The SiteVisit

How AI is Reshaping the Construction Industry for Good with Shawn Gray, CEO at ConstructIQ Advisory

• James Faulkner

Share episode

Send us a text

The construction industry is at a crossroads as it embraces the transformative power of artificial intelligence (AI). In this insightful episode, Shawn Gray delves into the pressing need for the construction sector, especially mid-sized firms, to engage with AI solutions that can substantially enhance productivity and operational efficiency. As labor shortages and demands for faster project delivery become critical issues, AI might just hold the key to unlocking greater profits and smoother project execution.

Shawn elaborates on how many firms are exploring AI but less than 4% are meaningfully adopting it. He shares real-world examples of how mid-sized companies are already using AI to automate time-consuming tasks, thereby redefining workflows and allowing teams to focus on higher-value activities. We discuss the unique challenges faced by field teams who are often reluctant to embrace new technology due to past failures, and how AI can bridge the gap by offering intuitive solutions that speak directly to their needs and pain points.

Listeners gain a deeper understanding of the significance of not just adopting AI tools, but integrating them in ways that significantly impact on-site productivity. The episode underscores that the risks of ignoring AI or not supporting its adoption can jeopardize a company’s competitive standing in an evolving market. This enlightening conversation is a must-listen for construction professionals eager to understand how innovative technologies can yield tangible benefits. 

Join us as we explore the best practices for AI adoption in construction and the profound implications it holds for the industry's future. Don't miss out—subscribe, share your thoughts, and contribute to the conversation!

PODCAST INFO:
the Site Visit Website: https://www.sitemaxsystems.com/podcast
the Site Visit on Buzzsprout: https://thesitevisit.buzzsprout.com/269424
the Site Visit on Apple Podcasts: https://podcasts.apple.com/ca/podcast/the-site-visit/id1456494446
the Site Visit on Spotify: https://open.spotify.com/show/5cp4qJE5ExZmO3EwldN1HH

FOLLOW ALONG:
LinkedIn: https://www.linkedin.com/company/thesitevisit
Instagram: https://www.instagram.com/thesitevisit

Speaker 1:

Welcome to the Site Visit Podcast leadership and perspective from construction with your host, James Faulkner, Recorded live from the show floor at BuildX Vancouver 2025. All right, Mr Sean Gray, a veteran of the Site Visit podcast.

Speaker 2:

Good to be back and my butt groove's still in the chair here. Great it is?

Speaker 1:

It is. I think we met here for the first time, correct? We did yeah, yeah, yeah, A couple years ago. I remember you were. I wouldn't say that you were. There's a certain type of intimidation that you had that I felt a little bit intimidated by you in the beginning.

Speaker 1:

And I'll tell you why it was. You were very serious. I'm like, oh, but this guy is like he's out to get it, and that's cool. Though, and we've gotten to know each other over time and I've got a lot of respect for your work ethic and you're very focused on things. So, very cool, it's good to know you.

Speaker 2:

Well, we're dealing with serious things here. We are. People need houses built, we need critical infrastructure built, and we don't have enough people to do it. It's pretty serious, yeah, so we can have a laugh.

Speaker 1:

Well, we can have a laugh, but you're right, it is serious. There's a lot of things going on. So let's just. You are the founder and CEO of Construct IQ. I would call you the metrics guy. You used to have a lot of statistics over time If anybody's listened to any of the old podcasts you had some awesome things that we covered last time, but today we're going to cover what your latest interest in is AI and AI development and construction, where you see things going, opportunities and how you see this playing out.

Speaker 1:

So first of all so I think you and I are probably aligned on this you see a lot of adoption of in a lot of the larger firms. They've got budgets, they've got lots of bandwidth to hire people to be able to just focus on something, make good revenue. They're just balls to the wall, busy all the time, not necessarily paying attention to what opportunities they have in front of them and are probably not utilizing to its full extent exactly how much more productive they could be. So maybe just take us through what your findings are in terms of you call it going from zero to one in terms of this stuff.

Speaker 2:

Sure, yeah, and thanks. And just for context of the audience here because you're talking about AI and that is the most salesable topic right now just to the audience to know that you're talking with somebody who's walked the walk on meaningful implementations of AI across $25 billion of construction over the last 10 years. So I know this is someone who's walked the walk on that, so I just want to make sure we set the stage on that for this conversation. When we say going from zero to one, I'm talking one being an instance of significant value discovered for the business in a state where let's throw some stats at you now in a state where over 40% of industry is exploring AI and what that means behind that, almost 85% of human individual beings are using some shape or form AI solutions in their personal lives or even in work. Most of them are using those Thanks to X.

Speaker 2:

Yeah, yeah and other things, but there is when we say going from zero to one. It's how we're using these high-powered tools In most use cases, barely scratching the surface of the value that they can actually bring to a business in terms of addressing the fundamental constraints preventing them from being productive and profitable. So when we say going zero to one, 40% of industry is actively looking at these things. Less than 4% of them are using these AI type of tools in a meaningful way.

Speaker 2:

I'm talking about. Hey, if you're using these things for reading and writing emails more efficiently, do that? Amazing. That's an incremental nugget of an improvement. It's not major needle moving value. So when we say zero to one, it's going.

Speaker 2:

Where can you be deploying these things to truly address the bottlenecks preventing productivity and profitability, especially at the job site level? And then, when we wrap this around those mid-sized firms, over the last year I just started this real intense journey at the time we chatted last, in the last year that was really focused around those small, mid-sized firms because, as you said, the big groups. They have data teams and digital teams that eclipse 10 times the size of employees these guys have. So bringing these things in and helping them understand that they can be as you said, you said it. These small firms are, you know, the scrappy groups. A lot of them are established on their 500 million dollar companies. Yeah, some of them are 10 or 1 million. More often than not, everyone's wearing multiple, multiple hats and their ability to do good work or take on more work is just capped by their human ability to produce. So in this last year, 40%. So I engaged over 250 construction professionals.

Speaker 2:

It was a busy, busy year, not much sleep, but we were on a mission. 40% were actively looking at things. Less than 4% were really deploying it in a meaningful way. Almost all of them, whether they said they were looking at it or not. Almost all of them were hesitant to actually go forward with anything just because of how they've been burned before on previous initiatives.

Speaker 2:

We've entered an age of kind of a double-edged sword. Here. We've got amazing ecosystem of technology, especially in Western Canada here over 150 technology providers, most of them now providing some type of AI enabled solution. We're number three in the world for construction R&D, but it's an emerging technology. There's no real established use cases or library of who's used these and where and when. So you've got folks that want to use them, but there's no evidence that these things are working well in certain areas. So it's a weird chicken or the egg situation that most industry is in and what I've looked at with these midsize firms their desire to do more work with less is tipping them over that point. They're embracing these new AI-enabled opportunities even before they have well-established centralized platforms and things like that, because you can provide immediate value to a job site stakeholder.

Speaker 1:

But what would that be? Give me something specific.

Speaker 2:

There's been some interesting, lots of different use cases, but what's kind of interesting about these things? A recent example from a highly innovative mid-sized group in your neck of the woods here they started by introducing AI agents just for the purpose of automating meeting minutes on meetings and on projects. Very trivial, actually didn't add too much value, but what that did was it fundamentally shifted the job function of the project coordinator, who typically would have been recording notes. They've now shifted to pulling up their BIM models, pulling up RFIs, pulling up relative project context in that meeting to make those meetings more effective. I'm going to tie all this together because these are all AI agents, so they're all integrated.

Speaker 2:

Yeah, so automating meeting minutes, the job site superintendents, safety officers, quality managers being prompted throughout the day on key areas of activity or risk so that those field folks can be putting in via text or voice what is actually happening on the project during the day. Being prompted to go pay attention to certain things on any given day that they may or might not know about is critical to the project. We know there's so many fires to fight, so helping them kind of understand where to be deploying their time and then when these agents, on this specific example when we had these agents deployed in meetings, deployed at the field, getting operator field level information because they're integrated, they started to understand a lot more project context between the different clay layers and they started you're able to build a lot more risk-based or contextualized tasking of activities during the day, a lot more risk-based or contextualized tasking of activities during the day.

Speaker 2:

That has been a fundamental when we look at the headaches that our field leadership have. A lot of the time they're just responding to people's problems. What's happened really, really quickly in a matter of weeks. Having these agents deployed, it's really helping them focus their time on really what matters the most and it's impossible for these small, lean teams to be cascading communicating the right information around and almost overnight these agents are taking on that task, so I think that where the confusion sets in with a lot of companies is the confusion sets in with a lot of companies is like, even when, when we start working with a company, they say where's my data?

Speaker 1:

Well, what? What service is it? Is it in the U S? Is it here? Where is it? So you know, we give them information on where their data is. How do I get that data? Make sure you're not sharing that with anyone else. So the notion that their information is going into an agent model, somewhere that is not private, how does that work?

Speaker 2:

So that is usually the first point of learning with these groups, because that's usually what they're concerned about the most, a lot of it driven because of media-based dialogue, but quickly overcome when they understand that I mean there's a difference between those open AI, chat, gpt type of systems and systems that are project or business specific instances. So that is usually the number one learning. It's closed loop, it's built for you. There are backend algorithms that do things, but it's just functional.

Speaker 1:

Where are those backend algorithms Building the AI comes from? The models have to come from something macroly, because you're not building those yourselves.

Speaker 2:

Some are building those themselves.

Speaker 1:

I mean, I've gone through this with makecom, for instance.

Speaker 1:

You go there and but that's going into a big database of stuff. So what we've been looking at is that we would have two specific databases. We basically would have one for capturing the data and then you have that information, depending on what the customer wants, then ports itself into another database for analysis, because trying to crunch it from this with all these different schemas is a gong show. So you basically have to, and you can have agents in between that are going to be doing some things. Yeah, so I think that I was looking at the idea of agents in a construction company, an AI agent, for instance, no-transcript, where does that change?

Speaker 2:

So that's one of the initial use cases for AI. When we look at, I would say what does AI even mean? I would say automation and integration. That's what it is right now for most groups. So that part we're going. Hey, you've got these closed-loop systems. That's what it means to them.

Speaker 1:

You mean Rather than artificial intelligence. No, I tell them that.

Speaker 2:

Oh, I get it. I said don't think about artificial intelligence, because these things aren't actually thinking. They're based on prompting and number ones and zeros. They're not thinking. But I would say, for the highest-value use cases for you right now, think of it as automation and integration. I see okay, and keyword is and Do them both. So those closed loop systems and platforms and some of them promise and sell that they are amazing integrators with all the other ecosystem, when we know that that's been validated as maybe not true or that's a big pain point of the community these days. So that is a function that these agents are playing. It's going. Hey, you can now bring in job function specific what we used to call point solutions, job task specific solutions and have agents running all of the different integration pathways between them.

Speaker 1:

Do you know what the big gong show is? Here Is the Google Play and Apple App Store.

Speaker 2:

Elaborate on that.

Speaker 1:

Because they're the bottleneck. You can't just have some.

Speaker 1:

They have to test your API in order to pass your app Interesting. So it's not like you can just put something out there with some totally open-ended blank slate application. It just won't get passed Right. So that's what we've been thinking about, these things as well. So where integration gets involved is where, if you're talking about multiple APIs through the App Store, for instance, apple's the worst. I mean, they're good in some ways because they're very controlled, but they're the most difficult to do. So if you were to have, for instance, if you had one app that was going to integrate on the fly to make that really work there on the mobile side, to get fast time, because you're talking about offline, you're talking all this. It's a gong show to actually do in practice. What's easy to do is if you are going on your own on the web, because I have a feeling that web applications are going to see a huge resurgence here mobile, responsive web applications because of this App Store restriction, because you can go real-time right away, start interacting with models.

Speaker 2:

Absolutely. The web apps are where you're getting the most horsepower right now, because of that exact statement you made. These are internet based. Well, let's call those more the open models. It's internet based information gathering and moving, so it has the best ones. The highest value operating applications are web-based right now.

Speaker 1:

Yeah, they would have to be, but that means no offline. Well, there is no offline.

Speaker 2:

Let's say you can maybe paint an artificial viewpoint of what that means to be offline, because some of these agents do things, let's say, in that offline state where I know what you mean.

Speaker 1:

But offline on an app, at least you can get the UI. You can't even get the UI on a web app without connection.

Speaker 1:

But no, I think this is very interesting having this conversation with you because it is. I look around and I see Sage here and they're not really talking about much stuff. They are here kind of maybe still talking about the same things they were 10 years ago. Good company, we're a partner with them. But yeah, it seems like there are solutions that are people with no customers who are trying to do things. That it's too. It's very risky. It's very risky to start getting into doing some of these things and the the processes in construction is things are going already Like even we saw this when we saw the transformation from paper to digital For a while.

Speaker 1:

They had to do both because they weren't sure the digital was going to work out, so they still had to fill the piece of paper. Now, let's say they go from digital to automation or AI. They now have to make sure that A it's not making any mistakes, it's an A, and if it is making suggestions, that it's not putting people in a wild goose chase either. But shit, that doesn't matter. So there's an interesting place in between. I think that is where the rainbow to the pot of gold is.

Speaker 2:

It's right in the middle, and I agree with you. And even when you said these different companies and platforms are or aren't talking about AI. Not everything needs to be about AI, and that's kind of what industry is trying to figure out right now is where are the right applications for this. But then when we're talking about what is that magical area? It is around more like those. I always help the groups understand. Where are these areas that are lower risk?

Speaker 2:

You're not making a major critical decision out of this, but where is something that, especially around let's call it the value of this? But where can you? Where is something that, especially around what's called the value of time? Where is something that can be take a significant amount of time away from your task, that can impact and have butterfly effects to many, many stakeholders in the easiest way possible, that doesn't diminish the confidence or require somebody that has 50, 40 years of experience to know is this the right call or wrong call? So that is where that, exactly as you said, that is that pot of gold at the end of the rainbow. I think some of us have been doing this for a bit, but even just last year, these groups that have kind of embarked on that avenue. They're finding those immediate pots of gold of value, and it's not from them coming up with ideas. It's these things. You have to use them, you have to have those. Especially the field stakeholders discover those pots of gold.

Speaker 1:

Do you find that the field is probably the most difficult? It's a weird environment Because the office seems like it would be simple.

Speaker 2:

It's a weird environment. The office folks are actually some of the most jaded right now because they are almost the actual most tech fatigued because they've been dealing with office-based products for almost 20 years in the industry. The field groups I think we talked about this maybe last time. The field groups they're sick of being forced technology that is really meant for data capture to help somebody else out in the office.

Speaker 2:

So when you bring in, they're actually a weird stakeholder to work with, where they've been jaded before, they've been hurt before by these initiatives. But coming in an approach going no, this is now about you, everyone else has the tools or should have the tools that they need. This is about you. How can these add value for your job? And putting a lot of equilibrium back to that ecosystem where you've brought in all that technology to help somebody else out. The ecosystem's unbalanced. So it's interesting how you bring things in that were directly resonating with a field stakeholder's job function. It balances the equilibrium out immediately.

Speaker 2:

We did an interesting project. This early stage startup group was doing some interesting report automation things, text to voice type of things. The chief safety officer think about the most stereotypical cliched construction person kicked the startup off the job the first day Within four weeks. This CSO was the champion because he understood. He saw right away that we just automated LEM, labor counts, capture from the trades, those two hours of this guy's day like walking around hunting this stuff down, then transferring it immediately. This person was like well, this took me no time, I didn't have to do anything.

Speaker 1:

So where did the inputs come from?

Speaker 2:

Right from the trades. So trades via text or voice, using AI prompts to their mobile platforms Right, but they have to download an app to do that. No, just well, I think. On this instance I think they use WhatsApp just because that was a common job communication tool on that project, but it can you know via SMS as well.

Speaker 1:

I see okay.

Speaker 2:

Yeah, so, but this would have been typical high resistor was the champion at the end because it was like this is for you, this is to help you bring to save time for you. So it's an interesting, weird thing where the field groups who had the bad rap of being the worst to deal with they're almost the most hungry for meaningful applications that can help them perform better.

Speaker 1:

Yeah, I mean I have. You know, obviously, the position that I'm in. I've been thinking about this deeply and I definitely have some. We're seeing history repeat itself, just in a different paradigm, but it's the same transformation when we first saw the iPhone, mm-hmm. To me it's. It's the same transformation, going from a baked keyboard to a not-baked keyboard. So I mean, I see it pretty clearly and I think that there is and this is mostly to fix the field, because the field is where this, this is where it's there's crazy inefficiencies. You also have to have uh, you got a revolving door of staff too, and you have all the way down to the lowest common denominator of the bring your own device situation. Because, as you said I think I think it was on the one, the podcasts you said that a high percentage of the revenue of construction is in projects that are small, right All over North America.

Speaker 2:

Yeah, Nine percent of well, let's call it over 90 percent are smaller mid-sized businesses, Exactly.

Speaker 1:

Right. So this isn't the building, the hospitals and airports. I know I say that a billion times but those big projects, institutional projects, that's not where all the money is. A lot of the revenue is being made. It's a huge amount of revenue, but the projects are very few and very big. But the smaller the company gets, the less organized it is, unfortunately, the diverse caliber of worker you're going to get.

Speaker 1:

Of course there's edge cases there, but on balance it's going to be a different kind of deal, right, and having software that is going to utilize these AI tools or automation tools that field part it's not going to really the office stuff is going to be ubiquitous. I mean, from doing a schedule you're going to have automation prompts and commands within software. Take the schedule I had last time, attach this to read the blueprints and give me a different schedule based on that. That intelligence is going to have to come from somewhere. I don't know where those models like you know they say, well, the AI, what AI? Which one Like? Where is this? You know who, what platform is reading this, what has enough servers and enough bandwidth to be able to crunch this? And it has to learn somehow right, and it can't learn the wrong thing. So you have to train models. You can't just like they don't just like suddenly start working.

Speaker 2:

It's interesting when you talk about they have to learn from somewhere In the early days of this journey last year. Any seminar I've done usually is about field productivity. Most of the audience is from pre-construction and wanting to know how can they improve estimates and things like that. My first answer to them is did you not listen to the entire seminar about the fact that what is happening during project execution is what is driving if you're going to make or break your estimate of profit? That is where you need to fix. We're not going to go talk about automating estimates other than maybe takeoffs and things, but we're not going to talk about predicting a project outcome because we already know what it's going to be.

Speaker 2:

It's about, as the theme of this conversation, it's fixing what happens during execution and with the hypothesis of if the people that have the experience to know how to proactively prevent challenges are busy doing something else, whatever that something else is free, that so they can go improve the project. Only, until you get to that state, we're talking even a 1% improvement. I think at the highest level, $8 billion GDP. Industry 1% improvement in productivity is $1 billion. Yeah, so we go just give it a nudge, just a little nudge that will translate into major project and profitability improvement. Then start looking at some predictive performance on the front end. But you can't fix something that's that's broken until you fix the right spot.

Speaker 1:

So we're in the sticky middle right now yeah yeah, and I think, uh, I think what you're doing it makes a lot of sense. And I think, uh, yeah, there there's lots to be done. I think, well, we can talk about this offline. I hate to say this to the listener, but I'm not going to give away everything I've been thinking over in the past number of months, but definitely something we should talk about more.

Speaker 2:

What do you?

Speaker 1:

see as some of the negative aspects to this. Negative in what context? What risks are there?

Speaker 2:

There's obviously opportunities, and then there's risks too. Well, I'll throw a weird risk at you. I feel the biggest risk is around groups that aren't actually actively engaging in learning about these and exploring, because this is unlike traditional platform era where you press a button and it goes the next step and does something the next generation of tools. You have to engage with them and you have to engage with them for them to build and learn and get better. So maybe this is a different area of risk that you were trying to flesh out, but I see these tools are things you have to use for them to get better.

Speaker 2:

The biggest risk that I'm seeing from a business is not just where is the data? What's the safety around that it's? Are you looking at these things? Are you people using them in a manner that adds value? Because if you're not, the risk is that these groups just as we saw with you know, wearable augmented reality devices, things that probably could have changed the game people didn't pick them up and use them. Those companies folded, collapsed and went somewhere else. So I find actually the biggest risk isn't what the AI can or can't do things around that nature. It's the fact that you have to be using them. You have to be supporting the companies that are building these things or they will go away. And what's interesting, and the reason why we developed the mission statement around the Prairie PropTech Association was we have a number of amazing technology providers in this Western Canadian ecosystem.

Speaker 2:

A number of them have left, either folded completely, or have left to go down to Texas or overseas. Because of that, what we're seeing is just that adoption gap and I go. That is the biggest risk. We've got these things that other companies have seen the significant value where it's freeing up time and directly translating it into productivity, which we need so much right now. But if they're not engaging and supporting these groups, they will go elsewhere and then you've lost that opportunity. And that's just my take on. Where is a big risk?

Speaker 1:

If the risk is you don't use it, you're going to lose it. Well, that makes sense. Okay, well, that's pretty cool. This has been good, always a good conversation you know how many more We've got to start.

Speaker 2:

We've got a whole series coming up now?

Speaker 1:

Yeah, I know. So how do people get a hold of you?

Speaker 2:

You can definitely check out the website, wwwconstructiqadvisorycom. Yeah, you can follow me on LinkedIn as well. There's a major initiative helping subsidize cost expertise, that walk the walk on these things lower technology costs, you know really helping those mid-sized firms be the first movers and establish value. You can see that on my website. Also check out prairesproptechcom.

Speaker 1:

Cool Right on. Okay, well, Sean, thank you very much.

Speaker 2:

Well, thank you very much. Always a pleasure. Yes, sir, thank you All right.

Speaker 1:

Well, that does it for another episode of the Site Visit. Thank you, contractors in North America and beyond. Sightmax is also the engine that powers this podcast. All right, let's get back to building.