The Reality of IT/OT Convergence in Manufacturing with Chris Lloyd (Syspro)

Scott Mann

Welcome to PodMagic, real conversations about solving real IT challenges. I’m your host, Scott Mann, SVP Global Sales at StorMagic. We’re always exploring how simple, reliable technology can benefit you and the people you serve, whether you’re running branch offices, retail stores, or supporting customers on the front line. Our goal is to always bring interesting guests, deliver some value, and have some fun along the way.

Today, that value is all about the collision of two worlds, traditional IT and operational factory technology, what the industry calls IT/OT convergence. It’s a massive challenge for manufacturers trying to modernize without losing control. And our guest today sits right in the center of this transformation, shaping solutions that keep our global supply chains moving. That is Chris Lloyd, Chief Solutions and Technology Officer at Syspro. Chris, great to have you on the show.

Chris Lloyd

Thanks, Scott. great to be here. I really appreciate it and looking forward to the conversation.

Scott Mann

This is a very an interesting one for me, just mostly because of how much of an expert you are. And I feel like I’ve learned a lot just in our preamble. So I’m excited for the guests to hear it.

From my side at StorMagic, we spend a lot of time at the underlying infrastructure, but infrastructure only matters because of those business-critical applications sitting on top of it, especially in that manufacturing world. And for a manufacturing business, application downtime, it equals dollars, a lot of dollars. From your perspective, what are the unique challenges around running essential ERP workloads reliably at remote manufacturing or warehouse sites where you likely don’t have an IT staff, an army of an IT staff on hand?

Chris Lloyd

I think it’s crucial for every manufacturer to consider each of the workflows that drives their business, especially those that have complex sites where they’ve got a distributed center, they’ve got a manufacturer that is remote, that has sometimes not great internet, whatever the case may be. Every manufacturer needs to map out what their workflows are, how important they are, that requires an always-on connection. Such that they can map out if that workflow can work on premise or can indeed be handled on the cloud. And for me, I think that it’s so important that the workflow placement is more crucial than a continued conversation of either we’re an all-cloud business or we’re an all-on-prem business. I think we’ve moved beyond that for manufacturing. I think it’s around the right workflow placement at the right place, especially when you think about manufacturers who often have very deterministic workflows.

That doesn’t require necessarily to have AI, but there’s scanning of something moving inventory from one place to the other. There’s no reason that that has to be overly complicated and it can’t necessarily always be predicated on a cloud connection, because if that doesn’t work for 20 minutes, your shop floor might ground down to zero, as an example. So you need to be very careful about what you run deterministically. And if you want to run that on premise to drive higher resiliency, that’s great. And if you have workflows that are running reports or whatever the case may be, or your setup allows you to have a purely cloud-based operation, then that’s great. Then you use that too. So I don’t think the conversation anymore is purely cloud versus on-prem as a universal, conversation. I think it’s always going to be a conversation about where to place the workflow at the right place, whether it’s deterministic, or if you want heavy AI loads and you want that probabilistic models, then yeah, cloud is going to be the best place.

Scott Mann

I think we’ve always known this. The world has always been hybrid. Edge computing is a buzz term now, but we’ve always had edge workloads. And it’s really just determining what is the most effective place for the workload to run. If you need reliability and it needs to be no downtime and compute needs to be readily available, then you need to bring the workload closer to the thing that needs it as opposed to the cloud. If you need the big data stuff and you want to have that something stored off site, then you need the cloud. That’s the reality of the world we live in.

You said something interesting there with it always on. And in the IT space companies spend a lot of their time measuring downtime and that differs a little bit in the manufacturing space when they’re considering that IoT OT convergence. How should companies look at it differently from that IT versus the OT on the manufacturing floor?

Chris Lloyd

For manufacturers, they don’t often look at downtime as systems connectivity as an example. Their biggest metric for downtime is more what they call OEE, which is Overall Equipment Effectiveness, which is really a percentage of how effective a piece of equipment is for them. So if they’ve got a 10 million dollar machine and it’s only being utilized 40% of the time, that for them would be downtime. If their forklift didn’t arrive on time to move a load on and that machine cannot continue, that is operational downtime for them, more than losing internet connectivity for a reporting engine as an example.

I think what we are seeing in in the world is that manufacturers used to in in many cases have this air gap mentality. So the shop floor was air-gapped, no connectivity. But with IoT, we brought edge very much into the center of consideration and that the shop floor is connected and is instrumented. And that data is vital to manufacturers. So they use IoT devices to tell them what’s happening on the shop floor so that when the planner goes home, they want to understand what happens at night, what happens when you don’t have the visibility. And that’s crucial for them to give that transparency between one site and the other. And that’s what IoT has brought to them.

Scott Mann

I was reading a statistic on downtime in terms of the manufacturing space. In the IT space we always talk about five nines of uptime, that’s crucial. But in a lot of scenarios in the IT world, infrastructure, there usually is a window where things can go down for a certain amount of time and there’s not a cost associated with it. But in the manufacturing space, there’s a direct cost and it’s massive. I was looking at some where it was two hundred and sixty thousand dollars an hour up to twenty thousand a minute if you have any downtime. So the concept of downtime versus always on, it differentiates because of that absurd cost of having any such instance like that, which makes sense why you’re talking about air gapped environments and you can’t be reliant on that network connection.

Chris Lloyd

For manufacturing, the zero downtime is not necessarily the right target. It’s around bounded downtime. It’s predictable, it’s agreed and recoverable downtime because you’re going to have machines that need maintenance at a certain point. But what you don’t want is that maintenance to have a meaningful revenue impact. It must be, as far as you can manage, planned downtime and where you want to limit the unplanned downtime because that has real world impact.

Scott Mann

You kind of answered my next question there a little bit, but maybe you want to expand on it. What’s your advice for meeting that always-on requirement and avoiding that costly production downtime you’re talking about?

Chris Lloyd

The always-on requirement might be something that sometimes drives too much complexities into certain environments. When manufacturers are looking at technology, you should always have a look at the workflow and what recovery time that workflow requires. As I say, if it is a shipping solution or scanning solution, something that’s doing inventory movements or something that’s on the shop floor, you can’t have a half an hour recovery time because that means your shop floor or your distribution center grinds down to zero. A truck won’t move, literally, unless you can scan. So the always-on requirement is great, but for manufacturers it might bring too much complexity. What you want to do is to make sure that your workflows have the right recovery times attached to them and that when you’re running on-prem, it can do that. And when it’s a cloud and can have that sort of recovery time attached to it, running a financial report or can take a couple of hours to recover, there’s no problem. But an inventory movement can’t wait a couple of hours, as an example.

Scott Mann

It’s not like a one size fits all when it comes to always on, it depends on the scenario and it goes back to the cost too. Like you’re talking about a truck sitting, that’s a lot of money that you’re waiting just to have somebody doing something and the most expensive cost in your company is usually the labor force.

Chris Lloyd

Yeah, exactly. I think the always on is maybe a bit of a misnomer for manufacturing.

Scott Mann

It’s more nuanced than maybe a single term that peanut butter spreads across everything.

What does it really mean in the manufacturing space? You’re talking about these different use cases, is it technical uptime that matters? Is it operational continuity? Is it something else? Is there a cool buzzword that I’m not familiar with in the market?

Chris Lloyd

I would go back to the OEE conversation. The three big things we’re seeing in the manufacturing at the moment is, the AI experimentation. Moving from experimentation to real operationalization. The supply chain complexities is through the roof at the moment in manufacturing, and the compliance pressures. So with those three in mind, I would say their biggest buzzword is to make sure that OEE is as high as possible to make sure the equipment is working as effectively as possible, to make sure jobs get out in time, that they can manage supplier lead times properly, they can handle pricing complexity. It is a really trying time for manufacturers at the moment. To bring predictability, to bring recoverability, is crucial for them because they’re under compressing margins. Every tariff impacts them. Every exchange control impacts them. Every compliance rule that comes in impacts them adversely. Manufacturing, especially mid-market manufacturing, is the growth engine of any economy and we seem to forget that.

Scott Mann

You touched on something there and we’re seeing more of that production data move closer to the edge. What are those new risks that are emerging? We’re talking about, like you mentioned, AI. That’s obviously going to present new great risks with data governance that we’re probably not at all prepared for. What are some of those risks that you see that are emerging now that maybe people aren’t thinking about as they’re building out their edge computing AI strategies?

Chris Lloyd

Cybersecurity is a real risk, especially for manufacturers, because the way that manufacturers work, especially in the OT side of things, is if something’s working, don’t change it. In fact, you don’t want to touch it in case it stops working. But it is important to recognize with IoT, with more and more connectivity, bringing AI to the shop floor, it means that you’re opening up manufacturers to more attack vectors, more different ways that hackers and criminals can try and gain access to the crown jewels. That is one challenge, around the cybersecurity. Recommendations to manufacturers are, just because it’s working doesn’t mean it shouldn’t be maintained. You do need to do maintenance, you do need to do patchwork, you do need to change default passwords. You need to make sure your posture is correct, but not only your posture is correct, that your policies are adhered to, changing passwords, maintenance windows, making sure you’ve got a segregated network for your IoT or your OT type infrastructure and your IT based infrastructure. It’s important to maintain those boundaries because we know that the OT side has a lot more attack vectors, let’s put it that way.

Scott Mann

That’s a good point. You almost need that operational autonomy outside in the OT space. It’s outside of the IT world. It needs to be able to operate in conjunction with, but independent of, at the at the same time. It’s different considerations to bring into effect.

You mentioned the shift in mentality that they have too and in the IT space you got Patch Tuesday and you want to update every Tuesday. In the OT space, if it ain’t broke don’t fix it. Don’t mess with it, but that’s a shift that probably needs to happen in the space now in order to create this convergence that we’re trying to go down.

Chris Lloyd

A lot of people on the shop floor, they cringe every time there’s a maintenance window or maintenance weekend, whatever the case may be, maintenance Tuesday, because something might not work. That’s always a risk, but it is crucial. It is crucial because the big movement in manufacturing is the IT/OT unification. To make sure both are safe and trusted is crucial because if you just have a segregated OT world, the best that you can get out of it is great insights. You could get the insights that this bearing might fail in a hundred hours. And that’s great if a solution tells you that. But you need to know that the bearing that is going to fail in a hundred hours, you need all the enterprise context from your IT systems, from your ERP, to know that this is a $100,000 job that that machine’s currently working on, that we have the bearing in stock to fix that machine. We need to know what certification somebody needs to have to fix that machine. We need to know what jobs are planned for that machine and if they can move out to create the maintenance window for ‘Mike’ who’s certified to get the bearings in stock to come and fix this machine such that we don’t miss customer SLAs on the job that we’re currently running and that we can shift the other three out, as an example. When you connect IT and OT data, then you have the intelligent context to drive real decision making as opposed to just an insight of the fact that this machine needs maintenance.

Scott Mann

It’s more than just technology working together. It’s more than just implementing IoT. It’s a cultural shift that needs to happen in that space to be able to adopt it and do it properly and securely and not create headaches for them ten years down the road.

I was doing some research, and it was staggering the percentage of the workforce that are set to retire by 2030. There’s two different studies, one alignment was that edge adoption is at 5% as of 2023 and edge adoption with AI is going to be 60% adoption by 2029. And then directly correlated to that is, you look at the workforce, and 40% of the workforce is set to retire by 2030. So it becomes absolutely critical to figure this out now so that you’re prepared for it. And you’re not prepared for it just to say, “hey, this IoT device can send us great data”. They need to completely operationalize AI so that they can be ahead of this absolute shift that’s going to happen in the workforce.

Looking ahead, like how do you see manufacturing evolving like that into the future?

Chris Lloyd

Let me first talk on the point that you raised, because that is something that has plagued or troubled manufacturers for a long time. They typically rely on three people in the organization. It’s typically the planner, the scheduler, and the supply chain manager. And any real changes in the business have to go past one or all three of those people to really understand the impact of a chain, even if it’s a supplier lead time, as an example. To understand the real impact of the business, you have to chat to those three people, which means that you’ve got the key decision makers of the business wrapped around three people for a massive manufacturer. Those people go home every night and that IP leaves the door, and it’s one of the biggest risks to manufacturing.

The way that I see manufacturing moving, especially with AI, is if you consider that we’ve been through three cycles of technology. Manufacturers have data everywhere, and the unification and bringing together of that data is the first success criteria. The second one is the ability to run a system that runs at the speed of decision making, whether that’s on premise or in the cloud, you need an event-driven system that can operate in real time at the speed of decision making. And now all of a sudden we’ve been handed AI that has both the ability to call probabilistic and deterministic models. When you put those three together, we now have a superpower of, all of the data of the business. We have the right system to drive the right decision and we’ve got AI, which can all of a sudden learn from those three key people in the business and take their skill set and scale it for the organization such that it’s no longer a risk. That’s where I see, especially ERP, meeting the need of manufacturing going forward, to remove the risk around three key man dependencies and having manufacturers scale with that enterprise intelligence behind it.

Scott Mann

The answer isn’t just one thing. It’s a lot of things having to come together into a big picture, which I imagine takes a lot of planning and strategizing around and understanding the businesses. Like you said before, every customer is different. Every piece of equipment is different on what they actually can sacrifice in their RTO. It’s really about understanding the big picture and figuring out how to baby step it to that goal by 2030 or whatever that time frame is that the individual organization has.

Chris Lloyd

Not every manufacturer is in the same place. We have customers that just want to run their manufacturing business. They’re not interested in insourcing crazy amounts of IT skills to create enterprise intelligence, they just want to make sure that certain outcomes are achieved, versus customers that are really trying to embed a technological maturity and to reimagine their workflows using AI. What we’re also seeing in the market at moment is almost a Maslow’s hierarchy of AI adoption. We have customers that want to use AI, ask and interact and do natural language insights. And that’s where a lot of people are in their AI-based maturity. All the way through to creating natural language rules engines, to creating headless agents, all the way to governing your business using AI. Different manufacturers are at different scales of that AI adoption. It’s important that everybody realizes that there is both an AI adoption based on trust in an organization, because the reason you don’t adopt AI is because of trust. There’s also a technological maturity, and you’ve got to realize manufacturers aren’t AI companies and they’re not technology companies. They are real companies just trying to produce a widget. Crucial for me is, to a partner and a company and a vendor that’s providing a solution, is to make sure you understand where manufacturers are on that journey and bring the solution that creates the outcome that they’re specifically looking for. Whether it’s reduced downtime, whether it’s competing against compressing margins. Whatever that business’s outcome is, the infrastructure world, the software vendor world, and the AI world needs to marry up what that manufacturer is looking for. It’s not a one size fits all. It’s not that there’s one answer for every manufacturer.

Scott Mann

It never is, especially AI, every company, every industry, it’s the same scenario. Manufacturing is probably just one of the most complex because there are so many moving pieces. There is so much of a labor force that gets involved in that as well. And it’s not just IT technology that can run it all.

We’ve kind of talked a little bit about the AI space, but if you had a crystal ball, what do you see? What do you see in the future for manufacturing and the IT/OT space?

Chris Lloyd

I definitely do see the IT/OT unification because those manufacturers that get that right within an intelligence context are the manufacturers that are really going to win. It’s the ones that can drive quick decision making on top of a unified context.

I would say to any manufacturer that’s currently looking at software, looking at infrastructure, looking at any different points, be careful of trying to buy a solution because it’s got AI in it. That’s not really the right question. It’s not really what people should be looking for. The question is, will that system give me the full business context to drive my outcomes. It’s not just a feature.

Scott Mann

You’re right. Everybody uses AI. What does AI actually mean in the context of business outcomes for that specific use case that they have? What are they trying to achieve? How does AI actually affect it? It’s not just a buzzword that they’re utilizing to get some better search engine optimization.

The only other buzzword we can explore is data sovereignty, which you did touch on a little bit, but I think we’ve done a good job of covering it.

Chris Lloyd

I do think data sovereignty is going to become more important. We know that ERP and manufacturing is often very overregulated, especially if you’re in the food and beverage space. If you’re in the aerospace or defense space, data sovereignty becomes more and more important. We’ve seen some things in the news, as recently as yesterday, with this sort of US and anthropic dysregulation of the latest AI models. That’s because people have found a way to jailbreak Fable, to weaponize it as a cyber attack tool. I wouldn’t be surprised if data sovereignty and regulation starts getting adopted in different geolocations, driving further complexities in IT and data sovereignty. It’s not just anymore about where the system runs. It’s going to be based on where you can provide your support, where you run your models. Where your IP is located. All of that starts to become more complex, down to even the providers that you’ve got. And your software bit of materials, it could go as far as that.

Scott Mann

That plays into like the conversation we had at the start too, bringing this back full circle. You’re talking about like the on-prem versus the cloud and what does that hybrid model look like. Data sovereignty plays a huge picture in that that as well. It’s not just about you as a business, what’s going to drive you the most, what’s going to be the most operationally efficient way for you to run those systems. It’s also about data sovereignty and geofactors that are outside of all of our control.

You see that in some regions as well. You see manufacturing done in in Africa versus the US versus Europe. It’s very different, and that’s just based off of regulations that they would have in in each of those regions that creates, what the manufacturers are able to do and how they need to do it. That’s going to definitely change over the years to come. That’s not a political conversation that I’m willing or able to get into, but it’ll be an interesting one and how it affects everybody.

Chris Lloyd

For manufacturers, especially in the mid-market space or in the enterprise space, they don’t want to be political experts. They don’t want to necessarily be IT experts. They’re just expecting that they’re buying into a solution that gives them the compliance and the data sovereignty that they need.

Scott Mann

I guess that comes into a big play for you guys at Syspro. As an ERP expert, it’s looking into how you can do this all for the company and take these things off of their hands because it’s not about them being the experts in IT. It’s about them being able to rely on somebody to do all of these things and manage the data governance policies and everything that they need to have and get the business outcomes that they want using the technology that they don’t care to implement themselves. And I guess that’s where a company like Syspro comes into play.

Chris Lloyd

Exactly. I couldn’t have written that better myself.

Scott Mann

I appreciate it, Chris. It’s been fantastic. Thank you very much for joining us today. And thank you all for tuning into PodMagic, where simple, reliable IT meets real world impact. If you enjoyed the conversation, be sure to subscribe and share. And we will see you next time. Thanks again, Chris.