The Future of Construction Data — with Michael Pompey

In our season finale, Fraser and Steve sit down in person in Scottsdale, Arizona with Michael Pompey, Chief AI Strategist at Arrow Electronics, for a candid, no‑fluff conversation about where construction data and AI are really headed.

Across two decades as a CIO, Michael has led digital transformation, infrastructure modernization, and AI strategy — and he brings that experience straight into the construction job trailer. Together, we dig into the realities behind the hype: data silos, ERP limitations, semantic models, data warehouses, and why “10 inches from the keyboard” is still the biggest barrier to innovation.

Michael shares stories from the public sector, insights on biometric data and digital twins, and a blunt take on how leaders will either use AI to elevate their teams… or race straight to the bottom. We explore the future of project management, the irreplaceable value of field expertise, and why construction firms must tie data initiatives directly to business value — not IT wish lists.

If you care about the future of construction technology, practical AI, and how builders can finally unlock the value buried in their own systems, this episode is the perfect capstone to Season 1.

Follow, subscribe, and stay tuned — Season 2 is already in the works.

Watch the video on YouTube: https://youtu.be/a3ocC682B04

Michael Pompey, AI Evangelist, Arrow Electronics: https://www.linkedin.com/in/michaelpompey/

Learn more about Steve’s work with eCMS Cloud Construction ERP Software at  Computer Guidance Corporation or Fraser’s work transforming construction data into actionable insights at Onware.

Transcript

SPEAKER_02 0:11

Hi there, Fraser here. It's Sunday morning and I'm working on editing up the show. And I just wanted to let you know that this was recorded live in person. We're gonna have the YouTube link in the show notes if you'd like to watch the video version. If not, on with the show.

SPEAKER_03 0:30

Welcome to Hard Hats and Data Chats. This is our season finale episode. We are live here in Scottsdale, Arizona. We've been doing this all along. Steve and I have been virtual. We're here in person today. We've we've got a guest. Before we give you a full introduction, I'm just gonna quickly say here that uh, you know, we've spent the season kind of digging into real challenges that the construction teams face. Talked about data silos, forecasting, KPIs, ERP realities, a bunch of different things, data governance, AI readiness, uh, you know, how we can make builders make better decisions with their data. Today we're closing out season one with a conversation about the future. Where is construction data headed? And so to talk about the future, we found a special guest, Michael Pompeii, is here with us. Uh, he's the chief AI strategist at Aero Electronics, helping organizations turn AI strategy into practical business value. He has 20 plus years of experience as a CIO leading digital transformation infrastructure modernization and strategic growth. So we're we're really lucky to have you here with us today.

SPEAKER_00 1:34

I'm glad you're part of the season finale of some of this. This sounds amazing.

SPEAKER_03 1:39

No, it's just me and Steve talking. So now we have someone else to join in.

SPEAKER_00 1:43

So we're gonna have like a we're gonna have like a cliffhanger or something, somebody gonna die. We have to wait for six weeks forever for season two.

SPEAKER_03 1:49

That's all yeah, exactly.

SPEAKER_00 1:51

Gotcha.

SPEAKER_03 1:51

Yeah, gotcha. So we'll have uh kind of maybe Steve will kick this off and uh we'll we'll go from there.

SPEAKER_01 1:57

Yeah, you know, I'd I'd like to start just by talking about ERPs in general. And you know, as much as us ERP software vendors would like to wish, ERPs can't do everything. So, you know, some stuff that's important to projects and managing work and construction, some of that data is not in the ERP. And when it's not in the ERP, you've got a silo, you've got duplication of effort, you've got bottlenecks, just takes extra time. That's the real problem with data silos. And and the best way to tackle those is with integration, good, clean integration. Hopefully it's a restful API or something like that. And data pipelines, data warehouses, data pipelines. So those things are becoming the real differentiators, what we're saying. Sure.

SPEAKER_00 2:48

Sure. So it's interesting hearing you describe that, right? Because as you as you as you're doing that lead-in, it just sort of takes me back to it's it's the classic case of an issue man 10 inches from the keyboard, right? So whether it's you supplying an ERP, whether if it's another hyperscaler coming in with their particular platform, if an organization doesn't take a hard look, 10 inches from the keyboard, they're going to find themselves constrained. NASTAP said it doesn't show up on a balance sheet, it doesn't show up on an RFP, it doesn't show up on a bid proposal, but that 10 inch from the keyboard work never goes away.

SPEAKER_01 3:21

And it, you know, and it's usually pretty critical, whatever that 10 inch from the keyboard's doing or providing to the to the field person, to the project manager.

SPEAKER_03 3:29

And we see so many times like the discussion focuses on software. It does. But it's not software, it's it's what what's your operational process? Like, you know, think about those things in addition to thinking about software. Changing your software is not going to change everything overnight.

SPEAKER_00 3:45

No, but sometimes the software is the easiest thing. It's tangible, something that's on the shelf, right? The CFO can actually see a quote proposal for it, a contract, money is left, right? You can actually see work has been done because you've acquired it. But the other piece, right? Bringing everybody along, talking the line of business, everyone understanding what the data, that's that's the stuff that's it's not as tangible to see, but it's stuff that's definitely required. If not only going to make your your ERP word, but if you actually want to leverage this thing people are talking about called AI, at some point we've got to get to it.

SPEAKER_01 4:16

Well, I'm glad you mentioned that, Michael. And now I I know you've worked with ERPs all of your career.

SPEAKER_00 4:22

Yes. Well, listen, on the public sector side, there's there's so much work that you've got to do, right? You've got so many moving parts. Not only that, but you also have sort of a legal mandated obligation to serve your citizens. And you need something to get it done. I mean, I've there have been times where I've actually had to design ERP systems because there was nothing off the shelf that would get what my organization needed to get accomplished done, as well as tile the data in. So we had to spin up a couple of them on our own. So I listen, I've I've tilted at that windmill for uh for far too long. Aaron Ross Powell, Jr.

SPEAKER_01 4:53

You're coming to us from a unique vantage point. So based on what you've seen, where do you see AI taking us? What is the role of AI with a near P system, especially with construction? Aaron Ross Powell Right.

SPEAKER_00 5:06

So so so here's the thing about AI, right? Because I know everybody loves to talk about it. And and I love too, right? I mean, that's that's that's part of my role. But there's there's also a piece we can't let go of is what are people going to do with it, right? Because AI is not going to do anything by itself. People are going to have to steer this thing in the direction that they want to go. So I can see two possibilities, right? I can see a possibilities where you know leadership decides to be courageous and brave and address some of these things and tell the truth and start leveraging the copious amounts of data that they already have as a true business asset and value, and use these systems to protect their staff, right? Preserve the work that's actually valuable and eliminate the mundane. I can see that. But I can also see a world where folks use this to run straight to the bottom. What's the quickest, most easiest way for me to shave budget? And if that means cut and head count to only hire them 18 months later because the system doesn't work, they will do that because it will look good to the border director. I see that's future happening at the same time. So I park my card over here, right? So I'm a dreamer, I'm a philosopher, I'm a lover, but I gotta admit that there's that other possibility that can also happen with this stuff.

SPEAKER_01 6:16

That leadership variable, which we've talked about a couple of times in our don't get away from it.

SPEAKER_00 6:21

That train is never late. I listen, I wish there was a product that I could sell somebody that would eliminate all those issues because I'd be richer than Elon and I wouldn't lose 16%, I wouldn't lose 16% of my value in a week. All right. Just because that need is actually established. So, yes, that's what I see.

SPEAKER_03 6:39

So that's what we saw like two, three years ago when this whole cycle kicked off. All of the tech conferences were saying, we're doing AI. It's coming, it's it's gonna be great. Then all the executives said, What are we doing with AI? It's taken a good two years for all these tech companies to actually decide, okay, this is what the products are gonna look like, this is how we're gonna incorporate. We're starting to see like, you know, these co-pilots and these tools being embedded in our tools. So, like, especially like, you know, I was saying in Word now, I can select some text in Word and ask it to make some changes to that text. And I'm not copying and pasting between two tools as much as before, seeing new tools coming out where I can just ask about my data directly within the tool. When you look at the construction industry, construction data, what do you think is an opportunity that the leaders are underestimating or they haven't seen yet?

SPEAKER_00 7:31

Well, you know, it's interesting because I can't tell what every single leader has seen, but you can tell what they feel is important. And the way you tell what a leader feels is important is just where they put their energy, right? Where they put their investments. Um, I think there's a tremendous opportunity and integrating data, right? Not only from your ERP systems, but I've seen some systems where not only have they done that, they've integrated uh data. We're actually we'll be talking about it in the keynote tomorrow, where they've integrated biometric data from the actual workers at the job site, workers uh using you know, guard hat with 360 cameras, actually filming the site, creating digital twins that they can use with forecasting, delivery, weather to predict when a site might actually have an issue. In theory, this could go to lower insurance rates because you are standing on top of a system that usually is reactive when it comes to injury, where now you're sort of predicting and forecasting, not replacing the safety manager, right? But giving them a set of eyes and ears beforehand, right? I can see it go in that direction. I think there's a tremendous opportunity. Like we were talking about earlier, Fraser, a lot of these organizations are sitting on a tremendous amount of data. You have to, right? You've got shipping deliveries, you've got staffing, you got contractors, you got all these things. But if they're all over the place and you don't tie that together, that's a tremendous amount of insight that you're missing out.

SPEAKER_03 8:44

I think that's where there were misfires at the start, right? Because people thought we're just gonna connect this AI up to our data and and think we're gonna get insights. And now it seems like there's been a backtrack, and now it has gone to like we need to build a data warehouse. We need to like clean up that data, we need to make it, you know, have actual descriptions and like a semantic model so that the tool can actually give us the results we're looking for.

SPEAKER_00 9:07

You know, and it's interesting too, because I because I remember, you know, when when this bell was rung, right? We just come out of the pandemic, right? The world was still sort of shaken off this monument, this event, and then you have Chat GPT, right, November of 2022 coming out, and all the hype, right? Schools are over, homework is over, your job is over, all these different kinds of things. And folks immediately jumped on that bandwagon, not realizing by doing so, right, they were sort of almost poisoning their own pill as it released to data related to what they expect to get out of it, right? And then you see leaders who know they've got to do something, but I don't know what that is. All right, wife's gotta get started, boards. Hey, I don't want to have my organization be the one that's left out, right? So FOMO's taking over. And listen, I hate to say it, but there are a lot of software companies out there that are more than willing to sell you something to bake a buck, even if it doesn't help you. And there are plenty of them out there.

SPEAKER_03 9:58

Or it doesn't even exist yet.

SPEAKER_00 10:00

Hey, I don't know anything about that.

SPEAKER_01 10:03

Solution of the moment.

SPEAKER_00 10:04

You know, good old vaporware.

SPEAKER_01 10:06

Yeah, exactly. You know, I really love the construction vertical market and the entrepreneurial environment that many contractors live in. But one of the things one of the things I think is interesting is the folks managing the work, project managers, the folks out at the job trailer. And as you think about AI and all the people losing their jobs over and that sort of thing, the worry about that, I think managing construction work is one of those areas that won't be replaced by AI. It it will be augmented, I think, that's my opinion, but it won't be replaced. And those people, first of all, they're highly paid, very smart individuals typically. They it's a paperwork-intensive job for the kind of person who needs to be out running the project. I wonder if, you know, I think AI really is gonna for them, the tools for them will evolve more like an assistant, you know, like somebody following them around. Let me research that for you, you know, that kind of thing.

SPEAKER_00 11:04

Well, it's interesting you say that, right? Because, you know, because when this thing happened, you know, the first fear that folks had was, you know, it's just gonna take my job, right? Um, you saw, you know, students who were maybe returning back to college after being remote during the COVID years, saying, is this degree going to be worthless if this tool can do the exact same work? Um, but when you think about where folks started to look, then you saw a tremendous interest in the trades and construction, like, oh, well, guess what? It'll never touch this because AI can't, you know, put in a toilet or AI can't hang in sheetrock. Um, you know, but on the other hand, you've got robotics not too far behind, right, coming through. But the interesting thing about, you know, whether AI can or should replace a person's job is the description you just gave for those employees. You shouldn't want to let go of somebody that's done this work for decades, knows the ins and outs, can tell by feel, right, when something is off. You don't want to lose that. But it takes a particular kind of leader, right, to think that way and to hold that value versus someone who says, hey, I see these as numbers on a sheet. I mean, we see examples now. I mean, uh, my favorite example is uh the example of Salesforce, right? Mark Menhoff, CTO, went on all the uh the shows. Oh, I'm gonna lay off 40,000 of my analysts because AI can do the same job, proceeded to do so, found out guess what, this technology is brand new, isn't as good. Quietly hired many of those back, yeah, right. Um now, hire a lot of them back to train the model to be better, to maybe do that again. I don't know, but that's a value that you I would hope a leader would not want to lose. I would hope he would take the mundane stuff that keeps them from being in the field and give that to the machine.

SPEAKER_01 12:46

Yeah, absolutely.

SPEAKER_00 12:47

That's that's what I hope he would do.

SPEAKER_01 12:48

Because the people we're talking about are also moneymakers. They also do business development. They you know, the reputation of the firm is based on their skills.

SPEAKER_00 12:56

That relationship, that trust.

SPEAKER_01 12:57

That's right.

SPEAKER_00 12:57

Exactly, exactly. And a machine will never give you that.

SPEAKER_03 13:00

Correct. Yeah. I get these emails from sales guys all the time. Right. They're the new rep for our account. It's this AI generated. You can feel it. You you know, well, because it it's just it's totally off. Like the guy the guy hasn't even read the notes on our customer, right?

SPEAKER_00 13:16

If he knew you, he would knew not to send this to you, which is automatic. It's funny you say that because I think, you know, with this technology showing up in so many different places, right? Like you mentioned, Microsoft and embedding it into your applications, and you're starting to see it in and some of the cabs um and some of the you know the different heavy machinery, it's going to force that other piece that it doesn't do well, right? That human value connection, it's gonna bring that to the forefront. And unfortunately, for some folks, they haven't spent enough time honing that.

SPEAKER_02 13:43

Yeah.

SPEAKER_00 13:44

But the person that can hone that, right, the thing that makes us really valuable and can leverage this piece in a place where it needs to be leveraged, is gonna write their ticket. They are just going to write their ticket.

SPEAKER_03 13:56

Agreed. We touched on this already. I mentioned initially there was this push, what are we doing with AI? Now we're we're walking things back, we're talking about building the semantic model, building the data warehouse, cleaning up our data.

SPEAKER_00 14:10

Some organizations are some organizations are still having that conversation, right? In fact, they've written the check to the big hyperscalers, got a ton of licenses for their staff, right? Spunt spent you know, a lot of money, now seeing the the the token usage go out the roof, and and it's no longer, you know, what are we doing, right? Yeah, what are we gonna be doing with AI? It's like, wait a minute for all of this stuff.

SPEAKER_03 14:33

For all the work, yeah.

SPEAKER_00 14:34

Yeah, and then and they and some of them are walking it back and saying, you know what, let me take another look at this.

SPEAKER_03 14:40

What what's the what's the value for for what we're doing?

SPEAKER_00 14:42

What were we, you know, and it's interesting because when you unpack some of those folks, if they tell you the truth, so you might need some uh uh you know liberal application at Alcong, um they never really had a plan.

SPEAKER_02 14:52

Yeah.

SPEAKER_00 14:53

They did not have a plan. Or or if they had a plan, it wasn't tied to the reality of what was actually happening in a workplace. Yeah. You know, they thought Janet on the shop four was doing X. They rolled this out and it's like, I haven't been doing that for three years. No one asked.

SPEAKER_03 15:08

Oh, and well, and it's the same thing. You go you go off and you build this data warehouse and you you spend a year and and all these resources. Yeah. And then what? Right? Unless you have a goal in mind from the start.

SPEAKER_00 15:19

Listen, and then and if folks and it's funny listening to you talk about building data warehouses because I've in my career tilted my uh at the windmills of business intelligence and data warehouses and ETLs and data lake house and things like that. And at the end of the day, if line of business doesn't see the value in making sure their data is called the same thing in all the three different places and they don't care and they're fine with it as is, you might be barking up a tree to explain the value of having this thing integrated and self-healing and adaptive.

SPEAKER_03 15:47

They don't understand. That's exactly it. I I talked to a guy, I talked to a guy last night, and that's where their IT went. And they they built all that. Right. And then they gave him access to to the tables eventually. And and he's like, I know exactly what I'm looking for. Like, just give me the access.

SPEAKER_00 16:04

Well, what's funny too is that you know, you build all these systems, right? You have all the integrated dashboards or whatever. And what's the one thing that the users always ask for? I want to dump with a spreadsheet. So I can throw it in Excel. I want to dump so I can throw it in Excel. So again, that that trains me.

SPEAKER_03 16:18

But that's that's that's our job is okay, what are you what are you doing in Excel?

SPEAKER_00 16:21

I'm not meal again. I just want to explore. If you just give me the data, at some point it's like, okay, fine, here, you get a dub, here's CSV. I stripped out all the pie, go to knock yourself out. I'm tired of arguing with YD board. You hear that conversation sometimes happen.

SPEAKER_03 16:33

Yeah, exactly. I've I've had that many times. There we go. Yeah. There we go. What's the one foundational thing construction firms have to fix before they can figure out what the value is and pick their direction?

SPEAKER_00 16:45

So the one foundational thing, it's the same thing I would say to any organization, right? The foundational thing you need to, and I won't even say fix, because it may not be broken. And many of these organizations have been running fine for decades without having to have any of this latest technology. But the one area that I would say you would need to address is tying business value in a known way to what you do so folks can understand. The interesting thing about being a CIO and being a CIO in a public sector is you know, when you're in that position, every dollar that I spent as a CIO was a political dollar. And it was a political because it meant someone else didn't get to spend it. So if I say, hey, I want to do a project for X, and I can't explain how my X is going to benefit their Y, then I just took money out your pocket. But if you can understand why I need to go in this direction, then guess what? We're talking to the CFO, he's doing all the conversation. I need you to give Mike this money because this is going to help me do X. And I think a lot of organizations know they've got processes, they know they've got work. But when it comes time to say, okay, this is tied to this particular business value, that 10 ish from the keyboard work might not have been done enough. Or it may not be well explained to staff so they understand why they're going in this direction and doesn't feel like, okay, great, we got another crazy mic project that I gotta do for some unknown reason. I don't know why. But they've been shielded from the realities of running a business, right? All they gotta do is take care of what's in the shop for.

SPEAKER_03 18:11

Yeah. When we talk to the construction, if it's an IT initiative, it's just an uphill battle. But if it's coming from operations and it's gonna help the margin, it's gonna help the profitability of the jobs, right? It's a whole other discussion.

SPEAKER_00 18:25

Listen, I look, I come from IT, so as I see I I'll tell the truth, I'll tell the IT truth, and I'll tell it to other IT leaders. You don't want it to be your project. Why why would you want it to be your project when you can't control risk? You can't make the line of business do what you want to do, but you want this project? Don't let them have the project. You come along to support it. Let this be a business decision. You know what I'm saying? Don't let it be an IT project. You don't want any IT projects. Don't have them. Just be support for the business.

SPEAKER_01 18:52

Yeah. You know, looking ahead, Michael, you know, from where you sit, what's your vision or prediction, whatever you want to call it, of where all this data the construction companies collect, where where is that headed in terms of AI and its use in the next decade? What what do you think the trends are gonna be?

SPEAKER_00 19:11

So you're gonna see two paths, right? And we and we talked about a little bit earlier, right? So let's take the the positive path, right? So so the organization with the leadership that sees that my data, not the stuff that's in the internet, not the stuff that the big frontier models have been trained on, right? But the stuff that's in my systems provides me value, you're gonna see them leverage that to build and integrate their own tools, right? To bring their own value versus taking their money, throwing it to a large-scale hypervisor uh that's done all the stuff for everybody else and just getting these generic results. But then you see the other path. You see another person that's gonna stick to doing what they're doing, not necessarily leveraging the best of their data and not running as efficiently as they can, lag behind. And we know what happens when you start to lag behind. At some point, the market shows up, and I don't care how big you are, it does not care. You know, uh listen, as as a Gen Xer, I I remember certain, you know, hallmark things from my childhood, right? Things like, you know, Sears and Kmart and all those other kinds of things that were just these bastions of, you know, they they're all gone. My kids don't even know what they are.

SPEAKER_03 20:17

Right. Well, like go like going to Radio Shack to play with the technology.

SPEAKER_00 20:21

Yes, yes. Oh man, you're working something. You know what? Oh man, I need to get a cable. Let me just run down to Radio Shack because nope, not anymore. Not anymore.

SPEAKER_01 20:28

Interesting.

SPEAKER_00 20:28

But exactly. So that's that's what I see. And and and again, I part of my mission and my job is to help, you know, our business partners and their clients understand, you know, not only where AI is going, but where they can actually bring value. And right now, it's gonna be ringing value out of your own data that you have because you know, a lot of folks have moved things in the cloud, which is great for efficiency and things of that nature, but there's still a lot of stuff in your own systems, right? It's hidden in the emails that no one's read. Right. Right. It's in the PDFs, it's in the all the documents. Right. Like like what I like to call the Afghanistan of IT, which is SharePoint, right? It's in it's in no one, no one's ever locked locks that down. I don't care how good you are, no one ever locks that down.

SPEAKER_01 21:07

Can I use that? That's pretty good.

SPEAKER_00 21:08

Just just just give me credit for it. But but no matter where it's at, that's where your values want to come. If you can tie that in together, that can be your superpower. But it takes some foresight and some willing to roll the sleeves to dig in there.

SPEAKER_03 21:20

This is the end of the season. It's been good talking to Steve, you know, one-on-one.

SPEAKER_00 21:24

But this is where Steve gets this is where Steve gets killed off. This is where Steve gets killed off the show.

SPEAKER_03 21:29

This is next level having you here. Tell me that. And so th thank you for joining us and sharing it. Yeah.

SPEAKER_00 21:36

I love it. Like I said, I I do this day in, day out. I love having this conversation. I love, listen, I love the fact that I could tell the truth with the two of you because I think a lot of times, you know, the reality gets overshadowed by the hype as well as the fear. But I think it's in between those where we can actually find some value in a path.

SPEAKER_03 21:52

Well, that's where we're getting past the hype and we're trying to get into like practical, right? Definitely.

SPEAKER_00 21:56

Definitely.

SPEAKER_02 21:58

Very good.

SPEAKER_03 21:59

So, yeah, we're already working on season two. Hopefully, we'll have some some stuff to announce soon on that. So thank you for being here today and uh listening. And please follow, please subscribe, and we'll talk to you more next time.