Hi everyone and welcome to a new episode of Great Question, a manufacturing podcast brought to you by Endeavor Business Media and the manufacturing group inside that company. I am Tom Wilk; the chief editor of Industry Week and we are coming to you live today from the IMTS 2026 Trade Show. With us today, our guest is Keith Dunnell, the VP of Manufacturing for ServiceNow. Keith, thanks for being on the podcast.
Keith Dunnell: Thanks, Tom. Yeah, happy to be here.
IW: This is the third day for me now at the trade show and I am picking up an excitement and a certain kind of momentum that I didn't feel two years ago. Are you feeling that in the air?
KD: Yes, A lot of energy here, a lot of excitement, a lot of curiosity too. I've talked to a couple of customers and they're saying, hey, I just got a new assignment. For example, I'm working on operating technology. I got volun-told I need to do it, so I need to learn about it quickly.
IW: For those who are new to ServiceNow, give us a little background on what your company's play in manufacturing is.
KD: Yes, definitely. So, we are really the orchestration layer for manufacturers, and we're helping manufacturers really leverage AI across their enterprise. So, for us, it's about starting with a data foundation. For years, manufacturers have had siloed systems and operated in a fragmented form from department to department.
So, we have what we call the AI Control Tower, which starts by unifying data records so that, hey, if the quality system has a record about a machine, it matches the production system's record about that machine and matches the supplier system's record of that machine. So, it's one asset, described one way, not three ways. And then the Control Tower helps to orchestrate work so that it can take action and help manufacturers both manage their plant maintenance uptime and downtime while optimizing customer orders and customer experience.
IW: That's great because that gives our listeners a foundation for the more in-depth technical questions, we're going to challenge you with today. And let's get started right away with one of the challenges we're hearing about, which is fragmented security and operations data. When A manufacturing leader or customer calls you and says, look, they're dealing with this data fragmentation issue, what's the first thing that you ask them about? Where should they begin to look at the problem?
KD: The first question isn’t, how bad is your risk? The first question should be, can you answer this right now? What changed the last time security fixed something on your production line, right, between the two departments? Most of the time, manufacturers can't answer that question. Security has the ticket in their system. Maintenance has the work order in their system. Again, different systems that don't talk and nobody owns connecting them. So, that's where you start. Not a big transformation. Just pick one recent security fix and one recent maintenance repair and then map what happened. Where's the data? How many systems did it touch? Can you build a complete picture of what changed? It doesn't cost anything and it may take an afternoon. And for most plant managers, that's the moment they realize they've got a problem and a clear way to measure what the problem is.
IW: You mentioned that security ticket is invisible to so many people except for IT or security. That is the truth. I mean, I put my own IT tickets in, they get resolved. I'm not sure where that information is stored, but it's not universally accessible. And that's the kind of information that operations and maintenance could use to figure out, okay, what is going on with the machine.
KD: Exactly. I mean, everyone has the same goal, which is to optimize machine uptime. But again, talking from different sources of data makes it difficult for these departments to actually work together on a line.
IW: I think what's curious too is that sometimes maintenance folks, they sort of inherit the job of patching the machine sometimes, especially in smaller shops, which makes the visibility problem even worse. Because you've got people, that's not really their core competency, it's not going to appear in the work order, but they do it anyway. That's right. So how do we know it got done?
KD: Exactly.
IW: Let's walk through a specific customer story too. Maybe you can pick one sector, whether it's CPG or automotive, where you worked with a customer who was having this exact fragmentation problem. Can you talk about what was broken on their site or what they were working with and what the outcome was when they fixed it?
KD: Yes, definitely. I'll do one for automotive and one for CPG. So automotive example, a tier one supplier has rigorous processes, safety and security departments, both well-staffed. but their operating technology or OT lives on one system, IT security lives on another system, and maintenance records live on a third system. So, there's an exposure on a controller that feeds the inspection line. Security flags it in their system. Maintenance says we can't touch it for three weeks. And production says customer needs parts on Friday. The security folks say we can't wait that long. But then nothing happens, right? The finding sits in a queue, and the conversation has nowhere to go. So again, two different departments talking from two different systems and it doesn't get resolved. But when you can unify the view, something changes. You can see the inspection line is running this order through Thursday and then ask the question, can we absorb a 30-minute maintenance window Friday morning? The answer is yes. Does that get enough time to patch and test before Monday production? And then the answer is yes. So now you have a situation where the CISO says that works and the plant manager says that works.
Again, unifying the data, now they both can see the constraints. And again, the goal is, hey, we want to optimize uptime. So, they're enabled to do that. So, the real outcome is once a plant can see everything clearly, they can find the same pattern repeating everywhere. Instead of finding sitting in queue for weeks, they can actually solve things very quickly. And there's no more disagreement because they're not talking from two different sources of data. They're both looking at the same data.
CPG, same pattern, right? There's a packaging line controller. It's flagged with a known exposure. Security wants to patch it. Production says the line is committed through the end of the week. We can't take the downtime. Same standoff, right? We need to do a patch. We can't afford to take the machine down, talking from two different places. One side sees risk; the other side sees a schedule. Neither has visibility into what the other right is protecting. Unified view conversation changes. Now security can see when that line may have a gap, change over, cleaning cycle, whatever it is. Production can see exactly what's being fixed and why it matters.
Instead of can't touch it versus can't wait, it becomes, here's the window and here's when we're going to do it. Here's how we'll confirm it worked. And again, there's not a queue where things are sitting for weeks. So, it enables the two departments to talk to each other, talking from the same language and working on the same goal, which is optimize uptime.
IW: On that too, a lot of times when you've got executive teams or management teams looking at implementing a solution like that, one of the things they're going to ask is how quick a win can we get? What kind of ROI can we get? In the first 90 days of deploying a solution like this, Keith, What's realistically achievable? How can people who want to implement a solution like this set the table and say, okay, here's what we can get done in those first three months?
KD: Yeah, it's critical to show momentum in 90 days, right? Because if you don't, the initiative can die. So quick wins really fall into 3 categories. The first category is visibility, right? So, running one discovery across the IT and OT environment, to discover what you're running, the equipment that nobody knew existed, shadow systems, third-party vendor access, provisioned years ago and never cleaned up. It's just a quick scan, but it surfaces all these issues. And again, it's across both IT and OT environments, which usually sit in segregated places, but looking at them holistically. So, that's very quick and then identifies, again, from a visibility perspective, tremendous opportunity and where there's risk.
The second piece, data connection. So, take one security finding, run it through the change process and document the whole thing in one place in terms of the equipment involved, what security flagged, what ops changed and what the outcome was and how long did it take. Do it once and you can see friction right in where the current processes are. If you do it three more times, you have a template for the next 20 times. You start to see patterns. So now, again, we're talking about where data needs to be connected, does not connect it. And that you can do that through two weeks, two to four weeks. So now the first part of the 90 days.
The third part, just one fast remediation, one fix to gain momentum. So, pick a security exposure. It's been sitting in queue for months, something that matters and something that's real. Use the new visibility to schedule it into a production window that works. And again, now we're talking about looking across fragmentation, right? So, the production schedule, the path that needs to be fixed, looking at those holistically and figuring out where we can sequence it in. So, it's not moving heaven and earth, it's just taking one example, connecting the data, saying, what can we fix? Go ahead and fix it, and then boom, you have a win. You've just illustrated how taking fragmented data, putting it together, and putting the departments together can get you a quick win.
IW: And you've just unlocked the future of the program too by showing what the capabilities are.
KD: Exactly. Now you start to scale it.
IW: Now you're taking me back. There was a maintenance technician I used to know who said his secret weapon was an ultrasonic probe to measure bearing health. And he said without fail in the 1st 90 days; he would always find something going on with a piece of rotating equipment somewhere that laid the foundation for future investment in the program he wanted to run. But it was that one win that made a difference. And it's got to be quick.
KD: That's right. Exactly. It's like they say when you play golf, it's that one good shot you hit that keeps you coming back.
IW: I'm still looking for that shot, man. Aren’t we all? Well, let's look past the 90-day proof phase. Manufacturers, of course, are going to be looking far down the road for metrics that are going to matter to both them and the wider company. So, what should someone expect to see down the road as things do evolve and change for the better?
KD: There's really two categories of metrics, operational metrics and business metrics. So, from an operational perspective, meantime to remediation dropping, unplanned downtime caused by unmanaged exposure dropping as well. You can run correlation analysis on your downtime logs against your unresolved security findings, and most will find a correlation they didn't expect. When you close the gap, the downtime goes down.
There's also a reduction in time for audit preparation. What used to take months of manual document assembly becomes a couple of weeks of pulling reports. And that's real time back for a compliance team to spend on strategy instead of spreadsheets. On the business side, in terms of metrics, can you absorb more change velocity on the production floor because you're not waiting weeks for security and ops to align? Can you patch faster, refresh equipment faster, add capacity faster? Like that example that we talked about, right? So being able to close the gap between things sitting in the security queue for weeks to actually making the patches.
For some customers, it becomes a throughput gain. And for others, it's operational confidence more than raw numbers. The metric that matters most is the one that I typically ask about. Do your ops team and the security team trust each other's data? Are they talking, right? Because again, eliminating that fragmentation of data enables the two departments to come together and collaborate. And it changes that conversation from a me versus you to how do we do this together? That's the most important thing. It's a culture change.
IW: We'll get you out of here in this one. When it comes to a product like AI Control Tower and looking at data fragmentation as a problem worth solving, in your work with different customers, what would you say is the biggest misunderstanding that you see manufacturers making when they think about moving into this space, into the solution space? What can you help them reorient to change from a misperception to something which is more accurate about what they're about to do?
KD: So, there's two things consistently. The first thing is they think the technical problem is harder than the organizational problem. So, a lot of energy goes into thinking about whether systems can integrate, data models can connect, and all that is real, but that's not the hardest part. The hardest part is the culture change around getting ops and security in a room. Saying, you know, you two are going to work together from one picture of the equipment from now on. That's where the culture change comes. That's where trust is needed, right? Takes a plant manager who says, I'm going to make this happen. Because if either team thinks it's optional, they'll retreat to their own system.
The second is expecting massive transformation overnight. We're going to unify everything tomorrow and everything changes. That's not how manufacturing works. Manufacturing is incremental and evidence based. So again, starting small, first 90 days, pick one thing, see if it works, get it to work, and then start scaling from there. Because manufacturers need to keep producing while they're changing. So, you can't just stop the whole manufacturing floor to do a transformation. You have to keep producing. So, it's important that you prioritize and pick and then scale. Pick one problem, solve it, show it works, then keep going.
IW: Wow. And how many people who are not used to being visible in the company are going to be uncomfortable when you realize, wait, the system is going to help everyone see everyone else work together more efficiently, more cleanly, much less any organizational, not conflicts, but organizational challenges with operations and maintenance where maybe they're used to being friendly adversaries.
KD: Yes, exactly.
IW: This will change that, right?
KD: That's right, exactly. Yeah, it's a real culture change. It's a real culture change. It becomes enterprise-level collaboration, and ultimately that's better for customers.
IW: Do you see also the executive suite observing these changes with sort of a quiet satisfaction? Maybe they've been wondering how to bring teams together differently? When you do a solution like this, are they surprised at that rollover effect? Like, oh, wait a second, here's everyone talking and interacting differently.
KD: I would say the ones that have experience in transformation are not surprised because they understand. It's always a three-legged school, people process and technology, but it starts and ends with the people. Right? It starts and ends with the people. So, the real indicator that is positive is seeing the people behavior and collaborate differently. They understand that really is the litmus test for we've made a sustainable change because you can change technology and change processes, but if the behavior doesn't change, then the transformation may not stick.
IW: Okay. Well, for those who want to learn more about AI Control Tower, where should we point them to?
KD: Servicenow.com. We have a lot of information and messaging around AI Control Tower. It's a real focus for the company. We're looking to help manufacturers really be able to get a competitive advantage in this era of AI.
IW: This is a huge trade show. Times at a premium. Keith, I appreciate the time you spent with us today. Thank you so much for being on the podcast.
KD: Thank you, Tom. My pleasure. I appreciate it.
Keith Dunnell: Thanks, Tom. Yeah, happy to be here.
IW: This is the third day for me now at the trade show and I am picking up an excitement and a certain kind of momentum that I didn't feel two years ago. Are you feeling that in the air?
KD: Yes, A lot of energy here, a lot of excitement, a lot of curiosity too. I've talked to a couple of customers and they're saying, hey, I just got a new assignment. For example, I'm working on operating technology. I got volun-told I need to do it, so I need to learn about it quickly.
IW: For those who are new to ServiceNow, give us a little background on what your company's play in manufacturing is.
KD: Yes, definitely. So, we are really the orchestration layer for manufacturers, and we're helping manufacturers really leverage AI across their enterprise. So, for us, it's about starting with a data foundation. For years, manufacturers have had siloed systems and operated in a fragmented form from department to department.
So, we have what we call the AI Control Tower, which starts by unifying data records so that, hey, if the quality system has a record about a machine, it matches the production system's record about that machine and matches the supplier system's record of that machine. So, it's one asset, described one way, not three ways. And then the Control Tower helps to orchestrate work so that it can take action and help manufacturers both manage their plant maintenance uptime and downtime while optimizing customer orders and customer experience.
IW: That's great because that gives our listeners a foundation for the more in-depth technical questions, we're going to challenge you with today. And let's get started right away with one of the challenges we're hearing about, which is fragmented security and operations data. When A manufacturing leader or customer calls you and says, look, they're dealing with this data fragmentation issue, what's the first thing that you ask them about? Where should they begin to look at the problem?
KD: The first question isn’t, how bad is your risk? The first question should be, can you answer this right now? What changed the last time security fixed something on your production line, right, between the two departments? Most of the time, manufacturers can't answer that question. Security has the ticket in their system. Maintenance has the work order in their system. Again, different systems that don't talk and nobody owns connecting them. So, that's where you start. Not a big transformation. Just pick one recent security fix and one recent maintenance repair and then map what happened. Where's the data? How many systems did it touch? Can you build a complete picture of what changed? It doesn't cost anything and it may take an afternoon. And for most plant managers, that's the moment they realize they've got a problem and a clear way to measure what the problem is.
IW: You mentioned that security ticket is invisible to so many people except for IT or security. That is the truth. I mean, I put my own IT tickets in, they get resolved. I'm not sure where that information is stored, but it's not universally accessible. And that's the kind of information that operations and maintenance could use to figure out, okay, what is going on with the machine.
KD: Exactly. I mean, everyone has the same goal, which is to optimize machine uptime. But again, talking from different sources of data makes it difficult for these departments to actually work together on a line.
IW: I think what's curious too is that sometimes maintenance folks, they sort of inherit the job of patching the machine sometimes, especially in smaller shops, which makes the visibility problem even worse. Because you've got people, that's not really their core competency, it's not going to appear in the work order, but they do it anyway. That's right. So how do we know it got done?
KD: Exactly.
IW: Let's walk through a specific customer story too. Maybe you can pick one sector, whether it's CPG or automotive, where you worked with a customer who was having this exact fragmentation problem. Can you talk about what was broken on their site or what they were working with and what the outcome was when they fixed it?
KD: Yes, definitely. I'll do one for automotive and one for CPG. So automotive example, a tier one supplier has rigorous processes, safety and security departments, both well-staffed. but their operating technology or OT lives on one system, IT security lives on another system, and maintenance records live on a third system. So, there's an exposure on a controller that feeds the inspection line. Security flags it in their system. Maintenance says we can't touch it for three weeks. And production says customer needs parts on Friday. The security folks say we can't wait that long. But then nothing happens, right? The finding sits in a queue, and the conversation has nowhere to go. So again, two different departments talking from two different systems and it doesn't get resolved. But when you can unify the view, something changes. You can see the inspection line is running this order through Thursday and then ask the question, can we absorb a 30-minute maintenance window Friday morning? The answer is yes. Does that get enough time to patch and test before Monday production? And then the answer is yes. So now you have a situation where the CISO says that works and the plant manager says that works.
Again, unifying the data, now they both can see the constraints. And again, the goal is, hey, we want to optimize uptime. So, they're enabled to do that. So, the real outcome is once a plant can see everything clearly, they can find the same pattern repeating everywhere. Instead of finding sitting in queue for weeks, they can actually solve things very quickly. And there's no more disagreement because they're not talking from two different sources of data. They're both looking at the same data.
CPG, same pattern, right? There's a packaging line controller. It's flagged with a known exposure. Security wants to patch it. Production says the line is committed through the end of the week. We can't take the downtime. Same standoff, right? We need to do a patch. We can't afford to take the machine down, talking from two different places. One side sees risk; the other side sees a schedule. Neither has visibility into what the other right is protecting. Unified view conversation changes. Now security can see when that line may have a gap, change over, cleaning cycle, whatever it is. Production can see exactly what's being fixed and why it matters.
Instead of can't touch it versus can't wait, it becomes, here's the window and here's when we're going to do it. Here's how we'll confirm it worked. And again, there's not a queue where things are sitting for weeks. So, it enables the two departments to talk to each other, talking from the same language and working on the same goal, which is optimize uptime.
IW: On that too, a lot of times when you've got executive teams or management teams looking at implementing a solution like that, one of the things they're going to ask is how quick a win can we get? What kind of ROI can we get? In the first 90 days of deploying a solution like this, Keith, What's realistically achievable? How can people who want to implement a solution like this set the table and say, okay, here's what we can get done in those first three months?
KD: Yeah, it's critical to show momentum in 90 days, right? Because if you don't, the initiative can die. So quick wins really fall into 3 categories. The first category is visibility, right? So, running one discovery across the IT and OT environment, to discover what you're running, the equipment that nobody knew existed, shadow systems, third-party vendor access, provisioned years ago and never cleaned up. It's just a quick scan, but it surfaces all these issues. And again, it's across both IT and OT environments, which usually sit in segregated places, but looking at them holistically. So, that's very quick and then identifies, again, from a visibility perspective, tremendous opportunity and where there's risk.
The second piece, data connection. So, take one security finding, run it through the change process and document the whole thing in one place in terms of the equipment involved, what security flagged, what ops changed and what the outcome was and how long did it take. Do it once and you can see friction right in where the current processes are. If you do it three more times, you have a template for the next 20 times. You start to see patterns. So now, again, we're talking about where data needs to be connected, does not connect it. And that you can do that through two weeks, two to four weeks. So now the first part of the 90 days.
The third part, just one fast remediation, one fix to gain momentum. So, pick a security exposure. It's been sitting in queue for months, something that matters and something that's real. Use the new visibility to schedule it into a production window that works. And again, now we're talking about looking across fragmentation, right? So, the production schedule, the path that needs to be fixed, looking at those holistically and figuring out where we can sequence it in. So, it's not moving heaven and earth, it's just taking one example, connecting the data, saying, what can we fix? Go ahead and fix it, and then boom, you have a win. You've just illustrated how taking fragmented data, putting it together, and putting the departments together can get you a quick win.
IW: And you've just unlocked the future of the program too by showing what the capabilities are.
KD: Exactly. Now you start to scale it.
IW: Now you're taking me back. There was a maintenance technician I used to know who said his secret weapon was an ultrasonic probe to measure bearing health. And he said without fail in the 1st 90 days; he would always find something going on with a piece of rotating equipment somewhere that laid the foundation for future investment in the program he wanted to run. But it was that one win that made a difference. And it's got to be quick.
KD: That's right. Exactly. It's like they say when you play golf, it's that one good shot you hit that keeps you coming back.
IW: I'm still looking for that shot, man. Aren’t we all? Well, let's look past the 90-day proof phase. Manufacturers, of course, are going to be looking far down the road for metrics that are going to matter to both them and the wider company. So, what should someone expect to see down the road as things do evolve and change for the better?
KD: There's really two categories of metrics, operational metrics and business metrics. So, from an operational perspective, meantime to remediation dropping, unplanned downtime caused by unmanaged exposure dropping as well. You can run correlation analysis on your downtime logs against your unresolved security findings, and most will find a correlation they didn't expect. When you close the gap, the downtime goes down.
There's also a reduction in time for audit preparation. What used to take months of manual document assembly becomes a couple of weeks of pulling reports. And that's real time back for a compliance team to spend on strategy instead of spreadsheets. On the business side, in terms of metrics, can you absorb more change velocity on the production floor because you're not waiting weeks for security and ops to align? Can you patch faster, refresh equipment faster, add capacity faster? Like that example that we talked about, right? So being able to close the gap between things sitting in the security queue for weeks to actually making the patches.
For some customers, it becomes a throughput gain. And for others, it's operational confidence more than raw numbers. The metric that matters most is the one that I typically ask about. Do your ops team and the security team trust each other's data? Are they talking, right? Because again, eliminating that fragmentation of data enables the two departments to come together and collaborate. And it changes that conversation from a me versus you to how do we do this together? That's the most important thing. It's a culture change.
IW: We'll get you out of here in this one. When it comes to a product like AI Control Tower and looking at data fragmentation as a problem worth solving, in your work with different customers, what would you say is the biggest misunderstanding that you see manufacturers making when they think about moving into this space, into the solution space? What can you help them reorient to change from a misperception to something which is more accurate about what they're about to do?
KD: So, there's two things consistently. The first thing is they think the technical problem is harder than the organizational problem. So, a lot of energy goes into thinking about whether systems can integrate, data models can connect, and all that is real, but that's not the hardest part. The hardest part is the culture change around getting ops and security in a room. Saying, you know, you two are going to work together from one picture of the equipment from now on. That's where the culture change comes. That's where trust is needed, right? Takes a plant manager who says, I'm going to make this happen. Because if either team thinks it's optional, they'll retreat to their own system.
The second is expecting massive transformation overnight. We're going to unify everything tomorrow and everything changes. That's not how manufacturing works. Manufacturing is incremental and evidence based. So again, starting small, first 90 days, pick one thing, see if it works, get it to work, and then start scaling from there. Because manufacturers need to keep producing while they're changing. So, you can't just stop the whole manufacturing floor to do a transformation. You have to keep producing. So, it's important that you prioritize and pick and then scale. Pick one problem, solve it, show it works, then keep going.
IW: Wow. And how many people who are not used to being visible in the company are going to be uncomfortable when you realize, wait, the system is going to help everyone see everyone else work together more efficiently, more cleanly, much less any organizational, not conflicts, but organizational challenges with operations and maintenance where maybe they're used to being friendly adversaries.
KD: Yes, exactly.
IW: This will change that, right?
KD: That's right, exactly. Yeah, it's a real culture change. It's a real culture change. It becomes enterprise-level collaboration, and ultimately that's better for customers.
IW: Do you see also the executive suite observing these changes with sort of a quiet satisfaction? Maybe they've been wondering how to bring teams together differently? When you do a solution like this, are they surprised at that rollover effect? Like, oh, wait a second, here's everyone talking and interacting differently.
KD: I would say the ones that have experience in transformation are not surprised because they understand. It's always a three-legged school, people process and technology, but it starts and ends with the people. Right? It starts and ends with the people. So, the real indicator that is positive is seeing the people behavior and collaborate differently. They understand that really is the litmus test for we've made a sustainable change because you can change technology and change processes, but if the behavior doesn't change, then the transformation may not stick.
IW: Okay. Well, for those who want to learn more about AI Control Tower, where should we point them to?
KD: Servicenow.com. We have a lot of information and messaging around AI Control Tower. It's a real focus for the company. We're looking to help manufacturers really be able to get a competitive advantage in this era of AI.
IW: This is a huge trade show. Times at a premium. Keith, I appreciate the time you spent with us today. Thank you so much for being on the podcast.
KD: Thank you, Tom. My pleasure. I appreciate it.
