Issue link: https://iconnect007.uberflip.com/i/1534953
48 SMT007 MAGAZINE I MAY 2025 we could make a more significant investment and go to market with a disruptive technology. I have always been in love with artificial intelligence. I have a long history in the opti- cal inspection market, and I always under- stood that when you go into it with a stan- dard mathematic algorithm, you will have limited functionality. Only the human eye can understand what you are checking in a very precise way. So, my point was to repeat the capability, the precision of the human eye inside an optical inspection system. We developed a unique optical head that works 30 centimeters from the board. All our proprietary neural networks are part of agnostic soware, meaning that they work with all optical heads and can analyze what- ever each optical head sees or records as data. Our initial idea was to go to market with something that did not yet exist that was unique, agnostic, very flexible, scalable, and disruptive, disruptive, disruptive. I understand you have a couple of machines. Can you explain those? Right now, we have two different machines. One is the Aton system. It is typically used in industrial automation, with components that are very tall or bolts that are very heavy. It is flexible. Aton's big advantage is that it has a fixed optical head, while the board can move under it, and the Z-axis can move the board almost 10 centimeters up and down. e second machine is Horus, which we have been talking about, where the opti- cal heads are moving and the board is fixed. Horus has a speed advantage compared to Aton. It's a machine designed and created to stay in lines where products are moving at a very high speed. Detection is very high, as is the speed that analyzes it. In the near future, we will launch Aton and also a very important new soware called Predictivity, which will allow both machines to have predictive capabilities rather than only detecting errors at the end of the line. Of course, to do this job, the amount of data you need to acquire is huge, and you need to process and analyze this data very quickly as well. For this reason, TowerAI has a huge advantage. It is why we are capable of doing this within our machines. Predictive is ultimately where we want to be. Yes, that is really the point of all this work. It's important to understand the error and be able to avoid it. And from an environ- mental sustainability perspective, it's ver y important not to waste raw materials and energy or create CO2 in an excessive way. It is ver y important to save and optimize all resources. e predictability soware will be capable of optimizing resources and allow for huge cost savings opportunities with these materials. With predictive analytics and digital twin upfront, we could achieve a lot around sus- tainability for electronics, which is certainly an area we should all be concerned about. Yes, this is very important; it's the future. Of course, right now, we talk about the produc- tivity. It looks like a simple task, but we need to understand what is behind it. First, how do you manage such a huge amount of data? Second, what do you have to do in terms of hardware and soware to be capable of man- aging this data? We are solving it. Roberto, it's great talking to you. ank you, Marcy. SMT007 " Our initial idea was to go to market with something that did not yet exist that was unique, agnostic, very flexible, scalable, and disruptive, disruptive, disruptive. "