전사 🔗
Jackie Slaght: Welcome to Beyond Tech, where innovation meets insight. I'm your host, Jackie Slatt. Today I'm joined by Eric Medicine, Director of ITOT Solutions, and Kevin Roomer, Solutions Architect at eTech Group. Eric and Kevin previously led our popular AI Ready or Not live session, and today we're continuing that conversation. We'll explore the real barriers that can prevent organizations from successfully adopting AI and answer a few additional questions that came out of that live session. If you missed that live session, don't worry, you can still watch it on demand for even more context and technical insights. We've included a link into in the episode description. With that, let's dive in and get started. So I want to thank Eric and Kevin. Thank you guys for joining us today. I very much enjoyed the live session. I'm excited to have you guys continue to talk about it. So I want to see what outcomes are manufacturers hoping to get with AI. What do they think it will improve and what's the most common request?
Kevin Romer: So in a lot of cases the customers are expecting to improve their efficiencies and that type of thing, understand more of what's going on, what what they can prove on improve on on their system. That's what they typically are are wanting to get. So so they know where to invest to improve. That's the big one.
Eric Medecke: I think what Kevin said there pretty much sums it up. You know, it's what people are looking to do. I mean that's the reason they invest in AI. They wanna because we're our people are aging out and there's a lack of resources. So the ability to have AI to help them to be do predictive maintenance, things of that nature, or to make changes on in their production environments.
Jackie Slaght: Yeah. So I know AI has been the topic for I feel like years now. Everybody has been talking about it. Everybody wants it. but despite all the excitement around AI, I've heard that a lot of AI initiatives stall before they really get implemented. what why is that? What
Eric Medecke: Well, I'll be honest, I think a lot of the people's AI initiate initiatives get stopped before they can even start because They find out that the infrastructure that they have doesn't really support everything. The infrastructures are aging. We talked a little bit about, you know, the people that have made the plants work today. You know, they've been able to keep the infrastructures, you know, piecemeal together, we'll call it. but then, you know, when they try to add more stuff onto these networks, things of that nature, they find out that it's too congested or they can't do it, or the they're out of IP address. processes or like all different types of factors right that we understand and know about when it comes to infrastructure. And so it it puts the brakes on it very, very quickly because they don't have the segmentation, the security, and all the things that are required to be AI ready.
Kevin Romer: That the the infrastructure is one
Load full transcript
of the biggest ones. There's some other stuff where when it comes to the actual PLCs and those type of things, legacy you know, when when you're dealing with legacy equipment, some things can't really handle it. Things are at their limits, you know. All all mostly comes down to infrastructure though.
Jackie Slaght: Yeah, that's well, one of the things this isn't really infrastructure, but I just want to wanna touch on it because I think it's something that that feel like you guys have both brought up already is that part of what manufacturers are dealing with right now is there's a lot of tribal knowledge that's either aging out or they can't kind of get the the new newer people coming in up to speed or they just kind of know if something's not working because they have that experience of years and years of knowledge and kind of making things work. if they can get the infrastructure, how can AI help take that tribal knowledge and bring it to to something that everybody can use?
Kevin Romer: So understanding what's there, starting with like an assessment to to try to build a a where you are right now, just to understand where we are at this point and where can we go in the future and that kind of thing. So a lot of people don't understand where they are right now and an assessment really helps out with doing something like that.
Jackie Slaght: one of the things I think we also talked about in the live session was what does it mean when a manufacturer says thinks or says that they are AI ready? and And what could they be overlooking if they think they're AI ready, but maybe they're not actually quite there yet.
Eric Medecke: I think some of the manufacturers, some of the plants and some of the people we've seen feel like they're AI ready because they think they have the correct infrastructure, they feel like they're on a different path because they've they've made that convergence between IT and O T and and they think that they're they're ready to go. Unfortunately It's the networks are no different than the highway at rush hour, right? We don't want rush hour traffic. Nobody wants to be stuck in rush hour. And we don't want rush hour on our networks in the OT side of the of the of the facilities, right? i and this is where the the the gap is, right? They don't they don't have all the segmentation, they don't have things V Land out properly, they they have high utilization on their networks, things like that that they're just not in tune to. They don't don't realize it. And so they think that they're ready, they think they're AI ready, they think they have connectivity to a lot of things that may exist in their facilities, but they they've not gone down that road of truly assessing things like Kevin mentioned earlier. And if if you have somebody come in and do an assessment, understand what's talking to what what devices are out there and understand what devices may or may not even be connected to your network to give you that ability to get that stuff connected. And that's part of the problem with people who think they're AI ready is they don't have the connectivity that they think they have because they haven't really done a true assessment or evaluation of what's currently out there.
Jackie Slaght: One of the things that I took away from that from the live session was how important it is to, you know, I I think what people want to get out of AI and to to make it as effective as it can be is really having the data set up. And if you don't have good data coming out of it, you can't really take all the advantages of of the results it can show you of like this is what might happen because we have all this data, but if you have equipment that's not on the network or you know, by itself, that's going to give you the wrong information. is that correct?
Kevin Romer: Yeah, so what happens there is basically a lot of systems for that have been put in place over the years have different vendors with different styles of doing different things, right? And a lot of times you'll get different types of metrics or or information that's not doesn't line up with something And so you have to do a lot of post processing to try to line up that data so you have a consistency across the board so that you can actually normalize all the lines or equipment or ha whatever it is that you're doing for a process and understand in the process what is actually the the weak link of the process. So that takes having the right data put in. And sometimes a lot of cases, you know, their plant might be connected and they th they have HMIs and they have OE displays even, but they may not have all the other information that you need to to have a proper AI evaluation of your system.
Jackie Slaght: this is kind of just a fun question, but I I want to see what you guys say. what are the biggest warning signs that facility may be displaying if they are not ready for AI? what would the facility look like?
Eric Medecke: guess warning signs, I don't know. Kevin co can help out here, but I think one of the warning signs is people that have connectivity that's intermittent or warning signs where, you know, hey everything out here in the plant's working, but you know, I don't know, once a month we have to shut things down and cycle the power and hey it's up and running again, you know, and we have that kind of a kind of a thing going on. The things like that anomalies really weird things that happen and it usually ninety nine percent of the time stones back to infrastructure, how things are wired, the types of cabling, how it's done. and that's really the you know, the foundational piece of the puzzle that we've been trying to trying to to explain to people that if you don't have the right infrastructure at the end of the day, you can't build the you can't build the thing.
Kevin Romer: Unexplained downtime is another one. you know, if they can't look back at something and say this is what actually caused that downtime and they
definitively know what that is, why that happened and then how they're gonna improve it, or they work with it and understand that that's what the problem is and they wanna fix it in the future. if they don't have that information, then a lot of times people will just blame the network. 'cause that's something people typically can't see. it's a common common thing to go go to on that. some other flags like like Eric was saying is when you hear one thing from one department but go on site and talk with the people that are actually keeping the system running and you get a completely different picture. We've seen that before too and that's that's kind of a well that that's commu it comes down to human communication in some of those cases and that's where we we try to bring everyone together to understand where everything's at and then what we can do to in the Pew to embrace.
Jackie Slaght: It's one of the things, Kevin. I think you and I like years ago were talking and working on a different project where you I just love the way you explained it that the IT and the OT teams have completely different like languages and priorities that they're trying to to both work towards. And I think where a lot of the challenges come by is that these two teams need to work together but they have very different priorities and just how they communicate so that is i think something else that i believe you guys are still seeing out in the field is that these two teams getting them on the same pages is still a challenge
Eric Medecke: I say it's challenging, but we're seeing more and more people and more and more companies every day collaborating more with their IT departments. where we've had a lot of rub between IT and OT in the past. I think we're starting to see a little more of that because IT is up against all the compliance regulations and things like that that are needed from a cybersecurity standpoint. all the things that they're required to report on to make sure that they can get that cyber insurance for the for the company. And I think you're finding more and more people that are in the OT space finding out that they're starting to learn what IT needs. and OT is learning and an IT is learning the opposite. They're learning what OT's needs are. So there's been cases where we see more collaboration and then there's cases where we still see people who are butting heads, we'll call it, and where we get involved, we understand, we help them, so we understand both sides of the fence at the end of day. And once we understand what the initiatives are of the OT environment And what the things in on their agenda, the things that they want to get done, then we usually ask for a meeting and try to get IT involved. And what we do there is then we start off by understanding and asking IT, you know, what is it that IT is doing to support the plant? What are they up against? and by understanding that and knowing what we know on the IT on the OT side and how to support them.
then we can work to start bridging the gaps because a lot of times the two parties don't understand what each party's agendas are and because that's our expertise, that's how we can easily go after this and help both sides.
Kevin Romer: That that's a really good point is is and it's a really good mindset I think that IT and O T departments they need to get into is is to how can we support each other in being more successful? Because there's a lot of talent on both sides of the fence there and coming together to you know, expand on those talents and really support each other, that really that's that's when things really thrive, you know. And it's it sometimes takes a long time, like I said, to get there. and some of the other things too is is budgets. You know, a lot of times O T is very stringent on budgets, their maintenance budgets, a lot of times they get cut a lot of times and they have to pivot and the why spend money on this all this other stuff that they they they don't need if they're running right now, right? So there's a lot of
Jackie Slaght: Wrong interface.
Kevin Romer: a lot of angles, a lot of talk and those type of things that that people don't understand and and it takes a long time to to kinda get the get those things to to change and and be able to try to work together. And that's the biggest thing. So working together.
Eric Medecke: Right. Like you said, Kevin, like with the whole budget thing, that's where we go back to understanding everything on OT or what their agendas are, help to go in and do those assessments and then help them to put a path forward, whether it's a one year, two year, three year remediation plan, whatever that looks like for them, help to put that together for them and help them to drive their agenda, but also maybe highlight the the the highest risks to the business and to be able to show them and to be able to show them and steer them to say, you know, we understand you want to do X, Y, and Z, but here's your highest risks and these are the things we think you need to do first. And a lot of times it aligns with what you know the plant's trying to do. So then by putting
Jackie Slaght: Yeah, that's it.
Eric Medecke: those roadmaps together and understanding what their agendas are and then helping them to fix that budget or to come up with that path forward, then it's easier to articulate that stuff upstream so that the sea level people can understand it right. You know, they may understand, you know, that we need to get X amount in and X amount out in production. But when it comes to getting in the weeds and understanding all the the things to get there, it's easier for us to be able to articulate it in in that way to make it I guess more palatable and understandable.
Jackie Slaght: I I think that's great. what should an organization evaluate first? The infrastructure, their data quality, cybersecurity, or organizational alignment?
Eric Medecke: So I mean You always want to know where your organizational
alignment starts, right? Are we aligned? Are we not aligned? Understand that, right? And then that helps us to lead into what do we do next and who's going to be the key players that are going to be involved in the next steps. And that next step is like assessing things. And we want to start with our infrastructure. So if you look it up today, it's like 70% or higher network problems are all caused by infrastructure issues. It's because it's cabled wrong, it's done wrong, it's physical, the foundation's built improperly. And so we start there, we start the at the next piece of the puzzle, after we have everybody aligned, we go after, you know, what it is that we need to look at first. And so that is definitely the infrastructure. And so that's the next piece of the puzzle. Jackie, what were the other two or three items you said there? You said organizational alignment, your infrastructure, and then what other piece?
Jackie Slaght: cybersecurity and data quality.
Eric Medecke: Right. So once you understand what your infrastructure is, Then we can start to talk about the cybersecurity piece of the puzzle because as we know, like I said before, IT is up against all the compliance stuff, and we need to make sure things are secure. So security goes hand in hand with what that infrastructure should look like. And then once you get all of the security and all of the infrastructure in place, then we can talk about what it is to get that quality data upstream. But without those foundational pieces of the puzzle, and you start to get and try to get data into that network. You know, you're gonna get you're gonna get missing pieces of the puzzle. It's kind of like the bridge is out at random times. And obviously you know you can't drive across the bridge if the bridge is out. so it's kind of like a drawbridge in a sense. So is the bridge down or is the bridge up? and those are the problems that we run into with poor infrastructure.
Kevin Romer: I couldn't have said it better.
Jackie Slaght: So I think you also sort of touched on this, but what role does standardized data play in successful AI initiatives?
Kevin Romer: So that that's having everything normalized. and and basically each each line, each piece of equipment, you get the same type of information so you can understand which one is is lagging or which one's excelling or you know, where can we invest to improve and those type of things. So without having everything evened out across the board, with understanding you know, the performance data from each type of equipment, you you can't really make a informed decision on w where where to invest that kind thing. So that's that's kinda the data. That one way that's one view of the data. the other view is is intermittent data that's coming in, which is comes down to infrastructure not being there, to support it and getting missing data and those type of things. There's a couple of different forms of data quality there, so
Jackie Slaght: So one of the examples we had in the live session. it was a facility that had all these like cybersecurity protocols and and I think it was the only one where the scenario was actually AI ready. but many organizations still see cybersecurity as a roadblock to innovation because because they need to be secure. So how does cybersecurity actually enable AI adoption and scale?
Kevin Romer: So you're talking about the scale you know, how how to scale up basically, right?
Jackie Slaght: not looking at cybersecurity as a blocker to innovation, but actually just something that is going
Kevin Romer: Yeah, interesting.
Jackie Slaght: to keep your facility secure while also allowing for innovation.
Kevin Romer: Yeah, that's that's kind of a I guess a little bit of a myth a little bit in some cases, right? People people see security as roadblocks because you gotta do all these extra things, right? unfortunately there the security is not always simple. there are some additional things and and a different way to look at things when you are actually just performing daily functions. You know, the things you did f in the nineteen eighties are different today. There's different threats, there's different vulnerabilities and there's you know, there's state sanctioned actors out there, and k yeah, people that are in companies and organizations that are organized that are attacking those type of things and you have to be ready for those type of things. So putting in place layers in security is what really helps identify the anomalies inside those layers before it gets to critical layers like the OT layer itself. And so that that's some of the things that you you want to put in place. it takes a while to get there, right, with with certain things and and again it starts with an assessment. but getting that foundational infrastructure set up to support all those type of things is very, very important. But also when you apply the infrastructure, you need to apply it in a secured way. And so there's there's there's some philosophies and those type of things and what we can do is help simplify how to get there by setting up a roadmap f to success basically. And so that's that's some of the main things is cybersecurity with issues that's one way to look at it. Yeah.
Jackie Slaght: So for a organization that is kind of just beginning or just getting started on their journey to becoming AI ready, what would you say are quick wins or low cost options that they could take to start down this path?
Eric Medecke: I think one of the things that a lot of people can do right out of the gate is to start by separating the IT and the O T with a firewall. It's probably gonna be your least cost effective thing. Even if you don't have all of the segmentation and things like that done in the OT side, you can at least put another firewall in or a couple firewalls in and follow the Purdue model where you have a DMZ, which is a demilitarized zone. And the goal there is that you've got a firewall out there
at the edge coming into your facility that's that's supporting the security for the office area. But then the goal would be to put another firewall of a different brand. different make model than what you have out at the edge to prevent like Kevin was talking about mul to prevent the infiltrations by having multiple layers of different types of technologies because if you continue with the same type of technology you know and you can break down one wall then you can break down the next and the next and so the idea there is to make it as easy for everybody to manage. as possible, but actually make it very difficult for the perpetrators to get through because you're you're changing up the types of security and how you do it across the board, all the way down, all the way into the access layer of the network. and by doing that you gain lots of visibility. And like Kevin said, I'll reiterate it, but when you have those multifaceted forms of security, then you have the ability to detect anomalies faster and to act on them and be proactive rather than it gets through and you're infiltrated and gosh, we're not in proactive mode anymore, we're now in reactive mode and we're down. And what are the down events gonna cost us? And so, as we know, they cost.
Kevin Romer: Another angle to look at things too is what what can you do now with basically having a design in place for a future state. and so if you're gonna add equipment, add a line, make sure you you add that firewall like Eric was talking about. But also anything new that you do, add it into the in in game design that you want to do with the segmentation. Go ahead and segment the things that are new coming in, but leave the other stuff there for the migration plan you want to do over a period of time. So that's another thing you could do right away.
Eric Medecke: Mm-hmm. Yeah, a lot of times, Jackie, coming from our assessments, we're able to make those recommendations on what to do. And usually what happens is we go right into a design phase like Kevin was talking about, and we're able to put that that part of that roadmap is a design. you can't build the house without a design or the building without a design at the end of the day. I mean, yeah, maybe there's some people out there that could do it, but I tell you what, it's gonna be more effective if you've got a literally a design and our design That we do are all validated designs. So we follow industry standards, right? Yes, we're using our expertise and our knowledge to give you recommendations at the end of the day, but we go back to industry standards every single time. Cisco validated designs, which is you know a converged plan wide Ethernet standards with Rockwell and Cisco, and we follow that Purdue model. We take those types of standards, we understand how your networks are laid out today. And we can implement and
Jackie Slaght: Yeah.
Eric Medecke: in integrate the things that are there into those models and those designs to help get you
to a state of secure, we'll call it, without basically the need
Jackie Slaght: So
Eric Medecke: for reinventing the wheel all the time. Now there are some cases where we come in and do an assessment and things like that, and we understand that the infrastructure is not in a very good place, which we find more times than not. But that's okay. we have the ability to put that design together,
Jackie Slaght: Okay.
Eric Medecke: go in and put in new infrastructures
Jackie Slaght: Yeah.
Eric Medecke: a lot of times side by side in facilities while they're running, therefore they don't incur any downtime, and then we work with them to create a migration plan of of what areas are we gonna bring onto the network little by little based upon the segmentation, because we all know that even if we've got endless money. that we can't just come in and wave a magic wand and make everything connect right out of the gate. But with those phased approaches, allow the allow the end user to to do things that
Jackie Slaght: Yeah, yeah, yeah.
Eric Medecke: are cost effective as far as upgrades and to be able
Jackie Slaght: that's one.
Eric Medecke: to get them to a more compliant and secure state at the end of the day.
Jackie Slaght: Okay. That's yeah, go. So one of the things I've heard you say in and just working together on several projects, from an ITOT assessment, clients get a a roadmap that is helping them rank what to do first, what to do next, and where they're gonna go. And I either I think we can do them for any sort of length, but I think most common I feel like you guys say is three years. how do you guys help a organization with this kind of roadmap phased approach balance the quick wins with their long term transformation goals?
Kevin Romer: It's it's adaptive, and understanding the company's goals and how fast they wanna do things. budgets, revenue, current state, well there's a lot of things that go into this. even company procedures, human procedures, how we how we actually do things operate on the system, those are some of the things we look at as well. There's a lot of things to look at, right? So And it it it's just understanding what the customers goals and timing for the goals that they wanna have. sometimes like Eric said earlier, compliance is a big one. You know, if you wanna have certain insurance compliance things and you have to hit those by a such and such date, there's your there's your roadmap. You need to need to respond within that period of time, right? So so that's Kind of a way I look at it.
Jackie Slaght: Yeah. So what are some risks that a client may experience if they are trying to implement AI on a kind of unstable foundation or infrastructure?
Kevin Romer: Well, if if you're adding more data to the system, sometimes that can basically create havoc on certain systems. There's there's a lot of systems that we've seen are are are so fragile that just putting a PC on the network, adding that extra l just a little bit of broadcast traffic takes down certain certain elements of the system. some of these are very, very, very fragile. So the risk
Jackie Slaght: Yeah.
Kevin Romer: Without understanding what you have, it's it's it's it's it's risky. So
Jackie Slaght: Well and I assume if it any data that it gives you is gonna be incomplete 'cause it's it wasn't built on a stable foundation.
Kevin Romer: Right. Yeah, that's right. Yeah. So missing data, you know, sporadic data, if you you know. A lot a lot of these older systems, the focus was not the network. The focus was production. that's all it was. And so you have a lot of network sprawl, that's what we call it. When you're just continually adding more things to the problem. Okay. And
Jackie Slaght: Yeah.
Kevin Romer: and the same the same network, same subnet, those type of things. and and it gets it gets out of control over over a period time. And it's not until, you know and and the maintenance maintenance are trying to keep these things running, right? They're doing all these things, jumping through all kinds of hoops to keep things going. And they need proper I would say design. understand how best to add those things in the future because things over time have changed significantly.
Jackie Slaght: yeah, absolutely. I guess I want to just make sure that like we I think understand that all of these facilities have as you said they're running for production and have were originally developed before this technology existed. So now we're I assume most manufacturers are kind of in this same situation of like Well, this is what we had, this is what worked, but now we have this technology that we want to take advantage of, but I guess doing it the the correct way so you can take take the most advantage of it is is important. what is the most common misconception companies have about being AI ready?
Eric Medecke: misconception I don't necessarily know like what the misconceptions would be like other than like some of the things we've already talked about. Like, you know, they they think whether they think they're they're AI ready when they're really not. I think that's probably the biggest misconception is they they feel like they're ready to go and they and they're just not there. I mean, they're playing like Jenga out in the plant and
Jackie Slaght: Ha ha ha.
Eric Medecke: you know, eventually it gets to a point where, you know, the thing's gonna fall over. gonna crash at the end of the day. I feel like that's what's going on and I and I feel like they don't all I don't think there's a lot of people out there that don't understand what's going on in their OT environment as long as they're making widgets and and they're wanting all of this AI, they're wanting all the reporting, the analytics, which rightfully so. You know, with that information we're gonna be able to produce more. And that's the goal at the end of the day. you know it's it's it's difficult. We find it very difficult to come up to management and say, hey, you
know, we need to replace some of these obsolete items and things like that. And it's easy for management to say, well, it's working. Do you have a spare? I know it's obsolete, but by changing it, is it going to make us any more widgets? Or is it going to make more widgets faster? What's the return on investment? And sometimes when it comes to networks and things like that, there hasn't been a whole lot of return on investment for it until now, where now we can try to get a plant where it's AI ready or where we're starting to use AI to be have predictive maintenance, things like that, where we don't necessarily need you know, individuals watching the stuff.
Jackie Slaght: Now you control one.
Eric Medecke: So considering we don't have the the staffing, then the AI can actually come in and and kind of pick up the slack a little bit, some of that tribal knowledge that it's that we can't find anymore because they're it's aging out.
Jackie Slaght: what's a common mistake teams should avoid during an ITOT assessment?
Kevin Romer: you you're saying the people administering the assessments or someone trying to do it themselves? I was Yeah, I was I was gonna
Jackie Slaght: I think either way.
Kevin Romer: say yeah, so lot this has to do with experience 'cause we we may have done these things or a few times in the past, but it's understanding the equipment that's there. You know, looking for things like PLC fives, Slick five fives, older equipment, even some TIA stuff. Th there's just different types of equipment you gotta look at and understand how it's connected to the network. And know if you can do a scan, those type of things. some of the other things is scanning tools that are used, a lot of people may already have some experience with having an IT scan on their OT network and it takes the whole system down, right? That's some of the the big hurdles that people are are concerned with. And so when you when you go to touch an OT system, it's very important to understand what it is you're working with, what what's there before you even touch it, you know? You have to kinda look at that and and evaluate things beforehand and say, Is it worth the risk to scan it? Yeah, we've we've gone to a lot of systems where it's like, you know, don't touch it. Yeah.
Jackie Slaght: Yeah.
Kevin Romer: We've seen we've seen full PLC five systems with like eight p eight PLC fives on there with all kinds of remote IO and everything and they're having all these issues, like I can't add one one more thing to that system and let alone try to do a scan. it's just not gonna work. But we already know what we know by doing by going out and taking a look at it, seeing what we see, we can already assess that. So there's the the the the way you assess things depends on the system that you have in front of you and and what you're you can you can do in those type of things. downtime, you know, if the system's running. Most of the time when we're working with systems, it's always running. It's always up. Yeah, most of the time a OT systems up. So we have to be very conscious of that. If we can get a downtime window, yeah, we can we can scan it. As long as it's not gonna break a a refrigeration system or something like that. That's that's working 'cause we've seen that too. You have to be careful those type of things.
Jackie Slaght: Mm-hmm.
Kevin Romer: Not personal experience. I'm kidding.
Jackie Slaght: Ha ha ha.
Kevin Romer: You you have to be very careful when you're when you're scanning a system and understand what's there.
Jackie Slaght: I yeah, that's it sounds like you have a lot of stories.
Eric Medecke: Yeah, we've seen a lot of different things, Jackie. You know, we come in a lot of times to pick up the pieces where other people have done those types of things that we're talking about because we use specialized tools. Like if we do a scan in an OT environment, it's not the same type of a scan that we do on an IT side of the network. because the scanning tools that we use on the IT side of the network can go out there and and look at all the devices all at once and kind of capture all that information very, very rapidly. However, if we try to use those types of tools on an OT environment and try to ask all the devices who they are, what they are, where they're located, and all the information all at once. Everybody's gonna r try to report back in and everything's gonna go haywire and it's gonna crash that network and it's gonna stop production almost every single time and we we're called in many, many times to come in and and and fix the problem because somebody has unleashed the kraken on the O T network and
Jackie Slaght: Yeah.
Eric Medecke: and hasn't stopped it. And so it keeps causing problems and so we come in and we find these these things that Kevin's talking about now where, you know, we gotta be very cautious and And we've seen things like refrigeration things been down or ammonia systems be down or just all kinds
Jackie Slaght: Mm-hmm.
Eric Medecke: of things and and and you know, there's dangers, there's risks to the business, there's risks to people. so our engineers are very, very, very skilled in understanding all of those risks and all of the hurdles and things that we need to know before we even go in and touch anything.
Jackie Slaght: I feel like we've talked talked a lot about kind of how to get started and kind of beginning with an assessment and getting that foundation ready so that you can officially be AI ready and ready to kind of take advantage of what AI can bring to a facility. But looking ahead to the kind of the generation that's coming into the workforce now, what do you think would be good kind of practical training or skills that that someone just entering the workforce should should focus on so they can I guess kind of either work in an OT environment or take advantage of AI. I'll I'll leave it open ended and however you would like to answer the question.
Eric Medecke: Yeah, I
mean, I don't know, Kevin's probably got his own thoughts on this too, but I mean if you're you're getting into this field and you're you're thinking about, you know, what are the things that I could do to help people get get to an AI ready state, I think the biggest thing is we harp on infrastructure. And if you're getting into the IT space at all, you need to understand the differences between IT infrastructure and OT infrastructure. And they are very different things, very different animals at the end of the day. The types of wiring, the types of things that we do in the OT environment do not exist in the IT environment. a lot of people think that they can run cable because they run cable or done things in the IT environment. And then when you get out there in the OT world and then they understand it later and they're like, my gosh, I learned all these different things and I'm so glad I didn't do that, or they learn trial by fire. Where they've gone out, infrast put an infrastructure in. We had a we had a particular customer who I know very well that actually had gone in and and their IT folks had cabled everything and they were very, very proud of it. But they kept running into all these problems and these anomalies. And we went in and showed them and then gave them a better understanding of what was really going on. And then they really once they understood it, they were like my goodness, I wish we would have really engaged you guys sooner because we had to now spend a lot more money after they had already spent a good considerable amount of money to go back in and and redo everything that they had done just the year prior. So a lot of a lot of costs incurred by not hiring the right people or even at least engaging the right people to start to have the conversation. think that is that's that's that's a big deal, right?
Kevin Romer: An another another perspective is if you're coming to work for us is is adopt adopt change. Be ready to understand things and change and and like Eric said, there's a lot of different ways to do things, but looking at the standards, the reason standards exist is because it eliminates a a lot of issues that have been learned over the years by many people that you know these standards, TAA standards organizations, those type of things exist for a reason. The reason is to try to melt
Jackie Slaght: I missed it
Kevin Romer: make things easier to prevent problems from happening. And that's why it's just it's really important to to kind of understand
Jackie Slaght: Yeah.
Kevin Romer: the s standards. So not just infrastructure standards but security standards and how those security standards affect the infrastructure. because like Eric said, they did a lot of things on that one plant. and they had to pivot later. And and a lot of people think, well, I'm just gonna run some Run some network cable down here and they'll figure it out later. Well, it's real important to understand the beginning of what your goals are with cybersecurity and where you want to go with things because some some infrastructure designs have additional cables required. And if you run a lot of times when you're doing a drop, running two drops, it only costs maybe ten percent more to run the second drop because there's a lot of labor involved to run that drop. So you might think about those things too. There's resiliency, there's redundancy, there's all kinds of things that you you might be interested in for the type of process that you do. Understanding a downtime event what that incurs on the system and those type of things. So putting those costs and and involved with back in the design, you can back that into design and it really kind of helps drive things to get you to the actual in state that's best for you that what works for this system and is in budget as well.
Jackie Slaght: Yes, yeah. Well, that is I I feel like awesome. I feel like I've covered all of my questions around the AI ready or not session. I want to thank you guys for being here. I do have two questions that I want to close on that are my fun questions. and the first one is, what is one AI myth that you wish would disappear?
Eric Medecke: I would say one AI myth that I think is that we can be we can just go out and
Jackie Slaght: We can do.
Eric Medecke: get some AI and just drop it in there and it's gonna solve all of our problems. That's a myth. It's not that simple.
Jackie Slaght: What about you, Kevin?
Kevin Romer: You can replace the entire workforce with AI.
Jackie Slaght: That's a that is a good one too. from what you have seen thus far, what's the coolest use of AI in a manufacturing environment that you've seen recently?
Kevin Romer: Troubleshooting.
Jackie Slaght: it's a good one.
Eric Medecke: Yep. I I think troubleshooting's a big one. I think I saw another one, Kevin, where somebody was making blown glass bottles and they used AI. So it it took a human a lot of time to go in and adjust the twist on the bottle or the shape or the size or whatever the glass bottle was. The environment was horrible because it's hot. and it would take somebody with a lot of experience to know what things to manipulate on machine. And so they took and did an analysis and figured out how to kind of automate it and implement AI to make those changes. And those changes that were taking like, I don't know, I'm gonna guess from memory, an hour or so to make or maybe twenty minutes to make, they were making in just seconds. So if something came off the variation came off just ever so slightly, the AI was actually in this particular test. that was very, very isolated in a very you know, they put a box around it. they 'cause you know, safety and things of that nature, they were actually letting it be like more than predictive AI, right? They were using generative AI to actually make decisions for somebody because it it it created a lot of good things and it allowed the free up that individual to do something else other than be in a dangerous kind of a situation or a A hot situation, right?
Jackie Slaght: Wow, yeah, that that is very cool. okay, I have one final question is what would you like to see us cover on a future episode of Beyond Tech?
Eric Medecke: Like what I wanna see? I don't know. I wanna see maybe something around Like I don't know, you I always want to harp you know, I always want to talk about cybersecurity and compliance, but I mean I think those are big deals at the end of the day. You know, what are the things that you're doing? how are you working, maybe how are you working with your IT department today? how are you how are you bridging the gap? Or the need
Jackie Slaght: Well I think.
Eric Medecke: the need the need to bridge the gap between IT and OT. I think that's probably a bigger topic at the end of the day. because we are, you know, bridging that gap every day and we have to do it to get where we gotta get to. It's it's it just has to happen. So I figure that's that would be good. It's it's not. It really isn't.
Jackie Slaght: Yeah, it's not really an option anymore. These they have to work together.
Eric Medecke: People are starting to realize it slow.
Jackie Slaght: Yeah. Well change is hard. what about you, Kevin? Or do we take all your good ideas?
Kevin Romer: I think I think I agree. Yeah, I agree. I agree that it's it's maybe coming to the table with some of the questions and answers and actually having some information on both sides to support what what things can each department and those type of things support each other on. Those type of things. That's what I would think would be a really good helpful one. Yeah. Collaboration.
Jackie Slaght: Yeah. Well that's great. I well I want to thank you guys for collaborating with me today on this episode. I do not have anything else, but just thank thanks for for spending a an hour with me.
Eric Medecke: No problem. I mean that's what we're here for.
Kevin Romer: Thanks.
Eric Medecke: I mean at the end of the day we wanna try to get the word out. We wanna help people. That's what we love doing.
Kevin Romer: Yep.
Eric Medecke: so it's fun.
Jackie Slaght: Yeah, I I always enjoy talking to to anyone on our team that agrees to come on a podcast episode with me.