parts and categorize defects. Thus, to sort parts, the piece must be viewed with its own eyes, possibly turned over, inspected on all sides, and then decide whether to discard it or not. For this, we have chosen cutting-edge technologies ranging from Artificial Intelligence to ensure uncompromised detection performance that must be as good as a human's, to PLC automation (industrial digital computer adapted to manufacturing process control) for real-time decisions, which allows, for example, to actuate an arm that ejects defective parts. Caution, we do not work magic; the defects must be visible. For this, we took inspiration from how operators look at parts and developed lighting and imaging techniques that allow us to inspect the parts, including shiny parts, which are very common in the cosmetics, medical, automotive, and other industries. And because we use artificial intelligence (AI), the solutions we deploy are like a new recruit that needs to be trained on the production line. We start by teaching the machine what conforming parts are. The AI then becomes capable of detecting anything that strays outside the acceptable variability of the line. A human then confirms what is acceptable and what is not before the AI plunges into the deep end. Obviously, humans put parts in the red bin less often due to the blue light from the forklift truck backing up that enters the imaging zone or because there is a drop of oil left on the machined part. It is precisely to help manufacturers manage these very concrete issues that we have set up a dedicated team. Our clients confirm that this signature brings immense added value. 

Which industrial sectors benefit the most from the Spark solution?

Hugues Poiget : We have always wanted our systems to be relevant for the largest number of people. Thus, our systems inspect everything from automotive gearbox forks to lipsticks. Historically born in the automotive sector, we are now also present in the cosmetics industry, aeronautics, packaging, medical, luxury, forging, and electronics, to name a few. Today, tens of millions of parts manufactured by Small and Medium Enterprises as well as large international groups have been inspected by our systems and are in the hands of end-users.

 How does Scortex integrate within the Trigo group, and how does this strengthen its positioning?

Hugues Poiget : We have always been very "hands-on" at Scortex, and thanks to Trigo, we have access to over 300 factories in more than 25 countries. A chance we did not hesitate to seize to continue moving forward together. What unites us is this global approach to quality. The costs of quality are numerous: scrap, training, penalties, sometimes recalls, etc. Trigo now allows us to operate beyond Europe: the United States, Canada, Brazil, India, and more recently South Korea, with an unbeatable level of responsiveness.

What are the main trends in industrial quality control that Scortex anticipates?

Hugues Poiget : Today, we wish to help quality controllers in each of their main missions: to avoid having to look at parts for 8 hours a day to find anomalies, accelerate their doubt resolution between an acceptable anomaly and a real defect, and provide them with quality data so they can quickly adjust the severity of sorting based on production volumes. The first major trend we anticipate is that AI-based tools will quickly become commodities. We all have in mind the recent example of Chat GPT. AI is entering our daily lives and is already making its way into factories. This is, among other things, what will allow a manufacturer to be more competitive than others. The other trend is the necessary reduction of the industry's impact on the environment. At Scortex, we understand that we must reconcile production imperatives with quality obligations. This is essential if we want to maintain a certain standard of living accessible to the greatest number. The use of tools like ours, by improving the efficiency of production lines, helps minimize the use of energy and raw materials.

Why do manufacturers choose Scortex for their automated quality control needs?

Hugues Poiget : At Scortex, we do not shy away from getting our hands dirty. It is necessary to move from proof of concept to a system mounted on a production line. Our clients have come to us because they were looking for a powerful yet simple solution that addresses the complex issue of visual inspection. Our Spark product is the combination of a tool usable by a quality operator directly on the production line and a control tower in which the quality data from the factory is aggregated, called the Quality Center, which allows for production optimization. Everything is provided by Scortex, from hardware to software, including the algorithm. And because quality is alive, our teams are ready to support manufacturers in all phases of the project. What a pleasure it is to hear one of our clients announce that they have gone from a 7% to a 2% rejection rate in just 1 month!

To better understand the cost of quality within a multinational, you can read our interview with Bill Black (Part I) and (Part II), former director of quality at the ABB group and previously at Airbus.

Here are some other articles you might find interesting:

• The applications of artificial intelligence in quality control

• Spark helps detect 100% of chips on glass containers in the luxury industry

• Spark Multi View Video: up to 4 cameras simultaneously to detect multiple defects in one pass