Issue link: https://iconnect007.uberflip.com/i/1534953
processes and sequencing, are amazing. Again, it's holistic—the machines, the people, qualifica- tions and certifications, ERP, shift management, etc. These are all part of the total MES equation. How is AI playing into all of this? Everyone's talking about AI. It's a global trend. It's impossible to have a conversation without talk- ing about AI. In terms of AI being a kind of a subset and a conse- quence of machine learning and developing the advanced algo- rithms that would lead to AI, we are incorporating AI with pur- pose, not just because we have a nice feature there. The basis for AI is in our MES already. We have machine learning capabilities because our architecture has capabili- ties related to the capture of all machine data, meaning that with the ability to integrate with machines and connect with dif- ferent systems, ERPs, physi- cal equipment, and people, we have the database or a basis for all that. Our technology allows us to capture and absorb tons of information. From that perspective and with this in mind, we have estab- lished two mechanisms. One is the supervised mechanism for machine learning related to input and storage of failures. A super- vised approach is asking a phys- ical person, "What is this fail- ure all about?" or "Is this failure related to X?" The other one is the unsu- pervised mechanism that cap- tures everything and automati- cally detects failures and trends. The intelligence is that with this machine learning capability, we will be able to detect and trans- form the contextualized infor- mation that we have. It's not only machine or people data, it's the combination of machines, peo- ple, process, and traceability. We will be able to see, "I got this failure, and I had this raw material coming from this spe- cific supplier. The machine pro- cessed this at a certain temper- ature. The speed was adjusted to that level, and so on." Going forward, if you have the combi- nation of all those variables, the probability of having a failure is very high. Once the system has learned the algorithm, it knows, "I have the combination of fac- tors that will lead to failure in five steps. I need to change some- thing now." Besides what I mentioned, AI is one of our main functional top- ics in our current short, mid, and long-term roadmap. We are and will continue to invest in AI capa- bilities and will deliver a better and more efficient MES with pro- gressive AI capabilities. That is part of the ROI equation right here. Exactly. This is a large investment, not just in the software, but also in the implementation. How do you measure ROI? That is one of the most challeng- ing aspects within an MES intro- duction project in a factory. This is a journey, a process that will take its time and require invest- ment. But sometimes it's difficult to prove that before the exam- ples happen, which would be the holy grail of an MES project. Remember, there is no visibil- ity of what's happening before the MES project or program starts because it's the MES that provides the visibility. It's very hard to com- pare apples with apples. In some situations, once the MES was up, some indicators looked worse. It feels dramatic because there were things you weren't measuring before, like maybe the technician had recorded a 30-minute setup that really took two hours. The MES brings this visibil- ity to the game. Sometimes, the return on investment on an MES can't really be shown until things are settled and in place. But I can assure you that the organizations