Issue link: https://iconnect007.uberflip.com/i/1528798
16 SMT007 MAGAZINE I NOVEMBER 2024 Suddenly you can have a dedicated AI agent as their lean manufacturing resource, on every single factory line. Just to be clear, this is happening on your specific factory data? Burke: Correct, it's happening on your factory data. ere are a couple of key things here. First, you mentioned the linking of data. You need to fuse together everything you know from the equipment: all the process parame- ters and things that CFX helps you collect. But it's also super important to know the context, what product is being built, and what step in the process is represented. at's data in your MES, your product data, your bill of materials (BOM) data, and maybe data from your ERP, which you then link all together. ere's another source here: What are all the observations in the heads of your frontline workers, like your quality engineer or your operator? Maybe the machine doesn't know something about the process, but your people see it, and they know. How do you capture that and then link it up with what the machines do know so that sud- denly you have a "data digital twin?" It under- stands the whole workflow of what's supposed to happen and has a lot of rich information on how every little step happens. Now you've got this asset, which is digitized and available. You can put up a dashboard, have data scientists look at it, and make KPIs for executives on top of it. But what can you do now that you couldn't last year? You can have an automated tool try to solve little problems and focus on one task. For example, "Drive down- time from four to three hours a day." I can know exactly what it is. I have my Pareto of positives. I can link all this data together. I know exactly which ones are impacting my overall efficiency the most. Is that everything in context of my production plan, maybe from the MES? e linking of data is important as an enabling tool. What's different now is the ability to auto- mate a decision or a recommendation on top of it using these little AI agents. at is getting people really excited. This is fundamentally different than what I was hearing at IPC APEX EXPO last spring. There's been a sea change. Burke: Yeah, there's been a shi. Even at APEX EXPO, we were not even talking about our- selves as an AI company; our core focus was as a data analytics company. e change came aer we did some soul-searching and asked, "Is AI hype, or is there something new here that makes something possible that wasn't possible before?" en we developed some use cases: ones around downtime tracking, capturing task knowledge with people's heads, and doing better predictive alerting in a way that's very high quality for experts. We took that AI road- map to our customers, whose response was, "Yes, exactly that." It resonated so strongly with them that we started leaning into the idea that something new is possible. AI makes data more actionable and useful than it was in the past, which makes it even " Maybe the machine doesn't know something about the process, but your people see it, and they know. "