Rob Spencer has spent more than 27 years at the intersection of construction and technology, starting as a laborer in a warehouse and eventually serving as CIO and Chief Administrative Officer across several large contractors. Today, he runs BuiltTech Strategy Group, where he helps contractors figure out how technology fits into their businesses and where it can actually create value. We get into when contractors should buy software versus build it themselves, where bespoke AI fits alongside systems like Procore and Sage, and how something an employee builds with AI can become a dependable company-wide workflow. We also talk about app fatigue, shadow AI, data and governance, human-in-the-loop workflows, why mid-sized contractors may need outside help building custom AI tools, and what construction technology could look like five or ten years from now.
Rob Spencer has spent more than 27 years inside construction companies, working his way from warehouse laborer to CIO and Chief Administrative Officer. He now runs BuiltTech Strategy Group, advising contractors on technology strategy, systems, AI and digital transformation.
In this conversation, we dig into how contractors should think about build vs. buy, why app fatigue is becoming a problem, where bespoke AI fits alongside established software platforms, and what it takes to turn employee experimentation with AI into dependable company-wide workflows.
We also get into data foundations, human-in-the-loop systems, shadow AI, technology budgets, and why Rob believes mid-sized contractors will increasingly need outside help to build and maintain custom AI tools.
Episode Notes
00:00 — Introduction
02:25 — Rob’s path from warehouse laborer to construction CIO
05:25 — What’s consistent across contractors — and what differs
08:25 — The acceleration of construction technology and growing app fatigue
10:35 — When contractors should buy software vs. build it themselves
14:30 — Getting the field involved in process and technology decisions
17:40 — Connecting technology investment to the top line and bottom line
20:25 — Leading edge vs. bleeding edge
23:25 — How Rob evaluates and pilots new construction technology
27:15 — What construction could look like 10 years from now
31:55 — How much contractors should spend on technology
33:30 — IT, innovation teams and frontline employees
37:40 — Where humans need to stay in the loop with AI
38:30 — Why the data foundation matters
43:00 — Shadow AI and the importance of governance
47:15 — Building an AI strategy around business objectives
50:45 — Giving employees freedom to experiment — with boundaries
53:45 — Turning an employee-built AI tool into a company-wide workflow
55:20 — Why construction may be heading toward fewer software vendors
57:15 — Where bespoke AI fits alongside the major platforms
59:15 — How AI changes the build-vs.-buy equation
1:00:40 — Who will build bespoke AI for mid-sized contractors?
1:01:50 — What does “AI-native contractor” actually mean?
1:03:45 — Rob’s new firm, BuiltTech Strategy Group
1:07:25 — How Rob assesses a contractor’s technology operation
1:09:30 — Repair, integrate or replace existing software?
1:11:00 — Using outside AI firms to build custom solutions
1:14:00 — Rob’s wish list for construction technology
1:16:00 — How contractors can get their data ready for AI
1:18:00 — Why owners should treat technology as part of delivering the project
Learn more about Rob Spencer
BuiltTech Strategy Group: builttechsg.com
LinkedIn: linkedin.com/in/spencer-robert/
Jason Jacobs: [00:00:00] Welcome to BuiltForward, the podcast exploring where construction is heading over the next decade, and how AI and other emerging technologies will and won't transform the industry. I'm Jason Jacobs, a longtime startup founder, investor, and the host. Let's get into it
Today's guest is Rob Spencer, who's spent more than 27 years at the intersection of construction and technology. What makes Rob particularly interesting is that he didn't come into construction from the technology world. He started as a laborer in a warehouse and ultimately worked his way up to CIO and chief administrator roles across several large contractors.
So he's lived through multiple generations of technology adoption from inside the industry. Rob recently started BuiltTech Strategy Group, where he helps contractors figure out how technology fits into their businesses and where it can actually create value. We get into a bunch of questions I've been wrestling with [00:01:00] lately: when contractors should buy software versus building it themselves, where bespoke AI fits alongside systems like Procore and Sage, how you take something an employee builds with AI and turn it into a dependable company-wide workflow, and why Rob thinks mid-size contractors are going to need outside help to build some of this.
We also talk about app fatigue, shadow AI, data and governance, human in the loop workflows, and what construction technology might look like five or 10 years in the future. Rob has spent his career actually doing this work inside construction companies, so I really enjoyed getting the perspective of someone who has lived through the evolution rather than just obser- observing it from the outside.
Here's my conversation with Rob Spencer. Rob, welcome to the show.
Rob Spencer: Thanks, Jason. Appreciate it
Jason Jacobs: Well, psyched for you to be here. Uh, we, we had a, a real good conversation, uh, I think it was two or three weeks ago, and, um, [00:02:00] and you sit in s- such an interesting seat because you've, you know, kind of come up through the ranks in the trades your whole career, and, and you, you come at it from the technology side.
And the technology side is changing so much that, um, I mean, it, it, it's, it-- And then a m- a lot of the people that I see that are trying to address that and kind of spearhead that are outsiders coming in, right? And you're, you're more of an insider who's been here all along, you know, not a fair weather fan.
So, um, uh, yeah, so I think your perspective will be super relevant for me and for listeners too, so thanks for making the time to share it.
Rob Spencer: Thanks for having me. It, it's definitely been an incredible journey getting to this point
Jason Jacobs: Well, for starters, maybe just talk a bit about the, um, about the journey. How'd you find your way into the trades and, and how'd you find your way into the, uh, technology leadership side of the trades?
Rob Spencer: Yeah, I, I wasn't necessarily looking to get into technology. I got out of the Coast Guard, needed a job, and [00:03:00] I started with this construction company just being a laborer in the warehouse. And on the side after hours, you know, my father-in-law had his own consulting technology business in a small town, and he handed me this thick stack of books, and was like, "You really wanna learn, like, go read these things."
It was all the Microsoft ones to go get your MCSE certification. And I did it, and I, you know, sitting there just... Intellectual curiosity was probably what, um, I'll probably say it a few times, but that was the one thing that probably propelled me, was because it was just asking questions and learning. And so anyways, to get onto the tech side while I was working for this construction company in the warehouse, I'd, I'd start fixing the issues in, in the warehouse on their computer, 'cause they had to wait, like, once a month for this guy to come out of Houston to come fix things for them. And then the whole office found out in that small town, and so they had me start doing it. And then the corporate office found out, and like, "Hey, you wanna do this for the whole, you know, corporate network?" And like: Absolutely. Um, and it just propelled from there. You know, the, the company that I was working with at the time, Orion, they were founder-led, and they eventually got into [00:04:00] private equity, and then they got publicly traded.
And through all that, we did, uh, 15, 16 acquisitions with, you know, one of them doubling the, the employee count. I think that one was around 2015. So I spent almost 21 years in the heavy civil side of construction, uh, leading technology, and, and it was a great journey and great transformation. And then I, I stepped out away from that company and, and went to Spaw Glass.
They're a large general contractor in Texas doing a lot of ground-up construction, but they have a civil component as well. Uh, and then, uh, 2021, late '21, I got recruited by a private equity firm, uh, to come in and, and join CAMP. Uh, they're in the multifamily space, heavy in renovation and restoration for apartment complexes all across the United States. Uh, so they wanted that high scalability growth, somebody who'd been there, done that to come in and help them out, uh, through that transaction. And that was [00:05:00] an interesting one, because not only did I get to do technology, but by the year-end, I was tapped to, to, in addition to being the, the chief information officer to, to be the chief administrative officer. So then I got marketing and procurement and risk and legal compliance and safety, several other departments thrown up underneath me. So it was a great learning experience to not just deal with the technology aspect of those departments, but just to see so much more that goes on in the, the organization for construction, uh, dealing with those other areas.
Jason Jacobs: Uh, and now that you've been in several different environments over a long career, um, what's your assessment in terms of how much of the workflows and the, um, ways of doing things are consistent from shop to shop versus different? And what are the things that tend to be consistent, and what are the things that tend to be different?
Rob Spencer: Um, you know, I'd say the, the consistencies are just, you know, project management in general, the, the core basics of you're gonna do your daily notes, you're [00:06:00] gonna do your toolbox talks, and your JSAs for safety. You're gonna do your schedules. Uh, you know, most companies are gonna bid in, in the same or similar format. Um, where, where I think you start to get into some differences is, you know, how one, you know... We'll take Civil for example. You know, it's all about equipment utilization. If the equipment's sitting in the yard, it's not making money, and if it's not making money, then, you know, is it an asset that you really wanna keep? And so, you know, they may see equipment utilization a little bit different than another company, and what rates, you know, mean good and bad. And, uh, so there, there's some differentiators there. Uh, when you get outside of that, there, there's also differences in how they see technology as an investment versus, "Ah, we're just gonna keep the, the lights on, right?
We're just gonna, we're gonna run technology, but we're not really going to be technology." And I think the phrase I heard from a couple CEOs over my career is, you know, "Do we wanna be a a [00:07:00] construction company that does technology, or do we wanna be a technology company that does construction?" And I think in the earlier days, they were construction companies that used technology. With, with the evolution in how things are starting to change, technology is embedded in every facet of the construction industry now. I mean, there's nothing that's done in construction today that doesn't have some form of, of technology. And so I think we're starting to see a little bit more of that.
They're becoming... You know, they may not like it, but they're starting to become technology companies that are, that are doing construction. And it, and it's not just because they want it. I mean, the owners are mandating these things, too. And, and we'll, we may pivot to this a little bit later or put a pin in it, but you know, I, I think that's an interesting conversation to, to dive into is, you know, as it relates to artificial intelligence and the technology, and these things that the GCs and the trades are starting to use, where do, where does ownership step in and start to put some [00:08:00] rules in play as far as what can you use, what can't you use? I mean, the federal government does it to some degree with, like, CMEs to CMMC, and FedRAMP, and regulations and things. But on the private side, are we gonna start seeing some ownership pushback or forcing, whether it's pushback or forcing them to start using more technology and artificial intelligence on the job sites to make them better, to ensure quality control, to make sure that we don't have slippage on, on the schedules and, and the budgets?
Jason Jacobs: And in your mind, when you, when you look backwards versus looking forwards, um, I mean, do you feel like the pace of change on the technology side has remained constant or, uh, or has it meaningfully accelerated over the last several years?
Rob Spencer: I would say it has, has meaningfully accelerated the past probably six years. You know, before 2020, I mean, maybe you get into, like, 2019. COVID, let's, let's take that whole year out, uh, which honestly, there was some significant investments and things [00:09:00] that came out that were really cool during COVID too with the way that they were using artificial inte-intelligence to detect, detect the social distancing and everything inside imagery. But outside of that, I think there's been major investment in technology in the past six, seven years that it, it was really kinda slow and up ramp up, up until that point. Now, that's a double-edged sword in that, you know, you, you go back to the 2017 McKinsey survey, you know, uh, construction only rated higher in, in digital than agriculture and hunting, which was like a slap in the face to me.
But you, you kind of fast-forward after that, and there's been all this investment and all these startups and everything happening. I think the downside to all the advancements is the field is starting to feel app fatigue. It, it's too much, right? And so we gotta, we gotta start to balance things out. And the analogy I like to use when it comes to technology is it's a tool, and you always wanna rightsize the tool for the type of work to make sure that the [00:10:00] crews in, in the field and the back office is the most effective and most ef- most efficient. You don't need a sledgehammer to go drive a nail in a wall, but at the same time, you don't need a claw hammer to go drive, uh, a, a railroad spike in a, in a tie. So rightsizing the tool is probably the, the biggest thing that we could do to, to help enhance the field now. And, and there's too many shiny objects, and everybody's got their own flavor for what they like. We need to, we need to get a little more standardized
Jason Jacobs: Uh, and, and when you've navigated these decisions and, you know, both the defining the strategies and the implementing the strategies in different shops over your career, do you tend to have the, the s- is, is your philosophy consistent in terms of build versus buy and, uh, you know, incumbents versus upstarts?
Does it depend on the workflow? Does it depend on the shop? Like, how have you found that you've, um, navigated that, and how has it varied, um, from, uh, shop to shop, if at all?[00:11:00]
Rob Spencer: Uh, I don't know, that's really varied shop to shop. My, my approach and methodology stays pretty consistent on build versus buy. I mean, when there's, when there's an off-the-shelf solution that matches exactly to what you're doing and how you wanna operate, you're definitely gonna buy. It, it's not worth going and building the customization and the upkeep and, you know, the infrastructure to, to support a build scenario like that. Where build really comes into play is, you know, if it is, if it is revenue driving and you, you just can't do it anywhere else, you gotta go build it. If, if the solutions that are coming off the shelf don't align at all or even close to what you're trying to do, what your internal processes are, then you don't do it.
I mean, when you go back to the people process technology, technology is that last piece that you throw on once you've got great people following a great, well-documented and followed process, because it should be an enhancement. You shouldn't go get a software that makes you f- forces you into redefining [00:12:00] and, and changing your entire processes. If, if so, then that's when you would want to go build because you just don't... can't find something off the shelf. Uh, and we've had several situations like that. I mean, I'll give you an example. You know, Agentforce with, with, uh, Salesforce, I think they call it something else now. They, they're like Microsoft, they like to rebrand every year on, on what their products are called. But, you know, Agentforce just, it wasn't that great for some of the things that we needed to do inside of Salesforce. And so we were looking at, you know, just going and building some simple tools to be able to take natural language processing so we could talk into an app, and it would come back and add notes into their, their, uh, sales opportunity and update fields. It-- for whatever reason, Agentforce couldn't do it, but a simple $5,000 build facilitated that, which not only helped to get the data where we wanted it to be, but it helped with forecasting pipeline and potential revenue and all these other things that were severely lagging because we just couldn't get the sales team to, to keep up with the data in an [00:13:00] effective way. Um, another build scenario was, um, we were doing an integration between Procore and Sage Intacct, and the out-of-the-box integration tool that comes off the shelf, and a lot of companies have used it, and I've heard, uh, the ones that do, they're doing a lot of firefighting. Um, but it, it requires a gatekeeper, somebody to look at the transactions and click a button to say, "Yes, you can move this data between these two systems." That's not an integration when you have to have human in the loop to that extreme. And human in the loop is, is needed at, to some degree. So we went and built our own custom integration between Salesforce-- I mean, between Sage Intacct and Procore that did everything that we needed the way we needed to do it, and there wasn't any reviews.
The only reviews was looking at the logs to see if something was failing and if so, why. And I would say probably 80, 90% of the time it was people not following process. It wasn't an issue with the integration in itself
Jason Jacobs: Uh-huh. And that, that brings up a good [00:14:00] question. When people don't follow process, um, uh, I mean, on the one hand, it's like, well, if they just follow process, it would work, but on the other, it's like, but if you can't get them to follow process, then it's not gonna work. So, um, you know, what have you found in terms of, you know, is it, is it, is it a screening criteria in the hiring process?
Is it carrot? Is it stick? Like, um, uh, when it comes to the workflows around the technology, and, and I, and I would put process, you know, kind of embedded in, in those w- workflows, um, how, how do you marry the two? What's worked? What hasn't worked?
Rob Spencer: I, I think when it comes to process and, and adoption, it comes down to a few key things. You know, one is who defined the process? What was it-- Was the field actually brought in and part of the development of that process that affects them directly, or was this somebody two, three layers up the chain of command that went and said, "This is what I want it to be," without any input, and then [00:15:00] just drives it down on the field? And the reality is the systems or the tools, you know, just can't support it, or they're just not optimized for it. So then the field rejects it, they don't do it, or they try to find workarounds where now they're having to do it in two different systems just so that they can do it the way they need to do it out in the field. But then they, they go and do the double entry into the system just so that, you know, they're checking the box that they did it the way that the, the process says. Um, another area that, that, you know, as it relates to, uh, process adoption is, you know, uh, how well... Let's see if I can get my frame of thought back here. It, it usually relates to newer technology and how long it takes to ramp up on that new technology with people. Because I, I, you know, I'll give a s- perfect example, last company. Uh, we were on, [00:16:00] uh, Dynamics SL and, you know, as soon as I stepped in the door, e- everybody across the company, everybody I talked to, "We gotta get off of this.
It's, it's a manufacturing ERP, it's not for construction." Everybody hated it. They want it off. And so we spend a, you know, inordinate amount of effort getting everything prepped, getting ready, we launch. Well, the reality is, once we launched, they wanted Dynamics because it's what they were used to, it's what they knew, and change is difficult for people. Even though it's, it's a difficult process that existed before, doing something new and different is just inherently difficult to people, right? And so it, it takes that reinforcement and just constant, you know, feedback loops and, and we did that. Like every day there was, there was daily calls, getting that feedback loop, helping them walk through and understand how to do things. And what we, what we'd saw was over 60, 90 days, all of the, the the resistance and everything, I mean, they knew it wasn't gonna go away, but they finally started to realize that, man, this is better. This is [00:17:00] the way that we need to do things, and the old technology d- does need to go by the wayside. So sometimes it's just about comfort more than anything. Um, I mean, again, you-- I, when I, when I do these projects and initiatives to bring in new technology, I always like to get that, you know, 25, 30-year superintendent. You know, these are the guys who've been out in the field, they've done nothing but paper their entire jobs and careers. I want them on my projects when we're l- rolling out new technology, 'cause I want to hear all the reasons why it's not gonna work. I want them to be the naysayers, because at the end, I'm probably gonna convert them into being a champion, and then everybody else is looking at them, and when they're advocating for what you just did and they were a part of it, everybody else starts to get in line and get on board
Jason Jacobs: Uh, I, w- when it comes to prioritizing and choosing what to go after, um, h- how do you [00:18:00] balance, I don't know if this is how you think about it, but I'll call it offense versus defense. Like, like fi- you know, finding ways to, uh, get more efficiency versus putting out f- fires and, and how do those priorities get set?
What, with what frequency do those priorities get set?
Rob Spencer: You know what, the interesting part is that the firefighting will always end up taking priority because it's a disruption. You know, something's not working or s- you know, the system's not, you know, processing things the way they wor- they should, or there's just resistance. So the, the, the tr- the trick is to try to identify those early on and reduce the amount of fires that pop up.
So, you know, good lean IT principles, uh, help out a lot with that, you know, focusing on the value adds because you, you really wanna be thinking about the, the changes to the organization. So as a CIO, I always look at the strategy for the organization. We're sitting [00:19:00] in the meetings, we're having the conversations. You know, I-- As, as funny as it might or simple as it might be, it's like I ask why a lot. You know, it's like, "Well, why do we wanna do it that way? Why does, why does-- Why is that the right approach?" You know. "What made you decide that this was the, the important thing right now when we've got these other three things on the plate?" And wh- when you kinda get through the stories and the whys, one, it helps you from a technology perspective because everybody through the chain of command needs to understand the why, and the why is different to every level of the organization. You gotta get the connection to the people. But the why is also important because then we can start to define out the prioritization of the initiatives. And from a technology perspective, we need to be thinking about the value driver. So when I'm thinking about technology and alignment with the business, everything usually ties back to one of two things. Are we enhancing the top line? Are we protecting the bottom line? So can we help drive more revenue in some form or fashion, whether it's efficiencies in estimating or, [00:20:00] uh, helping to close deals with the sales force and, and making them more efficient, uh, on the bottom line.
You know, how do we, how do we minimize fade? How do we make sure that we come on on time, on schedule, client satisfaction, you know, whether it's, uh, the efficiencies on overhead so that we can reduce the pull through to EBITDA? Whatever it is, those are the two buckets that I look at the most in making sure the technology is aligned with them
Jason Jacobs: Um, when it comes to, you know, early adopter versus laggard or somewhere in between, um, do you have a preferred spot in that innovation curve where you like to be in terms of, um, in terms of making technology decisions and, and integrating new stuff?
Rob Spencer: Yeah, I mean, when it comes to production, I like to be leading edge, not bleeding edge. Uh, when it comes to innovation, I, I do like the bleeding edge side of things. I like to, to, to partner with and try some of these startups, you know, because they're new and, and they've got these great ideas about how to solve some niche problems in, in [00:21:00] the construction industry. And the other thing that you start to get with them is you can kinda help frame and shape a little bit of the roadmap and direction that they're going. Uh, but when it comes to production, you gotta be consistent. You've gotta have good systems and processes in place, and a lot of times these, these bleeding edges just aren't gonna be it. Again, the startups are great. I love working with them, but you never know how long they're gonna last. You know, they, they've got great ideas and products, but if you go and you're throwing that into, you know, your, your $200 million projects or something, and all of a sudden they gotta shut the doors, they get gobbled up by a larger company, and, and now everything changes around everything you built with them for a couple years, you're gonna be in a major situation.
So let them develop and become more mature before you really kinda flip them into your mainstay.
Jason Jacobs: How, how do you weigh an additional login, an additional vendor, an additional PO, an additional diligence, like, you know, risk of staying [00:22:00] around, um, uh, you know, v- versus better product? Like, does a, does a better product always win, or in some cases, is there a better product but not better enough to, to, you know, to not just wait for Procore or Trimble or Autodesk or, or whoever to d- to get there?
Rob Spencer: Yeah. And I know the situ- uh, this answer doesn't always sit well, but it depends. Now, most of the time it, it's gotta have core IT functionality. When I'm talking about that, you know, it's gotta, you know, have certain sign-on capabilities. You want the multi-factor. You wanna make sure that there's tenant isolation.
All these things from a security perspective you gotta have in play when you're talking about, you know, anything that's of significance im- important to the organization. You know, when I'm talking about significance, I'm talking financials, critical operations, scheduling, that type of stuff. Uh, for, for doing your JSAs, for doing your toolbox talks, you know, those types of things, you can, you can have a little bit of [00:23:00] leeway in those because they're not as, as critical to the operations of the business.
It's not gonna directly impact, you know, revenue or profitability to the organization. So you get a little bit of a play there, but most of the time you, you've got a set core set of standards that they've gotta adhere to and be able to do, or they're, they're just gonna sit in a pilot phase or just one-off projects
Jason Jacobs: And when you do experiment, uh, do, um, if, does it tend to be like something that you're solving for where you're actually out shopping and canvassing if there's anything available? Or is it more opportunistic? Does it depend? And, and what does it take f- you know, especially given the vendor fatigue that you talked about before, you know, for, you know, for a new vendor to, to rise above the noise?
Rob Spencer: I'd say probably 80, 90% of the time there, there's a problem we're trying to solve. Uh, whether it's a, a problem with an existing solution we already have in place, it's just not working the way that [00:24:00] we want and we need to, you know, see is there a better solution out there? Or there's a new problem that we have to solve, and is there a niche, you know, software or, or solution that can take care of it? The other 10 to 20% is opportunistic. You don't know what you don't know until you start getting out there and seeing what's possible and, and what some of the new technologies are. And again, technology is running at such a rapid pace and there's new startups all the time. I mean, if, if we kind of stay in our little comfort zones, just like, you know, go back to the superintendents, even for us as technologists, if we stay in that comfort zone and we only deal with what we know all the time, we're never gonna see the possibilities that exist out there. So there's gotta be some experimentation and, you know, chasing some opportunity too
Jason Jacobs: And given all the stories I hear about, you know, shiny demos that, that don't, you know, that don't do, uh, what you thought they were gonna do where the rubber meets the road, um, uh, you know, how do you, um, how do you suss that out? Uh, you know, what [00:25:00] strategies have you found are most effective? And also, um, how do you test it in a way that doesn't introduce core operating risk but still gets you the data you need so that you're, you know, you're confident that it's not gonna be a dud?
Rob Spencer: Yeah, there is no fail-proof strategy to that. But again, when it comes into how you go about your, your RFP process and your selection process for s- uh, picking a, a new solution, you know, we, we typically have well-defined requirements. Uh, and I'm not talking must be able to process payroll, must be able to...
I'm not talking like detailed. I wanna be able to process certified payroll in this form or fashion. I wanna be able to process union payroll in this for- form or fashion. Or you get down into the nitty-gritty on the details and you, you find alignment. And then in your demo scripts, like you call out, "I wanna see it." And if you can't prove it, you know, especially critical items, uh, obviously you can't demo every single thing that you're looking for [00:26:00] because you're gonna spend six months just getting through demos to try to, to try to pick something. So you pick the critical things, and they've gotta show it. Like, and, and I don't wanna see...
You know, most of the time in, in the first iterations of these demos, you know, don't show me a highly customized, your all the shiny bells and whistles and all this crazy stuff. No, I'm gonna, I'm gonna give you a scenario. I'll give you some scrub project data, and I want you to run it through, and I want you to show it to me as if it's one of my projects type thing. And we're gonna, we're gonna flush it out. But even then, you, you're gonna miss something. And so the, the question starts to become, when you get to that point of you- you've selected a solution, you're in the, the design and build and implementation process, and you find that something that was the shiny object that just doesn't pan out the way you want, how quickly can you pivot? So it's always good to kind of have those couple extra solutions in your back pocket just in case something comes up. And, and they're usually the other products you were looking at or, [00:27:00] again, you know, when you're, you're kind of staying opportunistic, you, you know of other solutions that exist to be able to pivot into quickly
Jason Jacobs: Um, so this is, this is more of like a, a higher level question, but when you think about the future of construction, um, you know, if you look out, let's say a decade, um, you know, how, how different is it in terms of how the industry operates? And, and if you had to predict, uh, what things do you think will be the same and what things do you think will be different?
Rob Spencer: I'll start with what's gonna be the same. Uh, what's gonna be same is you're still gonna have people swinging hammers and using shovels. I mean, you're, you're never gonna get people out of the field, and, and we need those people more than ever, and, you know, there's the labor shortage and, and other things that kind of plague the industry that we've gotta, we've gotta change the mindset across, you know, I guess just humanity to, to try to solve [00:28:00] for. You know, and I remember when I was growing up, we had, um, uh, they called them vocational schools, uh, joint vocational schools tied to the high school, and your last two years of high school, you could go learn to be an electrician. You could go learn to be a plumber, or you could go learn to be a carpenter, and you were still doing the rest of your courses for high school that you needed to graduate.
But they were starting to learn the trades at that young age, and for some reason, it feels like we've, we've drifted away from that, and I think that's hurting us as far as what we need for the, the trades out there in the field today. We need to bring that back and, and I think that's gonna help out quite a bit. Where I, where I see things changing is, again, the, the back office is gonna be highly optimized through these systems, through artificial intelligence. You know, we're not gonna need 10 people in accounts payable processing invoices anymore. We're gonna need one or two humans in the loop to validate things because it's just gonna come into play.
Now, it-- for anybody that's worried about, "I'm gonna lose my [00:29:00] jobs," and all that stuff, I mean, I, I love the, the use case of, of IKEA, right? I mean, they had all these frontline people, and they, instead of laying them off like a lot of other companies were doing, they pivoted them into a way to drive revenue. So it doesn't necessarily mean you gotta go lay people off just 'cause you're dealing with artificial intelligence. Can you repurpose them in some form or fashion to, again, enhance the top line, protect the bottom line, uh, just do more, uh, value-added operations to the organization than data processing? So I think that's gonna continue to, to get there. I think we're gonna get a lot more tied in, uh, over the next 10 years with the technology in the field as it relates to the, the scheduling and the budget and the, the pool plans, and all these things kinda stitching together with the imagery and everything on the job site to make sure that, yes, everything is 100% running the way that it should. And we're gonna start getting into more [00:30:00] prescriptive and predictive. You know, the, the AI and the backend systems I think are gonna start analyzing all these data points and be able to produce that report that says, "Hey, um, yeah, your materials didn't show up today. Your crews here are behind." You know, it's gonna point out all these things on the job site and tell you, "Here's the things you need to do to turn it around." Right? 'Cause the reality is, you think about the labor shortages and all the other things that are plaguing the industry, it's not slowing down the timetables that owners are giving you to produce. They're still just as persistent as ever, that they gotta be done faster, they gotta be done cheaper. And so I think the technology is gonna help to optimize those things that...
'Cause y- you think about how much the, the human mind can process and, and deal with at one time, versus technology that's just gonna run all day, every day. And if it can come in and tell you, "Here's the five things you need to do today, and if you do them, you're gonna stay on budget, you're gonna stay on schedule," and you just make sure the superintendent go and execute it. I think that those are some of the [00:31:00] efficiencies are gonna come. There's gonna be a lot more advancements, I think, in, in drones. Uh, I love what they're starting to do finally with, with the automation, with heavy iron, uh, outside of mines. So, uh, I think there's... Trying to remember the job site. It's, uh, Sundt, S-U-N-D-T, is the GC that they implemented some, uh, automation with, uh, their, their yellow iron equipment on a project out here, right?
So it's unmanned, and it's not in the mine. We're finally starting to get to some of that. It's not just a tractor out in the middle of a field. We've got actual production equipment starting to be used.
Jason Jacobs: Uh, I mean, it, it seems like with, you know, with AI, with autonomy, with robotics, with drones, with that, um, that technology is really evolving in terms of what it encompasses in these organizations. So I have a few [00:32:00] questions on that front. For starters, um, how have you historically thought about the technology budget as a percentage of revenues, or, um, or how do you kinda right-size it for an organization?
And, and is it growing given that, um, you know, that more of these firms are starting to, in your words, be, um, technology companies that do construction versus construction companies that do technology?
Rob Spencer: historically, I would say most IT budgets have been around probably about .75 or so of revenue, which is, hey, you're barely gonna be able to keep the lights on at that. Yeah, we can go put out some laptops and desktops. We can get some printers. We'll get some networking and some internet services. But they're not being innovative.
They're not being frontline. They're not really driving anything. They're just providing the barest necessities for the company to operate. I think we've started to see a shift of that, where most companies are starting to fall in probably like 1 to 1.5% right [00:33:00] now. And I think the companies that are truly at the frontlines of technology, they're probably more like 1.5 to 3%. And some of them have even split out and have their own innovation divisions. You know, you look at Suffolk, you look at, um, uh, Rogers Brian, I mean, a lot of these companies, Beck has it. They've got their whole technology arms that that's what they do. And so there's some revenue drivers to some of that, too.
I mean, they're, they're self-sustaining. Uh, but again, it, it started with we're not just gonna stay in this one little bucket of budget and revenue. We're gonna... We believe in technology. We know what it's going to do, so we're gonna spend a little more money on it
Jason Jacobs: You, you actually just touched on an area that I was planning to ask a follow-up question, so I'm glad you, you did, um, which is y- you mentioned, um, some of these innovation teams. So I mean, it used to be that if a company was doing technology stuff, you know, maybe it was called technology, maybe it was called IT, but it was one [00:34:00] team, right?
Um, and now you have an innovation team. You know, some companies, especially if they're PE-backed, they might have a transformation team. Uh, um, and, uh, and yet at the same time, you know, I'm an outsider, so I'm just trying to learn, but what's being drilled into my head by all the insiders is start at the workflows and work backwards.
Start at the workflows and work backwards, and if you just come in from the outside and parachute the latest and greatest technology, it's not gonna work, right? And so, um, you know, h- how do all these departments play together in that world, and should they be under one roof? Should it be separate? Does it, does it depend?
And, and then same question in terms of all these departments and the field.
Rob Spencer: Yeah, that-- I think it depends on the individual company. I mean, it, it's the size and maturity of the organization that probably dictates more than anything. Is, is this, uh... is the innovation team, committee, whatever you wanna call them, are they a makeup of, you know, your CEO, your COO, [00:35:00] your, uh, in most companies, an IT director, your chief information officer, those internal people that are kinda taking the ideas and, and doing something with it? Do you have an innovation team that's made up of frontline people? I mean, this is one of the greatest things I felt like I, I'd done, was we had formed the construction technology group at, at Spaw Glass, where we brought in superintendents and PMs from all the different locations across the company, and we met quarterly, and we'd just talk about like...
I mean, most of it, while I was facilitating, they were, like, interacting and talking amongst themselves, the problems they were having, and what were they doing, and hey, there's this new cool app. And we got a little bit of, of budget to kinda go do and try things. And then there's the ones who are really trying to go advanced and beyond.
I mean, you, you start-- not to kinda digress too much into it, but just you've got, especially HVAC and MEP, I mean, they're, they're almost becoming manufacturers to some degree because they've-- of, of all the prefab that they're starting to do. [00:36:00] So now you've got this marriage between construction and manufacturing with the, the prefab side of things that takes a completely different mindset and kinda innovation team to go out on the floor and produce And then you've got the-- you, you talk about, like, the PE backed.
I mean, technology and everything is a thesis of their investments to their, their LPs out there. They want to see that technology is a driver towards those two things I talked about, top line and bottom line, uh, that, that helps drive the, the valuation and multiple up. So a-again, it depends on... There's a lot of factors that kinda come into play, and then you got publicly traded.
I mean, there's, there's a lot of, you know, SOX and regulations that they gotta follow that, uh, I wouldn't say stifle, but it definitely creates a little more red tape than a privately held company. I think what you're gonna see more than anything are these newer GCs coming up. I mean, these guys, they're, they're younger guys who have started a new company.
They know nothing [00:37:00] but technology, and so they're gonna be a disruptor because their, their whole company is built on artificial intelligence and top, you know, frontline solutions, and they're not on an ERP that they've been on for seven, 10 years now trying to figure out what's next. You know, the, these $500 million, a billion dollar GCs, the-these smaller ones that are coming up that's $30, $50, $100 million, they, they've got the ability to flex and play more because they're not as governed and controlled, which is a good thing.
I mean, they, they need the governance. They need to, to get their processes in line. But they've got more flex to kinda make those kinds of decisions as to what they wanna do. So again, I hate to say it depends again, but you know, every organization's just a little bit different in the maturity cycle and, and the controls in which they can operate in
Jason Jacobs: Uh, you mentioned before this concept of, uh, human in the loop. Um, when it comes to automation, um, what are the things you think are inherently human and [00:38:00] will, and will always be human? What are the things you think that the mach- or the areas you think that the machines can help? And then, um, how do you see that evolving over time?
And, and I have another question. I'm tempted to either let you answer that one, but I'll stick this one in too, which is just h- uh, like how much underlying infrastructure do you need in order for the machines to be able to do anything? For example, like documenting, you know, mapping workflows and centralized data and, you know, like I've heard this term ontology get, get thrown around.
Like, like how... You know, are these just buzzwords or, or is there actually like a whole, um, project just to, you know, just to get the data in a place where it can be usable?
Rob Spencer: I, I think the data is the key to everything, and I've, I've mentioned this on a, I think a couple posts I've done on LinkedIn. But, I mean, your data foundation is the key, just like a foundation of a building. If you don't have a solid foundation, you can throw up walls, you can do everything else you [00:39:00] wanna do, but it's just gonna crumble with a bad foundation. So starting there is the key to success with everything to deal with hyperautomation and artificial intelligence and everything that you're trying to, uh, to get across. So once you have that established, you know, you talked about what things are gonna, you know, be human, what things are gonna be automated. I think there's still gonna be a lot of human in the loop in the near term because we, we just can't trust artificial intelligence to, um, to go post transactions to the ledgers in accounting. We can't trust it to go automatically update a schedule. You know, we can't, uh, trust it to go order materials. We need-- We can have it make the recommendations.
You know, I'm gonna go build... Here's my updates to the schedule. Do you approve it? And a PM can look at it and, you know, fi- make the final decision, "Yes, make those changes." I'm, I'm gonna go book the transactions inside of accounting. Well, you're not gonna post them. You're just gonna create the transactions, and somebody in accounting is gonna review and, and post [00:40:00] and approve. And, and I think a big factor of that, that is the reasoning is it's not just, y- Let's say you got a great data foundation, and you could trust your data 100%. I don't think you can trust the models yet, and and a big reason for that is the models are changing so fast. I mean, you look at, uh, what was it? In, in the past three, four weeks, there's been five new models released, like major models with Astra and, and then you got, uh, Fable 5.1 and, and, every one of those models, nobody's going back and revalidating to see that everything they built, is it still working the way that we said it was supposed to, right?
So, you know, when you, when you think about from a publicly traded company perspective and the controls that you put in play, anytime you upgrade your software, you gotta do testing, and you gotta do validations and controls to make sure that all of your critical functions that operate within that still produce the same results.
And then you can promote it from [00:41:00] your, your test environment to production environment. You know, these, these models, they're just thrown at us, and there's no revalidation of anything that's done there, and they could be breaking things. They could be doing it completely different. I mean, even with... Y- You could still go through and, and do 100 different passes and validate it, but sometimes there's that one time that Claude or ChatGPT or Grok or whatever you're using still comes back and throws some curveball at you.
And you're like, "What are you doing? Like, we've, we've been over this 100 times. I've-- Like, we've written the rules, and you're still now producing something completely wrong from what we've talked about." Uh, so trust but verify, I think, is the biggest thing that, that's still holding us back from 100% automation.
Jason Jacobs: And, um, I mean, it seems like when it's doing what it should, human in the loop means, um, you know, offloading certain tasks and then human giving it a c- you know, a once over and ultimately saving a bunch of time without compromising quality. But if done wrong, it seems like it's [00:42:00] redoing everything and then wondering why you even invested in trying to get it to work in the first place because it would've been more efficient just to do it inefficiently than to do it twice.
So, um, like what determines, uh, you know, when it's done right and, and when it's not? Are there, are there, uh, you know, do you see consistent themes, um, in, uh, when it, when it's set up for success? Um, and then same, do you see consistent themes when it goes sideways?
Rob Spencer: Yeah, I mean, I would say as, as far as like when it's running best, I mean, one, you should never try to automate or throw something into these types of tools unless it's gonna be consistently repetitive. If it's a one-off, there's no reason to go build skills and models and everything else to go do something.
Just do your one-off analysis and call it good. Uh, so definitely, you know, first you gotta identify is it-- we kinda talked about is it worth the time or is it a waste of time? Well, yeah, I mean, could I go write a formula in Excel or could I ask Copilot or something to go write me the formula? And there, there are certain simple tasks like [00:43:00] that. Um, but when it comes to the, the major models and, and the processes in the organization, again, it still comes back to business alignment. You know, the-- I think the challenge that organizations have today as it relates to these things is the individuals and, you know, it, it's still technically shadow IT, but you know, shadow AI is, is like the word that's starting to come a lot more prevalent right now because everybody's got their own version, personal version of these tools. And if, if we're-- if we think they're not using them for company business, we're fooling ourselves. And so we, we've gotta... I think that's the biggest area of probably, uh, concern as it relates to the things that you're, you're hitting on there, is we need the governance and controls to bring it into enterprise control so that we're doing it in the right ways with the right models and the right, uh, systems.
The other thing that comes into play with, you know, the automations and the AI of making it consistent is the data that's being fed into it. Is it using [00:44:00] a four version old document or is it using the latest document? So the gov-- again, you go back to that data foundation and the governance and the controls, uh, is it d- is it using data that was, you know, confidential or classified? I think the-- where you'll see from consistency perspective, it's the ones who are following the rules and the governance and the controls with the right data are gonna be the most consistent. I think the ones that you can trust the least are the ones who's working outside the system and, and not following that
Jason Jacobs: A- as someone who's been doing this a long time, do you feel like, uh, this is, you know, the, like this is the s- y- you've lived this before with another paradigm shift, and so you have kind of a playbook and it's not different, it just, you know, like it's AI, but it's, you know, but, you know, and it's new, but it, you know, it looks just like this shift that came before and this shift that came before and the-- Or do you feel like it's, it's fundamentally different and needs to be treated different than, [00:45:00] um, you know, and that there's no bl- playbook for it?
Rob Spencer: Uh, I would say the, the underlying shift is the same, but the, the technology itself is fundamentally different just because of the possibilities and because everybody thinks they can just throw something in there and, and whatever it produces is, you know, God's truth. And the reality is it, it's not until you've gone through the validation checks, controls, things like that. So to me it's, it's, it's probably one of the more scary times as, as a leader of an organization because you don't know what's being put in front of you. Is it the truth? And you think about, go back to like human in loop. I mean, when somebody was building a spreadsheet, they, they made sure everything was correct.
Every cell had the right formula. When you've got four or five different spreadsheets feeding into a master spreadsheet, everything tied back, they, like, they, they did the work and the effort to do it. Now they just throw it into a model and it produ- produces it, [00:46:00] and you're like, "Here you go." It's like, it, it used to take you six hours.
Now you did this in five minutes. I'm not saying it's not right, but can, can we walk through it to make sure that you checked? And that to me, like I said, that, that's probably the scariest thing is can you, can you trust the documentation of things that are being put in front of you? We had our FP&A team, they were doing a great job and they were starting to use the AI and man, some of the, the graphs and, and the reports and the analysis that was being done was phenomenal.
And they were doing it so much faster than what we could take as, you know, the BI engineers and the data engineers that we had could could produce it. But you know what the issue was? They weren't doing the validations. And so there was data that was being presented as fact, and we're like, "Eh, like I know the data really well.
That, that number doesn't look right. We need to go back and take a look." Whereas from a data engineering and BI engineering perspective, we're doing massive validations to make sure that everything comes out correctly. [00:47:00] And so again, I... I'll, I'll, I'll leave it there
Jason Jacobs: So, so, um, you mentioned that if off-the-shelf is sufficient, like your first look is off-the-shelf, and you also mentioned that you've had some success with custom. I would imagine in the past, and I don't know about your specific shop, but in this industry, that if you're gonna do something custom, it was probably with an outside firm to build it.
Um, and um, so given that, uh, w- w- where off-the-shelf doesn't make sense today, um, where do, um... You know, what, what are your thoughts on the rank and file? So whether it's an estimator, uh, a superintendent, uh, like, you know, anybody in the trenches in, in, in different functions, um, using these tools, is it a good thing?
Should you be striving for more adoption? How do you measure, you know, consistency, controls? Like just, just like where does it [00:48:00] fit? 'Cause we haven't talked a, a lot about that kind of bottoms up, um, AI use. I mean, you, you just touched on it, but, but, um, but I'd love to... You, well, you talked about some of the risks, but it's, it's like, like is it all risk?
Are there benefits? Is it a, is it a strategy? Like how, how do you think about it?
Rob Spencer: It is a strategy. And, and every organization needs to sit down and think about what their AI strategy is and how it aligns to driving the organization forward. Anybody that's just doing AI for the sake of checking a box is doing it wrong. So, you know, just like any major initiative, you gotta take your business alignment.
What are our organizational objectives? You know, what's our revenue targets? What's our, our, you know, EBITDA profitability targets? Where do we wanna be in three, five years? And then can artificial intelligence assist with that? Or are there major pain points that we're facing today that we gotta reduce friction because, you know, maybe it's attrition. Maybe we're losing so many superintendents because we're just making things so difficult in the field [00:49:00] that, that, you know, we get them in and six months later they're out the door, not because they're getting paid, you know, $10,000 more, but because our systems and our processes are so difficult. So, so those are the things that you gotta be looking at from an AI strategy to make sure that you're... And besides, you know, all the, the the data foundations and all the underlying things, just fundamentally at a high level, you gotta look at the organizational alignment. And if, when you get down to, we talked like an estimator, there are some great use cases for artificial intelligence for estimators. But again, do you just turn it on for everybody or do you, do you try to find out like what are the use cases?
Is there a central model that we can go build on? Are we allowing them to go, you know, get prices and, you know, inclusions and exclusions from just a general, you know, personal chatbot that goes out and scours the web? Or do we want to use our internal resources and data to feed into it? And there's a lot of variability that comes into play.
I [00:50:00] think, I think every person in an organization has a valid use case for artificial intelligence today if it's done correctly. You know, again, the fear is that they're going to trust that whatever it goes and finds on the web is true, which, you know, we all know, you know, just because it comes from the internet, it's not true. Um, and where you've got your underlying documentation, you know, is, is that the source of truth? Because again, you could be using a four, four version old document as opposed to the latest version
Jason Jacobs: So when it comes to those security privacy and controls you were talking about to harness some of that bottoms-up innovation without, um, you know, without, wi- without making a mess, um, are there best practices to follow? Like, how do you learn, uh, how to set it up? And then given how quickly the landscape's changing, like you might set it up and be like, "All right, we're good," and then, you know, two months later you gotta do it again because the, you know, because another model came out or, or something else changed in the landscape.
So [00:51:00] like, as a, as a technology leader, what's your advice to technology leaders and how do you think about that?
Rob Spencer: Well, the, I think the best advice as an organization as a whole is you, you gotta give people the freedom to experiment with boundaries. So allowing your frontline people to go try a new app, allowing them to go try the artificial intelligence. But as long as there's certain boundaries that they can play within, you know, you can't take any confidential information, you can't take any intellectual property. But if you're gonna go do it, I wanna report back. Like, I wanna know what's going on. The other approach when, when you don't have that is you gotta get conversational. You gotta be out there in the field, you gotta be interacting with them. And it's, it's crazy the things that you start to learn w- as a, you know, me as a technology leader that I would learn as I'm out there in the field and you're talking to the superintendent like, "Yeah, so you know, I was working in ChatGPT the other day and, you know, I, I took our spec book and I threw it in there, and this is the [00:52:00] fantastic things that it produced for me.
Man, it really greaten the... It shortened the amount of time I had to do to go find these three answers that I was looking for." And when you start to, instead of responding, if you just ask questions, "Well, well, why did you feel the need to do that?" You know, "What, what, what was the outcome you were hoping for versus what you actually produced?" You know, "If we could roll this out across the entire organization, what would you think it would look like?" 'Cause you, you start to unravel what the underlying issue was that they felt the need to have to go do something like that. And it, it's been the same way with shadow IT for, you know, 10, 15, 20 years.
I mean, the... Somebody who's working around the systems and processes, they're not doing it because they wanna be, you know, arrogant or, or be that person that just, you know, fights the system. They're doing it because they've found friction, and they're trying to reduce that friction. They're trying to make their job easier. So we can seek to understand, you know, the underlying reason why, and then, [00:53:00] you know, find out is that an organization-wide problem, and m- can we go change that across the entire company? So that the conversations really start to unravel and uncover the shadow IT, the shadow AI initiatives that individuals are starting to do. And when you, when you have that and you, and you're not going in there with, "Okay, now I'm gonna throw the book at you because you violated the policy and you went and did something." If you go in there with that seek to understand mindset, it, it builds that rapport with them, and now you start to get these advocates and these champions, uh, with the frontline for when you start to get into those major initiatives across the company.
Jason Jacobs: So it used to be that if I was out in the field and I had a pain, I might go and, you know, bug you to allocate resources, either internal resources or, you know, bring in an outside development shop or someone to build me the tool that can solve my problem if there, if it doesn't exist off the shelf. Now people empowered to just use Claude or ChatGPT and build it themselves.
But let's [00:54:00] say it works and it wants to spread across the organization more broadly and, and you want it to actually start getting into critical path with some operating workflows. We need those controls around it. Is there a pass off? Um, you know, like, like what is, what is the process to go from, you know, the equivalent of an MVP, which is just me in the trenches building something for myself, to something that you do want to maybe roll out more broadly?
And, and has that changed from how it used to be?
Rob Spencer: Uh, I feel like it's changed. I mean, again, every company has a, a slightly different, uh, stage of the maturity cycle, but if you've got good change management or even decent change management, I mean, that stuff should surface into there. It gets into... 'Cause when you start talking about rolling across the entire company, you know, what's the total cost of ownership?
Is there any return on investment? You know, how is this gonna help us achieve goals? It starts to kind of start to model into your normal initiative and change management cycles, uh, to, to [00:55:00] get there. Because you gotta make sure, just like with anything, if, if this is gonna be a great solution, you know, what are the underlying data systems it's going to use? Uh, is it developed with the right code and controls in place to protect the organization? What's the adoption cycle looks like? How are we gonna measure success? So all those same things start to come into play. It just goes from this one-off into more of a, a traditional, uh, change cycle that happens in an organization naturally
Jason Jacobs: So directionally, in five or 10 years, are we gonna have, uh, more vendors on average in these shops or less when it comes to software?
Rob Spencer: I feel like we're gonna shrink back down a little bit. I, I think, again, right now we're in this explosive stage where, you know, we're, we're no longer just in front of agriculture and hunting. We're, we're a lot more, uh, on the bleeding edge side in the construction industry with a lot of companies that have come up. But I think at, at, at some point everybody's gonna start to recognize that, um, [00:56:00] it, it's too much. You know, the, these, again, these startups and these, these tech companies, while they're... a lot of them are great, they're... I think they're getting a little bit too much reign. I think we're starting to see, and, and you probably know this a lot better than I do with your background, you know, that the angel investors and the venture capital, they're not just freely giving out as much money without the controls and the things in play that they used to. And so I think that's gonna stifle a little bit of, of these startups and these innova- innovative companies coming up because, you know, they're not just getting $20 million to go play and have fun. Like, they've gotta have, like, a full, you know, um, you know, uh, minimal viable product. They've gotta have everything lined up.
Like, "How are you gonna make us money?" I mean, the, the... E- everybody I've talked to in the startup industry has said it's completely changed, and it's, it's becoming a little more difficult to raise that venture capital. So I, I think that i- in a, in a good way, I think that has helped [00:57:00] not just have every, uh, person out there thinking they can go start up a company and, and just throw technology at the construction industry, because we don't need that.
And, and again, I think that's why we've gotten into this app fatigue kind of mindset with the field. It was just too much. So I think the dialing back, if anything, is gonna help us
Jason Jacobs: And then what about, uh, vendor software versus bespoke, uh, uh, or versus software built in-house?
Rob Spencer: Um, it kind of goes back to that, you know, buy versus build mindset. I mean, there's always gonna be the big players. They're, they're not gonna go anywhere. They're, they're gonna continue to grow. Uh, what's interesting is you, you go back before Procore went public, you know, they were, "We're gonna, we're gonna build, we're not gonna buy." And then, you know, shortly before they went public, they started buying the estimating tool and a few others and rolling them in, and now, I mean, that's, that's really their mindset [00:58:00] is, "We're gonna, we're gonna buy these companies," you know, with the, what, two major ones this year, DataGrid, and then now DroneDeploy. Uh, Autodesk is doing the same thing. Uh, the... And, and the good thing is they're getting better at tucking them in. You know, when, when these products were first brought on or, or bought out, it was, it was still this one-off solution that you, it really didn't play well with their own technology. And so I, I think on the, on these bigger players, they're always gonna be there. Uh, there's, there's always gonna be a need for bespoke, and I think if anything, the, the bespoke is going to come more with artificial intelligence. And I think the reason why is you, you get back into, you know, fundamentally there's, there's core processes across the construction industry, and, but every company has their own certain ways of doing these niche areas, and those niche areas are where the bespoke artificial intelligence solutions are going to come, and that's where [00:59:00] some great companies are already starting to exist out there to, to help with that. You can sit down with them, and you can talk about the pain points and the challenges, and they can go solve for it really quickly
Jason Jacobs: Uh, and, and what about, um, off the sh- you-- what about off-the-shelf versus Uh, you, you hear a lot about how in this age of AI, uh, uh, you know, as the cost of software continues to fall and approach zero, uh, that, you know, that every shop's gonna have tooling that fits them like a glove versus just, um, uh, you know, having to put up with what gets stuffed down their throat from the vendors.
What do you think about that?
Rob Spencer: Uh, again, there's a use case for everything, but, you know, when it comes to the point solutions, every, every technology solution out there is now checking the box and saying they've got artificial intelligence. You know, a lot of it is just automating some rudimentary task that people are having to do. Uh, where the agents really start to come into play is gonna be the, the differentiators, and I [01:00:00] think a lot of that is gonna be more on the build side, because you're gonna need an MCP that can, can tie multiple systems together to do a s- a specific thing in the way that you want it done. There, there's still a, a need for the point solutions, uh, because a superintendent probably doesn't need an MCP to go do something simple inside of a Procore or an Autodesk or a Viewpoint or a CMIC. They need to be able to go, you know, do something quickly in a daily log. They need to be able to, you know, quickly look up, uh, documentation inside of an RFI or a spec book. They can do that inside the point solution 'cause it doesn't need to go anywhere else. But I think your, your PMs, uh, your back office, your estimating, all of these guys are gonna need something that just can't be done natively inside a solution, and that's where the bespoke build type situations are gonna come into play
Jason Jacobs: And then how do those get built? Do they get built by, by the estimator? Do you have an internal team that builds them? Do you use a vendor?
Rob Spencer: I hope [01:01:00] they're using a vendor. And, um, I mean, if they, they've got the, the wherewithal and the resources to bring in, uh, an AI engineer, then that's great. I mean, you-- there was, there was case studies, um, I'm maybe quoting it wrong, but I think it was Suffolk who had talked about they're putting an AI engineer on the projects, and they're sitting in these meetings, and they're listening to what's going on, and they're understanding things, and they're, they're pushing these changes to schedules and everything because, you know, why have the, the, the PM go do it when they can sit there and listen and then just automate right there on the job? But these are, you know, these are multi-billion dollar companies. You know, these are the few in the, the thousands that exist across the construction industry. I think everybody else outside of them, they're gonna need somebody to go build some solutions for them because they're just not gonna have the resources and capital to go hire an AI engineer on every single job site
Jason Jacobs: So the, I mean, the term AI native contractor, if there is [01:02:00] such a thing, is that a BS term or, or does that term mean something to you?
Rob Spencer: I, I think it's something new that's starting to come up. Um, I hadn't seen it in a, until probably the last 30 or 60 days. Um, and I'm, I'm still not quite sure what that means. You know, to me, when I, when I see it, and I, I
kinda
Jason Jacobs: like blockchain.
If you just, if you just, you know, say it at the right time when it's frothy and puff your chest out and act like you know what you're talking about, then, you know, someone will believe you.
Rob Spencer: Yeah. And, you know, I, I haven't heard the word blockchain in a while. Um,
but, you know, the AI native, uh, it- it's probably these, these more startup type GCs and contractors that, again, they're, they're younger, they're technology focused, and they're probably technologists who are doing construction, right? And they're, they're leveraging the technology to make them, you know, a, a better, faster, potentially cheaper GC than what's tr- you know, traditionally been out there, uh, because they are more technology focused and not, [01:03:00] uh, the other way around. So I don't know. It'll be interesting. I, I think it's just... I think it's a buzzword right now.
I don't think it's gonna sustain because owners aren't gonna care. Uh, actually, yeah, I, I go back to, I think right now owners need to be questioning if somebody is an AI native, because I think that raises some risk and exposure that they probably need to question those GCs versus the ones who are tried and true with everything they've been doing. So they might wanna, you know, potentially rethink that, and I don't know. I'm, I'm not an AI native contractor to know is that winning me more work or is it losing me work. But, uh, I would still caution the owners. It, it's something they need to have conversations with companies like that about is, what exactly are you using it for?
How are you using it, and how does it affect me?
Jason Jacobs: Well, I know we're running up on time, Rob, but if we, if you have a few minutes, I'd love to just hear a little bit about, about your new or newish practice
Rob Spencer: Yes, um, appreciate the, the segue there. I mean, [01:04:00] yeah, I've been inside of, of construction companies for over 27 years, and as I, as I stepped away from the last company and I did a little bit of soul-searching on, you know, what's, what's next. I, I wasn't quite sure what I wanted to do, and as I, I did that soul-searching, I thought about all the great things I've gotten to do over my career usually stemmed around from problem-solving and, and helping companies, and especially when we were heavy into M&A, being able to touch new companies and understanding different challenges and problems and how we tuck them into our systems and everything else. And so I, I started, uh, Built Tech Strategy Group. We are a construction native. We, we're not gonna, uh, have any advisory or consulting services to any other industry. I'm, I'm staying with construction companies, GCs, specialty trade, um, really around technology strategy and, and helping organizations that, you know, don't have that expertise in s- in-house, who are, are trying to become that [01:05:00] $30, $50 million company to grow to $100, grow to $250. That $250 million a year company that, you know, wants to hit $500 million. That billion-dollar GC that wants to become a multi-billion dollar GC. And the, the range is, is pretty, pretty wide, but it's all around alignment of technology and making-- you know, helping them to, to set things up with the right foundation, the right approach, a- and the right governance and controls to lead to long-term success and not just short, uh, windfalls
Jason Jacobs: Uh, so, um, s- s- so, um, at the stage you're at now, I know it's relatively new, is each engagement wildly different or do you find that there's a, you know, kind of a sweet spot for what you're doing? And then same question around is it a consistent process from engagement to engagement or is it, is it all over the map depending on what you're doing?
Rob Spencer: Um, you know, what was interesting is I, I thought it would take me probably six months to a year to be where I'm at right now after just a couple months. [01:06:00] Uh, the great thing about starting this was, you know, the, the power of network and the, the outreach that's come of people looking for the services. Uh, so I, I would love to say that everything's cookie cutter, but I mean, that's a great thing about consulting and advisory is again, every company's at a slightly different stage, but a lot of them start with the, with the same approach.
We, we gotta come in and we gotta do an initial assessment. We gotta figure out where you're currently at, where the friction exists, what the tech stack is that you're running before we can, you know, go and, and say, "This is what's going to solve, uh, some of the challenges and issues," or, "This is what's gonna drive you to that next level." Uh, so a lot of them start in that same phase. There's others that, um, you know, fractional CIO type role. They are a mature company. Uh, they've got a, an IT team. They just don't have somebody who has kind of been at that executive level that understands strategy and vision and alignment, and they, they need to kind of help, uh, not just, uh, focus the [01:07:00] organization, but help to coach and build the team up internally so that they can, they can be that next tier company. Uh, and then there's some-- there's always the one-off engagements around, you know, ERP replacements or evaluations and major systems. Uh, the other great thing that is kind of going on that is interesting is I have had a few startups that I'm, I'm on the advisory board for and, and that's cool because, you know, I get to see these, these companies from a different perspective.
I'm not a buyer, but just seeing where they're at in their life cycle and then helping to, you know, not just create network connections for them, uh, but then advise them on, you know, the approach and the strategy and, and everything else related to the products that they're delivering.
Jason Jacobs: So when you come into a customer, um, for that initial assessment, is it you? Do you have a team of people? If it's a team, what are their skill sets? And, and then how much of it is shadowing out in the field versus, um, uh, you know, ver- versus [01:08:00] talking to the owner and leadership?
Rob Spencer: Right now it, it's me. Uh, you're-- when, when you come to Bilt Tech, you're getting 100% me, 27 years of experience. Uh, it'll be great, you know, I, I keep telling, you know, people that, that I have conversations with, and even my wife, uh, I'm probably a year out before I'm ready to bring on an-another consultant.
I'd, I'd love for it to happen sooner, but I wa- I wanna get things established. I want, you know... I'm, I'm a process and a, and a controls, and, like, everything's gotta be documented so that when somebody else comes in, you know, the same experience that you would get with me, you're gonna get with the next person I bring in, and the next person I bring in. Um, as far as it relates to the interaction, I mean, these assessments, they, they're interviewing, like we talked about. I'm gonna go talk to your frontline people. I'm gonna talk to your accounting team. I'm gonna talk to your IT team. I'm gonna sit down with the executives. Because everybody's got a different perspective on where we're at with the maturity cycles, what's working, what's not working, where the friction is. [01:09:00] Uh, I used to kinda joke in, in the corporate world, I was like, "Adam, I feel like I'm the corporate bartender sometimes" in, in the fact that, you know, I'm not going around and handing out drinks, but, you know, everybody likes to talk to me about their problems and what's going on, and it's, it's not the personal side.
Like, I, I get the brain dumps of everybody and, and the issues that they're facing. And the great thing about that is, you know, being the technologist and being a natural problem solver, then my-- I start trying to figure out how do we, how do we eliminate that, you know? If they're telling me there's a problem here, then there's gotta be a solution to fix it.
And when you start to see those consistent, uh, themes start coming up across multiple conversations, it helps to elevate those conversations to the executive team, so they can recognize that, you know, we, we gotta make a change here
Jason Jacobs: And, um, when you come in, has it tended to be so far helping them make sense of the vendors they have, figuring out where they can consolidate, maybe switch to a different one that's gonna play nicer with the, you know, with the rest of the tooling that they [01:10:00] have or, or are you building, um, uh, or, um, or is it some mix of those things?
Rob Spencer: Uh, it, it's all of that really. I mean, I'm, I'm vendor agnostic. I'm, I'm not here to promote any particular company. Uh, I've got no stake in any company. I'm not taking incentives. I'm not getting kickbacks. I get no referral fees. I've had vendors try to offer me that. I'm absolutely not gonna do that. It's just not how I operate. And so there's some, you know, there's one engagement where we're having a conversation with a client and they're not happy with the solution that they're on today, and it doesn't mean that we're gonna come in and replace it. The-- You gotta start asking the question, "Well, why? And where does the friction, you know, where does the friction exist?
And is it the solution? Is it the processes? Or is it you've got this great solution, but you're just not integrating with other applications, so when, if we get those integrations done, does that solve the issues that you're looking for? Or is it that your, your friction isn't with the platform, it's with the company that supports the platform? And, you know, maybe there's a third party that we can work through instead of working directly with [01:11:00] that vendor to help give you better support." So there's a little bit of different discussions around, uh, the different technology, but there's definitely not a... U-unless they are 100% dead set from the very beginning, we want this ripped out, it's usually an evaluation of all those different factors that make a determination of is it time to repair, is it time to replace
Jason Jacobs: Uh, and and when you do build, uh, is it you building? Do you bring in a firm to build? Um, and then once it's built, how does it get maintained once you leave?
Rob Spencer: So I, I, I have partners. Uh, again, nobody specifically. I'm, I'm not promoting any particular company, but I've got multiple AI vendors lined up, uh, depending on, you know, w- their availability and the type of projects that we're working on, 'cause some of them are really good at, you know, artificial intelligence related to, you know, estimating and pre-construction.
Other ones are really great at the operations side. Some of them are better at the, the back office. So having those [01:12:00] partnerships, uh, is gonna help out from the build perspective. And again, the client ultimately decides who they wanna go with. I'm just there to say, you know, "Here's a good, better, best scenario.
Here's where they excel." You know, they go through the interview process and, and they select. Um- Sorry, I think I lost the back half of that question
Jason Jacobs: Oh, yeah. So if they build it, do they then maintain it over time?
Rob Spencer: Uh, again, if, if, if we're using a third party, it, it depends on the complexity of the build as to whether or not, you know, anything's gonna be maintained in-house versus externally. I would say 90% of the time the third party is gonna continue with the maintenance and upkeep and the changes as needed. Uh, again, just because the internal team either doesn't have the bandwidth and capacity or they just don't have the skill set to keep up with it
Jason Jacobs: Uh, do, do you find that most of the vendors that you are purchasing from are still with a traditional SaaS model, or are you seeing that the business model has been evolving under your feet?
Rob Spencer: I, I think [01:13:00] it's what kind of model?
Jason Jacobs: Oh, the business model. Yeah. Is it still SaaS or, or are you seeing other models emerging?
Rob Spencer: Yeah, I mean, most of it's still in the SaaS model. Uh, you know, most companies are, are, are just not bringing things in-house. Um, I'm, I'm a cloud-first guy. I, I don't like, uh, infrastructure on-prem. I think there's still a purpose for it in some very extreme use cases, but the SaaS model's become phenomenal because it's allowed us to shift our resources from an IT perspective from having to have, uh, a multi-person infrastructure team just to keep hardware running and just to make sure that we got failover sites and we're testing those failover sites and everything else. With the SaaS model, I mean, they're keeping up with all the infrastructure. We just gotta make sure that we got the r- right rules and controls in place as it relates to data migrations and APIs and data exfiltration. So we've been able to focus more of our in- IT resources into data and applications and less from, from infrastructure
Jason Jacobs: [01:14:00] Uh, w- um, last question. As you look forwards, do you have a wish list for the tools landscape of, you know, are there, are there particular things you wish existed that don't? Uh, you know, do, you know, do you want the landscape tools to look, look different in the future because it'll make your life easier and your clients' lives easier?
Um, just any thoughts around, um, uh, you know, 'cause we talked about what you think it will be, but, but now I'm asking what do you want?
Rob Spencer: That's, that's a very good question You know, I don't think that there's a-- There's not a silver bullet. There's not this one product that I think is just going to be, um, the game changer. What I want the construction industry as it relates to technology to get a grasp on is their data. I go back to the data foundation more than any t- more than anything, especially now. With everything that's happening, with the rapid pace of technology and artificial intelligence, you gotta get your data clean. You [01:15:00] gotta get the data foundation set because everything launches off of that. And without it, you're-- y-you got a house of cards, and I hate to see it come crumbling down. Uh, uh, my other wish list I'd probably say is let's get more companies above that one point five percent of revenue. I know I'm probably gonna take some heat for that because, you know, we're already working on tight margins and-- But, I mean, like any time, there, there's ways to job cost it. There's ways to roll it into the project. It doesn't have to be strictly an overhead expense. And then when you think about the enhancements to revenue and protecting the bottom line, the, the profitability and how that contributes back, I mean, there, there's ways to model it out that those gains is what wins you that additional money. That additional money is going to bring the advancements in technology in your organization. It's going to make you a competitive advantage. And so I think those are the two big wish lists I could have out there for the construction industry right now
Jason Jacobs: Uh, so on the d- uh, I just wanna, I guess, [01:16:00] I, okay, I hate to use the word double-click 'cause it's so cheesy, but, uh, just wanna double-click on each of those, and then, and then we can wrap. On the data side, so in order to get the data how it needs to be, how? Um, you know, are there firms that do it? Uh, you know, do, do, do, do you need to bring on new skill sets in-house that didn't previously exist?
Like, how do you take what's trapped in so many people's heads or on pen and paper or, uh, on da- on data silos with different vendors that don't talk to each other and, and get it in a way that it's usable and organized?
Rob Spencer: Yeah, th-there's definitely firms that do it. Uh, you can bring the in-house expertise in there. There, there's some great data architects, data engineers out there that exist that'll help you, you know, if you bring 'em in-house, they'll help you get things back on track. Um, but just like any- anything, you know, the best time to do it is now.
Don't wait another six months. Don't wait another two years. You, you gotta start right away. And, you know, part of it starts with just a mapping exercise of where does the data exist today and and where are the manual touchpoints, where are the integrations, where do we have the [01:17:00] failures? And then the second phase is, okay, now that we know where all the data is, how clean is the data? You know, there, there's still systems that exist out there that lets you put a date into a dollar value field because they don't have the backend controls that stop it, which is ridiculous on some of these large ERPs that exist out there that do that. I'm not gonna drop names, but you know who you are. Uh, but you gotta cleanse the data because just because you've got good, you know, mapping of where everything is, it doesn't mean that you've got good data. And then the last piece that, that helps with that data foundation has gotta be the adoption. You may have clean data, but when you've got gaps in your data because the entries just aren't there, or things just aren't being processed because people are resistant to the change or they're just, you know, working around it, that's the last piece.
If you can drive the adoption on top of, uh, a good clean data set with a great foundation and everything's mapped and tied together, that's when you can start to leverage the artificial intelligence to do great things
Jason Jacobs: Uh, and then [01:18:00] on the, um, on the driving that one and a half percent number higher in terms of the technology budget, just speak to owners for a second. What's the pitch?
Rob Spencer: I mean, it, it-- everything that it-- let's-- I think I said this before. Everything in construction is being done through technology today. there's still some laggards who, you know, are, are still doing some paper-based, but very few companies are still touching paper. Um, where I would say the next challenge is we gotta get people off of Excel, you know, for running critical operations of the business.
But when you're running through technology, when you allow them, you know, allow the GCs to charge the costs of, of Procore and Autodesk and Trimble and CMIC and all these systems that they're using to run your projects, there's a reason. They're doing it not just because it's a shiny object. They're doing it because they want it to come in on time.
They want it to come in with low quality and issues. They want it [01:19:00] to be done safely. So you've got to embrace the technology as a part of what it takes to build your project and allow the GCs to charge it to the project as part of it. That's going to help to get to the one, 1.5%, because then it's a job cost, it's not an overhead expense. And it's just gonna make things better for your project too, not just throughout the life cycle, but now you get to see the... You know, you think about all the imagery and the job progress, and then the handoff to the, to the operations team. You know, when you have all that detail and that data, you know, in today's world, if you don't have all the technology, if you just got a set of as-built PDF plans, you know, that, that's great. But the reality is, if with all the reality capture and everything that's happening out there, the plans say that there's a pipe here, but the reality is, it's six inches to the right. And so operate-- you know, your, your field and operations team, uh, or facilities and operations team coming in to do some maintenance is gonna, you know, punch holes in the wall over here when really it's over here. So anyways, all that stuff starts to play into not just the constructability [01:20:00] and the quality and the, the scheduling and, you know, everything else that comes into play, but it's gonna help with the, the maintenance side of things too. So be a little more generous to your GCs and let them charge some of this to you. It's, it's gonna benefit you. It's gonna benefit them
Jason Jacobs: Well, this was such a wide-ranging discussion. Is there anything I didn't ask, Rob, that you wish I did, or any parting words for listeners?
Rob Spencer: No, this has been another great conversation. Um, I'm sure we're gonna talk again. Um, we- we've had some, some good dialogue up to this point, and I know that you're still learning, and I, I'm eager to keep listening to the podcasts that you've got because you've had some great speakers on, and everybody's got these different perspectives. So there's little nuggets of wisdom that come out of every one of them. I appreciate you putting this on and doing it
Jason Jacobs: Well, uh, yeah, it's all which direction you look. I feel like in this, in the, um, several months that I've been doing a deep dive here, I've learned an incredible amount, and every week I can see myself just climbing the learning curve. So [01:21:00] on, on that, uh, on that hand, I've made a lot of progress, but when I look relative to what I need to know to actually be productive, right?
I think talking to people like you that have been decades in the industry make me realize just how much I still don't know. So, um, it's gonna take a while, but I-- that's why I'm not jumping to build and it's like dues paying and, uh, and hopefully by, by not just doing the work, but showing my work, it's, you know, it's maybe a way to, uh, you know, to make up for the domain-specific resume that I will never have.
So, yeah.
Rob Spencer: Going back to that intellectual curiosity, I think that's why we connect is that curiosity and wanting to ask questions and just learn and learn and learn as much as we can. It, it just makes us better as people
Jason Jacobs: Well, thanks Rob. I really appreciate you coming on the show, uh, and sharing your wisdom. Uh, best of luck with the new practice and, uh, and like you said, uh, eager to keep in touch and, uh, and continue to compare notes as we both get further along
Rob Spencer: Thanks, Jason. I appreciate it
Jason Jacobs: That's it for this episode of BuiltForward. I hope you enjoyed it. If you found it useful, [01:22:00] share it with someone building in the industry and follow the show so you don't miss what's next. Thanks for listening and see you next week.