Episode Transcript
[00:00:00] Speaker A: Foreign.
[00:00:06] Speaker B: Hello everybody and welcome back to the Control Alt Manufacturing podcast, Resetting and Rethinking Manufacturing. All right, little level setting. Here's a brand new podcast. I'm going to remind you what we're talking about here. So this podcast going to be exploring the people, technologies and strategies driving the digital transformation of manufacturing. Each episode is going to feature conversation with an industry leader, a system integrator, an engineer, innovators, and we got a good one today. But before we jump to that, I am one of your humble hosts, Gary Cohen with WTWH Media joining me as always the better of the two hosts, let's be honest, Stephanie Neal. Stephanie, how you doing?
[00:00:43] Speaker C: I'm good, Gary, Happy to be here.
[00:00:47] Speaker B: I think we got a good one today because we're going to be talking to John Lee, senior manager of Manufacturing intelligence at Matrix Technologies about uniting disparate data and, and other things. Should be fun.
Full disclosure. John and I saw each other very recently at the CSIA event in San Diego. The last time I saw him we were both up on stage and he was getting an award, so he's pretty fancy.
We've also both been doing a little bit of traveling. I think you were recently in San Antonio. I was in San Diego. Tell me about what you learned at your conference.
[00:01:21] Speaker C: Yeah, well, I've been traveling a lot. I'm very tired right now.
[00:01:24] Speaker B: But let's be honest, part of that traveling was for fun.
[00:01:28] Speaker C: Though I did have a little vacation in between.
I don't, I'm not going to talk about that right now. I don't want to make you jealous, Gary.
You know, I'm already somewhere in Europe.
But anyway, yeah, for work I've been traveling, went to, you know, a couple conferences. I mean some of the big ones in Detroit, Automate, and then, you know, some, some other, you know, vendor specific conferences. But everywhere I go, everywhere everybody goes, right now all you hear about is AI. And I gotta admit, I'm a little scared.
It's like taking over the world, but I mean, it's making everything easier for manufacturers, for operators, which is great. I think that's the ultimate goal here, especially with digital transformation.
But you know, I'm scared because like now we're getting to the point where it's acting autonomously and making recommendations and then eventually making decisions or being able to program that PLC and do we actually need the engineer anymore? The answer is we do. Because you have to be able to trust the data that is underneath all that AI and so you need that human in the loop. So that's what I'm hearing and I think it's a great segue into the discussion we're gonna have with John today about uniting disparate data. Because obviously again AI is driving everything and it's really important in the manufact plant right now. But you have to trust the underlying data and it has to be clean and connected well.
[00:03:05] Speaker B: And manufacturers, companies are producing more data now than they've ever produced in the history of the world. There's data coming in and it's really what do you do with the data? As you said, is it clean, is it reliable? You know, collecting all of this data and then not using it doesn't do anybody any good. So yeah, it's AI hopefully will help sift through some of that. But. But yeah, get it. Getting through all of that data is going to be, is going to be a challenge. Which segue is why we have our expert with us today which is John Lee, Senior manager of Manufacturing Intelligence at Matrix Technologies. John is responsible for the successful implementation of manufacturing intelligence and manufacturing operations management projects. He has in depth experience in manufacturing and information systems design and has been with matrix technology since 2007. Also a recent award winner. John, go ahead and join us, we're happy to talk to you today.
[00:04:00] Speaker A: Yes, hello Gary. Thank you. Thanks Stephanie.
[00:04:03] Speaker C: Hey John.
[00:04:04] Speaker B: Very happy to have you here. So yeah, the last time we saw each other we were both dressed fancy and standing on a stage getting our picture taken.
[00:04:12] Speaker A: Yep. Yeah it was a great time out at CSIA and yeah, we're definitely thankful for everybody and receiving the award. So it was a great event. Week was very fulfilling.
[00:04:23] Speaker C: Did we say what the award was? We gotta say what the award was.
[00:04:27] Speaker B: I was just gonna say that. Do you want to say it John?
[00:04:30] Speaker A: Yeah, sure. So Matrix Technologies was selected as the Control System Integrator of the year in our category.
We are the fifth time winner hall of famer. So very excited about that.
[00:04:44] Speaker B: Yeah, we have our Control Engineering, Plant Engineering System Integrator of the year awards every year.
This year in the mid size category integrator category, Matrix Technologies was a winner. So we had a couple people come up and accept that award which was fun. But let's start out with just you John, tell us a little bit about yourself, how you got into this position and if you want to a little bit about Matrix and what sets you guys apart.
[00:05:09] Speaker A: Sure, absolutely. So I started my career in information systems. I worked with a commercial company for quite a bit of time. I moved over to Matrix and manufacturing as you said. Back in 2007. So I've been here since then and it's evolved, right? Starting with just SQL databases, data warehouses, ingesting data, creating different types of data streams, and then into the MES operations management space, really just making data work, helping with decision making, analytics. So I spent my entire career doing that with information systems and now specializing in industrial manufacturing and what we call digital manufacturing systems.
So Matrix, we are a full service engineering firm. We have three divisions, Engineering, industrial systems and manufacturing systems and solutions.
So we are kind of full service turnkey. We can help with engineering, construction, EPC type work. We do controls, automation, and then all of the information side for digital manufacturing infrastructure virtualization as well.
[00:06:17] Speaker B: I always enjoy talking to system integrators because you guys know how to make things work, which I appreciate.
We talk a lot on this about digital transformation. Do you look at manufacturing intelligence as synonymous with digital transformation?
[00:06:32] Speaker A: It is somewhat. It is kind of just more on that information side. There are a lot of components to it. You know, one of the biggest ones is cybersecurity.
So that's why we have different groups that specialize in different areas that allow us to bring, you know, that value and expertise on the projects. So those different disciplines.
But yes, so it is, is kind of all falling under that umbrella now within digital manufacturing. Right. And the transformation is just the process, that journey of getting there.
[00:07:02] Speaker C: So that process and that journey, John, Gary and I were just talking about AI taking over the world and how we get there. But how are manufacturers addressing the challenge of disparate and siloed data in digital manufacturing?
[00:07:19] Speaker A: Yeah, sure. So we continue to see, you know, the, the plant floors still have a mix of new and legacy systems. Even some of the greenfields, you know, still don't have the latest and greatest. Not everybody wants to be on the cutting edge. So a lot of the data comes from different OEMs. So a lot of manufacturers, you know, bring in different OEMs for different pieces of equipment that have their own systems or controls and PLCs on them.
And what we do is kind of look at the overall architecture and how we can start bringing that data together when we go into existing facilities. Some of the assets, the devices aren't even connected to Ethernet and they might not have a very secure infrastructure down at that OT layer. So we have groups that can help with that, make recommendations and help bring them up. So a lot of the disparate data is just coming from different machines that talk with different communications protocols.
And we're really looking to kind of you know, standardize that kind of level the playing field, bring in some of these newer technologies that can allow us to bring the data in reliably.
We call that data integrity.
Also allow us to kind of cleanse the data a little bit for what roles and where the data is going.
Some data will go to historians for process data. Other data is contextual. We'll go to, like, database systems, and then some will be ingested into the cloud. Right. So as you look at this transformation in this journey, you also want to keep in mind how are you going to extend it and how are you going to scale it? What do you want in the future so you're not painting yourself in the corner. We call that technical debt. You want to avoid that, and you want to be open and agile and be able to move with change.
[00:09:06] Speaker C: I want to ask you about the protocols you just mentioned, but before I do, you mentioned that some of these things aren't even connected. Is it because the manufacturers are afraid of a cyber threat? Is that why they're not connected? Or is it.
It's working, so why should we fix it?
[00:09:27] Speaker A: Some are just delivered as individual work cells and individual assets. They're not always connected in a line with supervisory PLCs, so in that sense, they may not have connection. They may just still have IO connections that control the equipment within.
And what we want to do is, you know, access the data that can be available. And in some cases, we're even adding some sensors. Right. There's this IIoT industrial Internet of things where we can have very lightweight sensors for vibration, proximity, different kind of monitoring temperature.
So we can bring all these things together and start, you know, getting the data to flow throughout the facility.
There are protocols that help, you know, whether you're having, you know, this disparate of, you know, different PLCs, and you have the sensors and you can have different pieces of equipment.
So there are some that, you know, just lend itself to that. So you can use edge devices or different types of, you know, protocol converters. There are gateways for opc, ua. We also have MQTT nowadays, which is a good protocol because it kind of cleans everything up and can use brokering technology, which is more.
More report by exception type of, instead of pull response that doesn't overload the network so much. So there's a lot of advancements that are helping us do that today. A lot of new tools.
[00:10:49] Speaker C: Okay, so I think you answered my question about, like, what some of the protocols are that are used to integrate heterogeneous Data sources. Are there any other things that you're working with?
[00:11:01] Speaker A: So some other technologies that we've seen, you know, that we are looking at, some of these are just proof of concepts, right? As people are just getting started with stuff.
But certainly like hiveamq is a really good broker that we worked with at mqtt. We have worked with like kepware, opc, ua. A lot of the software vendors have their own stacks. You know, Rockwell has very good data streams and data Mosaic, Aviva has Connect. So a lot of them have that. We also use Ignition quite a bit. It has a lot of protocols that access the control systems themselves. And then we can, you know, I guess set up brokering technologies and different things to really unlock that data and create that ecosystem and start to get data available to all the consumers.
[00:11:46] Speaker B: I come from a mo mostly a cybersecurity background. So whenever we talk cybersecurity, I perk up a little bit.
You were talking about legacy systems earlier. How complicated does that make the integration when you're dealing with these legacy systems? That A can be very old, but B were not designed with things like security and communication and data gathering in mind.
[00:12:09] Speaker A: Right. So our system infrastructure group specializes in that. And what they look at is, you know, how do we keep things isolated within the OT network?
So the plant process control networks and the plant networks they, you know, very much look at, you know, how do we use the managed switches to set up the different VLANs, keep things isolated at those lower levels. And when we talk about moving data around the data that, you know, does need to move or be promoted, we look at, you know, using edge devices to do that. That can be secured, and that's us outbound only traffic. Right. So we have to have the layers of security with the firewalls in place, and we have to architect that in a certain way.
[00:12:50] Speaker C: Well, I just want to jump in for a minute.
You know, there's been a lot of discussion too about like cybersecurity and managed services or whatnot. But when Matrix Technologies goes in and implements a cybersecurity type of solution, ultimately who's responsible for that? Is it like, do you guys get involved once you've implemented it, or is it, you know, we keep talking about it ot that, you know, they're, they're sort of coming together a little bit more. Who's responsible for the cybersecurity solution that's implemented?
[00:13:23] Speaker A: We do offer services with that to do some management of the cybersecurity and mostly on the OT layer, a lot of manufacturers we work with, they do have, you know, groups that provide those services within the IT to help monitor the upper layers of that. But when it comes to the ot, we certainly do take some of that responsibility on.
Yep, got it.
[00:13:48] Speaker B: Back to manufacturing intelligence solutions, how do those enhance data integrity and improve reports and dashboards and the things you need to be able to track all this data in manufacturing?
[00:13:59] Speaker A: Yeah, certainly. So, you know, the data integrity, when we look at the different systems, they can have, you know, data entry from manual or different sensors or automation, they could provide feedback, you know, within itself to make sure that they're getting accurate readings. We can detect different things, like if there's anomalous or invalid readings. Even process historians have had this for a while, you know, where they have clean data or not, or they're getting good signals. But manual entry is one of those bigger things too, Right. So when we still have some operator entries, you know, they could have, you know, malformed strings, you know, invalid characters, different things, and then we still have some interaction. Even when you think about like taking weight scale measurements, right, that are not fully automated, you know, operators are still responsible to make sure that the scale measurements are right. So as we automate and take some of that out, that helps maintain that data integrity from the source of where the data is generated. And then as we promote it through the system, a lot of the protocols now can tell whether or not the, the packets are receiving are clean or, you know, they have an issues with the data.
And I would say the other thing that we look at whenever we architect systems and want to move the data around is just making sure that it has the right context. Right. So instead of just sending values that, you know, may or may not mean a lot or can get out of line if it has, you know, kind of records with it, a header message, you know, in the right context of where the data is coming from, it's easier to maintain the integrity of that from the source to the end.
[00:15:35] Speaker B: So we talked about this avalanche of data that manufacturers are collecting these days. Why is having collective information readily available important for everything from production changes to issue resolution and manufacturing?
[00:15:48] Speaker A: Yeah, So a lot of manufacturers, right, so there's this decision making process, right. We're always looking at continuous improvement, process improvement type things.
But when you're looking at the data, you know, even hours later, at the end of the shift or weekly reports, there's no time to really take action and really accelerate and make some changes or improvements.
And it really empowers, you know, kind of democratizing that and getting that data available to more roles, all the way down to the operators in the floor, allows them to be involved with some of that decision making. And it can really expedite, you know, the changes, the improvements, you know, set points need adjusted or anything. It gives them the ability to do that. Right. And as said earlier, we're, we're getting to new technologies actually can recommend changes.
So having that data available really accelerates that value of being able to make change.
That's why the real time data is important, is really to make decisions in real time.
[00:16:49] Speaker C: Yeah, And I had mentioned earlier having a human in the loop, like everybody gets so nervous. We have all this automation. Am I getting automated out of a job? But really like when you said, John, you need context around the information you're collecting. I mean, that's where the operators and plant managers are really adding value. Right. And they have to be there to make sure that the data is correct for sure.
[00:17:15] Speaker A: You know, they are still some of the most, most val. Most valuable decision makers. Right. So they are truly the stakeholders. Right. They understand the systems. These tools just enable and help. Right.
So they're providing some guidance. But ultimately the decision, the decision would still be for the operators. Right.
There is, you know, a lot of talk about, you know, automated process control. Right. And having the AI be able to do that.
But I think for the most part, you know, we're going to continue to have that human element involved, you know, but it just helps in the system and enables them to do a lot more and have a lot more information to make better decisions as well.
[00:17:55] Speaker C: So speaking of making better decisions, how does manufacturing intelligence drive and enable things like predictive maintenance or AI enhanced operations?
[00:18:09] Speaker A: Sure, absolutely. So, you know, AI can help with some of the anomaly detection things, but then we cross into the machine learning part of AI. Right. So when you can model things and look at different variables and different conditions, you can actually detect a lot more, I guess, you know, drive more process change. Right. So you can see more of the bigger picture and it can process it so quickly that it can help make those recommendations or can show you when things are about to occur. So it can start to predict things when it starts seeing certain variables and situations and it can predict those outcomes, and that's very helpful. Right. So predictive maintenance, you mentioned, is a huge area right now because, you know, you can have a lot anomaly detection, you know, within vibrations and temperatures and different things on the assets and equipment themselves.
And actually we see where OEMs are starting to offer this Software as a service that can help with maintenance and monitoring those things and it can help provide notifications and warnings predictably, like well in advance and start to doing some prescriptive. Right. Offering some solutions. Right.
So it's really been a game changer. It's really added quite a bit of value.
[00:19:25] Speaker C: Where are your end users in terms of the, of their deployments and are there any customer examples that you can cite? I mean, you don't have to give away the names, but I'm just curious, like where you want to.
Yeah, give us the big names.
But like early in their journey here of digital transformation.
[00:19:46] Speaker A: Yeah. So most that we work with are still somewhat early. I mean, they, they are, you know, they have their transformational strategy in place, but they're still early in some of those steps. And as mentioned and the technologies on the high end and we talk about the cloud, the AI, the ML, you know, those are there and those are ready to do pilot programs, proof of concepts and initiatives. It's the connectivity of the plant floor that most people are in right now that we work with and they're moving quickly. But it is a large effort and undertaking. As you look at connecting up different PLCs and controllers, you're finding that some don't have the upgraded firmware and some don't have the upgraded cards and the necessary things to actually connect up the entire plant floor of the entire facility the way that you'd want. So some of the end users we have, like in life sciences, they're bringing their labs up to scale. They're a little bit further in their journey.
They're really looking at building out unified namespace uns.
So where they're building an ecosystem where everybody's providing the data through brokers, it is ingested into the cloud. And they're already able to run different types of analytics and different types of machine learning on what they're looking at.
And a lot of the bigger manufacturers, you know, those are kind of still looking at connecting up plant floors. Right. So everybody's at a different place in their journey. But I'd like to see, you know, it's great to see everybody is, you know, moving forward. Right. Moving towards digital manufacturing and higher automation.
[00:21:21] Speaker C: So it's one of those things that you have to do. Right. Like you're going to get left in the dust. You're not going to be around in the future if you don't embark on this journey now.
[00:21:29] Speaker A: Yep, for sure.
[00:21:31] Speaker B: On the flip side of that, I wanted to ask, do you still have customers who Push back against some of this. Why do we need everything connected? Our systems are air gapped and or are we at a point now where everybody understands that they have to be on this journey?
[00:21:44] Speaker A: I think there are still, you know, there is that smaller presence, right, of people that still feel like, you know, we're doing great, you know, we're making product, we have a lot of processes we've grown over time for continuous improvement. Reliability programs, you know, they're a little bit slower to adopt.
But I do think as things progress and we see more and more successful transformations and use cases, business cases, you know, they are starting to take a look at, you know, what value can this provide me and to. And in some cases they are reaching out just for consultations. Right. They want a transformational assessment just to see where they are.
They may not be ready, but they're thinking about that strategy.
[00:22:27] Speaker B: So and I have to ask because Stephanie and I started, started this podcast by talking about this scary Skynet. The machines are going to take over future.
How much is AI impacting your day to day job and performance right now and how much has that changed in the last few years?
[00:22:46] Speaker A: Yeah, I mean certainly there are ways that can accelerate, right. So there are AI tools that we use to help with coding and testing different things there, you know, in the office we have like the Copilot, you know, the large language model, LLMs that can help, you know, with some documentation and wording and that helps make us more efficient, accelerates a little bit. So as a company, you know, we have a lot of internal initiatives as well. Just looking at where it can help with our internal processes. And these are the same things and strategies that, that we talk with our customers about.
[00:23:18] Speaker B: Yeah, I mean even as a media company we use it fairly regularly to help us through certain processes. I mean it's not, again, hopefully it is not replacing people at this point, but it is definitely streamlining workflows for us.
[00:23:33] Speaker C: And it's one of those things again that like you have to get in front of it because it's not going away.
So figure out how to use it to your benefit. So, you know, that's what we're doing internally. It sounds like, you know, that's what you're doing internally, John. That's what your customers are doing. I mean, I asked an executive recently because I'm like, this is scary. AI scares me a little bit. I shouldn't say that out loud, but I do.
And I'm like, like, should we like is there gonna get, is it gonna get To a point where it's, you know, making decisions that maybe aren't correct and we should be worried. And he's like, oh, no. Remember when the Internet came out and everybody was like, oh, no, this is going to ruin the world and look at us now.
[00:24:20] Speaker A: Yep. For sure. I think it's just a matter of adoption. Change management could be a process for anything, especially one of these big disruptive type technologies. Right. And certainly this is currently and will continue to be right as it evolves. And we'll start seeing newer and newer generations of AI come out that continues to do more and more. But again, you know, it's really as a tool, it's an assistant. Right. It can really empower, enable, and really help promote as far as taking over and making all decisions, I'm not sure how far away that is, but I think it's pretty far.
[00:24:58] Speaker B: I mean, as people say all the time, AI is the worst it's ever going to be at this very second, it just gets better and better and better. Although I will, Stephanie, I'm not sure. It depends on what day you talk to me, but I think I can make a solid argument that the Internet did ruin the world.
[00:25:12] Speaker A: Yeah.
[00:25:15] Speaker B: So I don't want to refute the expert you were talking to, but I think I could give you a couple of data points.
[00:25:21] Speaker C: True.
[00:25:21] Speaker B: Yeah.
Anything, Stephanie, Anything else I want to ask John? Just a John Lee question. Anything else you want to talk about about disparate data AI machines?
[00:25:33] Speaker C: No. I mean, you know, this is a great, you know, good conversation, good starting point. I think we should have John back at some point on our podcast to maybe talk a little bit more about other things because everything's evolving so quickly.
But, yeah, no, thank you, John, for being here.
[00:25:50] Speaker A: Thank you.
[00:25:51] Speaker B: You're not off the hook yet, though.
[00:25:52] Speaker C: Yeah. You're not up the hook.
[00:25:53] Speaker B: I just want to ask. I love talking to engineers, system integrators.
What is just personally, what is your favorite thing about what you do? What gets you excited to come to work every day?
[00:26:05] Speaker A: My excitement is just seeing things evolve. Right. So helping people solve problems when we can bring a solution in and those light bulbs start clicking and people really start seeing the value, you know, those are great times. Right. So big accomplishments for me and my team and Matrix as a whole. So, you know, I really enjoy that. That's what I've always, like, enjoyed is just helping people. You know, at work, in my personal life, I like to help people, and so I get to do that every day. Right. You can use technology we can turn data into information, help solve problems, make decisions.
I've always enjoyed that and we'll continue to enjoy that.
[00:26:44] Speaker B: Fantastic. John, huge pleasure talking to you again, I should say, because we just talked about a week ago, but really, really a pleasure to have you on the podcast. Thanks so much for coming on.
[00:26:54] Speaker A: Yep, thank you. I appreciate it. Have a great day.
[00:26:57] Speaker C: Thanks, John.
[00:26:59] Speaker B: All right, some great information there from John Lee of Matrix Technologies.
Yeah, it's the, the, like we said at the beginning, the data thing, there's so much of it coming in and it's what you do with it. If you're not using it properly, if you're not collecting it, if you're not cleaning it, if you're not making actionable decisions off of it, then it's just. Yeah.
[00:27:22] Speaker C: And you know, manufacturers really need system integrators. Like, I really feel like sis are overlooked, like in the grand scheme of things and their role is becoming more and more important, especially with digital transformation. So, I mean, again, everything that John was talking about and the protocols and I mean, a lot of this is new technology that, you know, operators and plant managers, they're just trying to make sure that their equipment is running and they're getting product out and they don't have downtime. Like, they're not always on top of all of the new advances in technology. And you know, the system integrators are out there, you know, in the trenches and talking to the technology providers and they have great partnerships and keeping ahead of the cutting edge stuff. So yeah, it was great conversation with John. And the data is so important, I think it gets overlooked because again, as we keep talking about AI and analytics and yeah, that's great, but if it's running crappy data, you're not, you know, what is that term? Like, not like if there's garbage in, garbage out.
[00:28:35] Speaker B: Yeah, one of my favorites. I like that one.
That works for AI too.
And I totally agree with you about system integrators. Like I said, I was at the CSIA conference in San Diego in June and the thing that I like about system integrators, and it's not like all engineers aren't problem solvers, they are. But system integrators tend to have a really good global view. So they're looking at, at everything from cybersecurity to data to automation to everything that's going to come together to make a system work. And yeah, I just, I tend to find them very practical and kind of fun to talk to because they usually.
[00:29:09] Speaker C: Have a problem very humble. Very humble. Right.
[00:29:12] Speaker B: Like, except the John Egomaniac, that guy.
No, very humble. Yeah. No, no, it was, it was great.
[00:29:18] Speaker C: Yeah.
[00:29:19] Speaker B: Appreciate having everybody out there listening to us. Obviously. It's a new podcast. If you want more great information on anything of the topics that we talked about today, check out our suite of engineering brands, whether that's consulting, specifying, engineer, control engineering, plant engineering, all sorts of engineering brands where you can find this stuff in video form. We have other podcasts, we have great articles on all of this information and we really appreciate you stopping by and listening to Stephanie and I ramble here at the end of the podcast.
[00:29:48] Speaker C: Yeah. Thanks for tuning in.
[00:29:50] Speaker B: I always want to say we'll see you next time, which is nonsense because we're not going to see them. So you'll listen.
[00:29:57] Speaker C: You'll see us next time.
[00:29:58] Speaker B: You'll see us next time. That's it. Thanks for joining us, everybody. Bye. Bye.
