Revolutionizing Construction: How AI is Transforming Job Sites

🔗 LINKS

Join the group coaching program and Master Construction Cash Flow with Profit First

Visit the episode page at for more details and a transcript of the show.

Get Profit First for Commercial Construction--OUT NOW!

ℹ ABOUT THIS EPISODE

In this episode, Wade and Stephen delve into the transformative impact of AI and technology on construction job sites. They discuss AI's role in improving safety, estimating, and project management, while highlighting real-world applications in job safety, video monitoring, predictive maintenance, and smart procurement. 

The episode also explores the integration of AI for legal document analysis, bidding, and forecasting, ultimately showcasing how these advancements can save time, enhance efficiency, and increase profitability in the construction industry.

The book is available now on Amazon!

Subscribe to get notified as soon as new episodes go live.

⌚️ In this episode:

• 01:35 AI for Safety and Monitoring
• 03:37 AI in Insurance and Legal Documentation
• 06:56 AI for Project Management and Estimation
• 14:12 AI in Equipment Maintenance and Theft Prevention
• 24:01 AI for Procurement and Material Management

THURSDAY, JULY 10TH, 7-8PM EST 

Join us for a FREE live breakdown of the new 2025 tax law, which introduces major opportunities—and a few new rules—you can’t afford to miss.

Click here to add it to your calendar.

Join the Profit First for Construction community!

Find all episodes and related links at ContractorSuccessForum.com.

Join the conversation on our LinkedIn page: https://www.linkedin.com/company/CarpenterCPAs

FIND US ONLINE
Wade Carpenter, CPA, CGMA | CarpenterCPAs.com
Stephen Brown, Bonding Expert | SuretyAnswers.com

Transcript

Wade Carpenter: [00:00:00] If you think AI on the job site means a robot swinging a hammer, think again. It's not replacing your crew, at least not yet, but it is showing up as a foreman who never sleeps, the estimator who never forgets, and actually pays attention. We're Today we're digging into how AI and tech are actually changing construction on real job sites, in real time, and the tools that are saving time, money, and sometimes your backside.

This is the Contractor Success Forum. I'm Wade Carpenter with Carpenter & Company CPAs alongside Stephen Brown with McDaniel Whitley Bonding and Insurance. Stephen, let's face it. If AI can help your crew show up on time, fill out a time sheet, and stop leaving power tools on the roof, it it might be the most valuable guy on the payroll.

What do you think?

Stephen Brown: Yeah, AI can totally control your employees and make them behave like you want them to. Absolutely. you said, it's It's not a robot controlling a hammer, Let's L et's get the science fiction part out of it, and let's talk about what supercomputers [00:01:00] do to take massive amounts of data and compress it and be able to answer your questions from a logic standpoint.

We're gonna talk to our listeners about all the different ways that AI can help contractors, more specifically what is it you want to accomplish and what can it do?

Wade Carpenter: Yeah, and like you said, it might seem like science fiction, but this stuff is really coming to light and that's why I wanted to throw out some things that we just kick around today, some real world things that we can point to that actually is happening and making a difference on job sites and the profitability.

So first one I was wanting to throw out, which you had given me a couple examples too, but I had seen some real time video with AI monitors that detect your safety violations, unsafe zones, some near misses on an excavator, stuff like that. There are some things out there that's called, Everguard.ai's Sentri360 platform. It's analyzing video by itself or [00:02:00] some of these wearables to alert supervisors instantly when things are possibly gonna cause a safety, violation or somebody get hurt.

Stephen Brown: More importantly than the violation is somebody getting hurt. It just amazes me how AI can also take images, like you were saying, the visual of a job site, and do some predictive analysis on the safety issues of it. And also for equipment maintenance, which is the smallest thing. What What is the downside? Why wouldn't you do it?

If you're You sent me an outline that kind of showed the pros and cons of implementing this AI in different situations for contractors. But I would say if you're not looking into it and you're doing it, you're really hurting yourself and possibly hurting one of your employees.

Wade Carpenter: Like big brother with a hard hat, basically watching your every move. Sometimes you think about what the cost of an injury is. It's not just the lost time, the productivity on your, your job site, your workers' comp might be going up, those kind of things.

I had a client, [00:03:00] that I worked with for several years and five years ago, they were doing this. They were originally providing security guards on job sites, but they went to where they got cameras and not just, it was like, infrared cameras and stuff like that, so you didn't have to have a security guard on site.

But they got it to where, even back then I. that, AI started looking at some of this stuff and they didn't have to have as many people. So thinking about, hey, we could prevent some theft, we can prevent an accident. These kind of things are actually happening now.

Stephen Brown: Absolutely. It's there, it's happening and it works, it's effective. Wade, what do you think our listeners need to know about AI platforms, different operating platforms?

Wade Carpenter: There's a bunch of different operating platforms and a lot of them may need to be built specifically for this particular purpose. But, this may be a question for you, now this one I was particularly looking at that I just mentioned, they were saying in their materials it says they can cut [00:04:00] insurance costs.

I don't know if anybody's ever approached you about some kind of system like this, just like your telematics in a car, could it possibly help their insurance rates?

Stephen Brown: Well, they're all using AI and their insurance rating now. I think of predictably, what would tell you that a driver is not a good driver? You're MVR, your past driving record tells you a lot. And also the telematics that you put literally on the dashboard of your vehicle that the employee's driving, that's giving real time data about the speed, how they drive, whether they're erratic, their attention span. All these Kind of build into a predictive analysis in real time.

You're Gonna see insurance companies down the road that are gonna be able to take drivers based on their realtime driving habits. Instead of saying, well, they're too young. They've got a couple of speeding tickets on their MVR we're not gonna ride it. They'll predictably be able to analyze that driver, and decide on whether they're gonna take the risk or not.

What concerns me is that the [00:05:00] insurance company starts saying, we're seeing some habits we don't like. So we're not insuring that driver anymore in real time. That's scary to me because you buy a policy for a year and it should cover you and the drivers that the underwriter has agreed to cover for that year. So yeah, that comes up a lot.

No, another thing to me, Wade is so fascinating about AI is as a bond agent, I get so many documents. Instead of just getting a bid bond request and a copy of the bid bond, I just get a link to the specs. So not only do you have to figure out what the liquidated damages are, and you have to let the bond underwriter know about what size price ranges this project. A lot of times that's all you have to go on and we have to take all that data and analyze it.

And using AI now, all the surety companies, I use it all the time to quickly scan legal documents and look for language in there that my customer needs to be aware of before they sign it. But also [00:06:00] taking all that data from the beginning of the job. You're starting the job, you're downloading all that data into an AI system like ChatGPT or something, and then you're able to access information from that data.

Every contract you do, for example, you're out on a project there's a problem occurring, you can literally get on your cell phone and ask AI, who is the subcontractor that did the so and so work at this particular project? And what's their contact information? And literally pop it up and call them on the phone and say, Hey, you know, the roof's leaking here. You need to get out here and fix this.

And then not only you doing that, but it's documenting the conversation you had and, what they said about it. So you got some legal recourse when they don't do what they're supposed to do. All that just blows my mind, Wade.

Wade Carpenter: Yeah, and you just unpacked several things I wanted to talk about today. I guess going back to my point about if we had some kind of [00:07:00] system like that was monitoring for safety. We're gonna talk about some of these other things on this job site too, but just like insuring property, if you had an alarm system on that property, I imagine that would help insurance rates.

Being able to put all this information in there, some of the stuff like forecasting materials needs from the job history or the weather. There's so many things you just unpacked I love to talk about here. But, there are systems like there's something called Atlas Technologies, and then I think there's Togal and there's Toggle AI.

Toggle AI, it's got this backlog driven supply analytics, those kind of things that can predict when you need to get it on the job. So it's like trying to create this crystal ball and maybe not having a dedicated purchasing agent that's basically goes on vacation and things get dropped and those kind of things.

If it can do some of that, it can definitely save you time as well as make you more efficient.

Stephen Brown: Yeah, we talked about data management, [00:08:00] right? But what about bidding, estimating?

Wade Carpenter: There are some that, if you can just throw the data in there, even if you created your own system for this, it take your past bids and put it in there and then if you could collect it and say this is this type of job, it could probably help you predict some things that, maybe have gone wrong on the job or whatever, or maybe this wasn't as profitable as you thought it was gonna be. Some of those insights are well worth the time it takes to do it.

Stephen Brown: Yeah.

Wade Carpenter: And probably save you a ton of time on trying to figure these things out.

Stephen Brown: To me, what's great about it from an insurance standpoint is that AI is gonna help take those underwriters that have one certain thing that happened 20 years ago that they will never insure because that happened, out of the equation because it's gonna give real data information that kind of steers that underwriter from a company standpoint of what kind of risk they're gonna take.

And it's the same with bidding, you historically bid things over and over again. There's all this [00:09:00] information in that bid. There's all this work you did to put the bid together. It costs money to put a bid together, and there's all this data that goes in on unit prices for a job that you're bidding. And if you can take that information and you can store that, and you can use that history of that to help you bid better, why aren't you doing that?

Wade Carpenter: Absolutely. Great idea. I'm thinking about some of the real world things that I found and looking at, there's something called 3D sync we talked about BIM and drones and sensor data, but there's this company called Bentley iTwin and NVidia that makes all the NPU processors for AI.

They created some kind of system like infrastructure builders that can go out there and run through the job and model it for you. It's like dress rehearsal before you even drive that first nail in there.

Stephen Brown: When you talk about these different options that are out there, I'll just say listener, Wade Carpenter's absolutely the best host to have on this because [00:10:00] he gets deep into this stuff, and he loves tech more than anybody you've ever met. And systems. You know about Elon Musk, building this big supercomputer cooling facility here in Memphis. But one interesting thing was while they were reaching out to contractors to do the work, a lot of those contractors I was talking to were just intimidated. We don't have these AI systems that can fight their AI systems to make sure we're complying with everything. And that was their fear.

Wade Carpenter: Well, I could see that, like this one I was just talking about. This has the capability to go out there and simulate changes before they cost you money or forecast where crews are clashing how you know, they, the timing's not lining up. It's amazing what this stuff, these virtual reality walkthroughs it can do, and it is not science fiction anymore.

Stephen Brown: Right before we started this podcast, we were talking about what AI can replace and what it can't replace, and in real time, how fast this is moving.

Wade Carpenter: Yeah, and [00:11:00] going back to what you were just talking about too, I was talking about the Toggle AI, but this is where we were talking about before we even started this morning, that Togal AI. And there's another one called ProEst. It does the same thing. It puts all this stuff into one system, and it is not science fiction. You don't have to create it yourself. There are people out there building this stuff for you.

Stephen Brown: I would think the most difficult part was just being able to get all your data into that operating system, and then you would say, is that data secure? This is some vital data for our company that we don't want a general open platform to share with everybody.

Wade Carpenter: One other thought that you already sort of mentioned too, I I was actually on AI panel in Baltimore last month and we did also go around the room and this is one of the things they were using, a lot of contractors may not have thought about it. You can dictate to your ChatGPT and voice to a report, like field [00:12:00] logging, what happened today?

AI generated field reports. Instead of typing a whole bunch of stuff in, they hate doing it. On the truck home, get on your phone and we could have a chat bot call you and ask you the question like, what did you get done today? And this stuff, there's, the Open AI whisper, there's something called Tactiq , T-A-C-T-I-Q that goes along with Procore like apps, stuff like that, that can actually generate these job site reports. And like the podcast we recently did with Ariela Wagner, she was talking about documenting all this stuff. This is a great way just talking and write me a report. This is the things I want you to do every day. And whether you even pay for one of these things or just use ChatGPT, it's not hard to spin up a custom GPT that says this is the format of the report, these are the things I wanna tell you every day.

Stephen Brown: When I think about the most important tool of a project manager and estimator, you're gonna be able to [00:13:00] access all of this from your cell phone on the job.

Wade Carpenter: Yeah, that's something we learned a long time ago is trying to get cost data out of our clients and that they're on a job, project manager, we get it on their cell phone. We have a lot better results than trying to wait until they go home and they're tired and they don't wanna mess with stuff like that.

If they could document this stuff in the field, it saves a lot of this administrative time. Maybe you got somebody in your office that's trying to chase somebody down and has to do that every week, it reduces a lot of errors and, eliminates bad handwriting from other things.

If you could put all this in there and has it all written up, typed up, and so maybe the owner has all these project managers trying to keep up, they can actually be getting some kind of updates in real time. What's going on with the job?

Stephen Brown: Absolutely. So you got communication about the flow of the work. You got the ability to ask important questions that you need to answer to quickly, and you have the ability to dictate, in the phone, real questions or problems that came up while you're thinking [00:14:00] about them instead of when you get home at night and you're exhausted.

Wade Carpenter: Same thing. That's why trying to get receipts and stuff. I could tell on and on about how to make the phone a a lot more friendly to getting the information you need.

But, one of these things that you had brought up already in here, and as well as I know we were talking before, is the predictive maintenance on machinery or some of these sensors, whether it's just with a predictive maintenance or you see some of the whole cage is starting to-- and maybe that operator didn't care or didn't point it out, you don't wanna ruin an excavator because somebody didn't pay attention. These sensors can actually be installed. There's a one called StruxHub, which I think is Caterpillar's system, it's some kind of predictive system that they can actually install and as much as it can cost to fix a lot of those things, if you can fix it before it's broken, it's just a whole lot cheaper that way.

Stephen Brown: Isn't that truth?

Wade Carpenter: Leave you stranded.

Stephen Brown: It's amazing, . It gonna predict something that nobody knows? Believe me, it's based [00:15:00] on an immense amount of historical data on that particular make and model of equipment that you have. I had a customer that puts some tracking information on their equipment, and generally that's done. If it's stolen, you can find it. And there's an art form to hiding those sensors where thieves can't find them.

But in this situation one of the subs went back on the job site and he was using that forklift and the scissor lift in the middle of the night to try to get caught up on some stuff. And they were immediately able to go out there and deal with that situation.

Another situation where one of the owner said my excavator is heading down the road at 60 miles an hour. It's on the back of a flatbed truck.

It's being stolen, and get the information that they arrested them.

Wade Carpenter: Yeah. And again, that's some of that GPS technology makes a lot of sense, but in this case, when we're talking about putting some kind of sensor on there to predict some of this, it is like your excavator whispering to you, like, [00:16:00] fix me before I puke oil all over your brand new port slab.

Stephen Brown: Yeah. Look at your car. These warning lights, drive you absolutely insane and you don't know which ones are important or which ones to disregard. You know the old days, the light wouldn't go off, so you just put a piece of electrical tape over your dashboard.

Wade Carpenter: Yep. I found another one that was I already talked about that, Alice AI. There's another one called nPlan, dynamic AI scheduling. Responding to whether delays those things like, the crew pace. What it's doing is allowing you to take that Gantt chart you learned and basically have AI do it for you and shift all that stuff and taking a lot less of your project managers.

Stephen Brown: Wow. Yeah.

Wade Carpenter: So it's kinda like Tetris for your timeline. You stop, fill gaps.

Stephen Brown: Yeah, I mean really it's just, they're estimating, bidding. There's a legal aspect, there's your equipment, there's your safety. It just doesn't, end, does it? And a lot of these platforms that are good, I've noticed that they just get better. [00:17:00] Because they're a platform that specializes in that platform. Kinda creepy, but wonderful.

Wade Carpenter: I can't say I've really tested any of these things so I can't vouch for them. These are the ones I came across. What I hope our contractors are gathering is that these things are out there, maybe they're in their infancy, but with ChatGPT, we've come a long way in two years.

As they said it all along, this is the worst it's gonna be because it keeps getting better and better exponential rate. But in this particular case, if we can keep our crews active, reduce the idle payroll, and boost our predictability on our jobs, I think that goes a long way to being a lot more profitable.

Stephen Brown: You're gonna wanna go with an established platform because also you're gonna wanna know what kind of agreement you're signing with them. Who does your data belong to? And what kind of safety procedures do they have in place to protect your data before you do business with them?

Wade Carpenter: Yeah, there was some other ones. This one's sort of the same kind of [00:18:00] concept. It's like predict not just when you need crews or materials on the job, but labor forecasting and matching. There's a couple of ones it's called Bridgit Bench and something called Smart Barrel.

And again, it helps you think ahead, you're gonna need three framers in week nine, so it's basically trying to match up people. It's kind of like Tinder for trades, match you before you swipe the wrong way. So, but it cuts a lot of, scrambling and wage inflation. And some of the things that we see on the jobs, they're not profitable. Sometimes we stay out on those jobs too long. And if we could eliminate any of those factors, I think goes a long way.

Stephen Brown: Yeah, this great, great stuff Wade.

Wade Carpenter: You already said this, talking about uploading contracts and flagging risk and there are some systems that you could build your own, there's custom GPTs, but there's ones that I found was like Luminance and Kira Systems, K-I-R-A Systems. But these are [00:19:00] some real technology right there that decipher all this legal gobbledygook and basically this indemnity clause will cost you a kidney.

I don't know, you shouldn't be replacing a lawyer. I would say that all the time, but it is that lawyer that doesn't charge you by the hour and talk straight. I would say get some legal interpretation, but it's a good first step. Something like, hey, this is, mean, I've, I've I've done this myself. It's like, throw the terms of service, like, act as my lawyer and protect me and say, what are the things I should be looking out for? And it's not a construction contract, I do that on like software subscription, signing up for?

Stephen Brown: Well, I mean I mean you're dealing with people, and AI can't do that for you, but look at the common sense of doing business versus what you can use computers to assist you with. That's gonna be debated forever, isn't it? I was just thinking, what if you had the entire contract with all the terms and conditions with the laws and the states [00:20:00] you were operating, and you decided to stop work on a project because the contract that you signed, they weren't complying with what they agreed to do, and you're tired of it?

So you just ask AI, what are the repercussions on this job if I just quit the project? and it'll just give you a bunch of things to think about before you call your attorney, and get their advice on it. Because I always say, if you've got a really good construction attorney, they're making you money. And you say, well, they're so expensive. A good construction attorney is making you money. And so you wanna use AI to help you get the facts to them quickly and you can see what you think your options are and ask them about it, and they'll tell you exactly what AI forgot and what you may not have been thinking about. From the law and from a human standpoint, and hopefully knowing the players involved.

Wade Carpenter: Lawyers may hate this, but you run it through there and maybe draft a first draft of [00:21:00] it. I may have told you that story that about a month and a half ago I had somebody that became a new client, but he took my contract and came back with all these questions from it and like, okay, well let me create my own addendum for it.

But he asked some really good questions and I was able to answer and address his concerns. I feel like I've got a great long-term client that seems like a great fit. But that allowed him to ask some good questions. So yeah, it was a little annoying at the time. It was like, okay, well you're questioning this and you don't understand, but I want you to

understand. So I guess I'm going off --

Stephen Brown: No, the best professionals want you to understand what you need to know. Best professionals aren't intimidated by you having knowledge. The best professionals have experience.

Wade Carpenter: The next one I had was they all sort of like fit together in pieces, but, we talked about those BIM systems and stuff like that. But there's these drone and plus LiDAR, which you know, and plus the AI surveying a job [00:22:00] site, checking the grade.

We talked about a couple times that putting it on an excavator and reducing the time it cost to grade something out. There's something called Built Robotics, again, these are just, survey drones, those kind of things. These things are out there right now and again not advocating any of them in particular. It is just maybe you should go out there and explore some of these things if it's one of those pain points that could save you a lot of money.

Stephen Brown: Absolutely.

Wade Carpenter: If you weren't out there with a tripod and a chalk line and just flying the drone to get your site plan in several minutes, it'd probably be well worth it.

Stephen Brown: And just the photographic real time images that you get from that drone from above, amazing, they say a picture's worth a thousand words. You know what's a drone image worth?

Wade Carpenter: That sort of leads into the next one I had, not just drones, but photo based change order detection, and this AI can compare job site photos or scans to flag like deviations. Okay, this wall is in the [00:23:00] same wrong place and the sooner you fix those things, the less it's gonna cost you to fix them.

Stephen Brown: Mm-hmm.

Wade Carpenter: You may feel like this is sort of invasion of privacy, but it's like somebody wearing the snitch cap, just pointing out this hidden duct that is behind drywall before it costs you a lot more money to rip that drywall back out.

Stephen Brown: Absolutely.

Wade Carpenter: But if it's is out there and you know, or you're the owner or whatever, and you're paying attention to this stuff, or AI is paying attention for you, and you're doing stuff, or your crew's doing stuff on the job, it can catch some of that scope creep instantly.

Okay, we're doing this, but that wasn't really in the specs. If you wanted that, maybe we should talk about a change order.

If we could boost these change order captures, help with audits help with protecting yourself if you ever get into disputes, any of this documentation. I think what, three, four years ago we had Chad Gill on here talking about taking pictures of these kind of things, but we didn't have all the AI stuff.

Stephen Brown: That's right. It was a brilliant idea and it still is. The last thing that we need to talk about is, [00:24:00] the procurement aspects of it. Tracking material supplies and orders. Also how much you're you're ordering in your billing, use AI to track that for you.

Wade Carpenter: AI can compare vendor pricing, surges in pricing. We've already got some things in place that manually people used to have to do. that We put in place with our systems here, working with payables and purchase orders. But, used to be a manual process for somebody to, and it still is for most contractors, honestly. You had a purchase order out there for X amount, and then the bill that you get from that vendor, always marked up $10 a yard or a cubic yard higher, whatever it is. But some of the real ones I found in this area, something called Bazaarvoice and something called Procurify integrates with several of these systems.

Stephen Brown: Bazaarvoice sounds cooler than Procurify. That sounds kinda geeky, but it's very accurate at describing what they're doing. It is [00:25:00] definitely taking the bizarre outta the procurement process.

Wade Carpenter: It's almost like you got this whistleblower accountant checking over every vendor for you. So it's like trying to let's get away from this. Oops, we paid extra. All these excuses that happen, we can avoid that. That can save material costs and the fraud risk and, use some of this benchmark data, it, it's amazing what these kind of things can do now.

Stephen Brown: Absolutely. And I wasn't dissing Procurify, I was just saying you name two. I like it. Were talking about two polar opposite names for this. If I had an AI platform, I would definitely give it a name that people could remember about its purpose, wouldn't you?

Wade Carpenter: I'd probably have AI come up with a name for AI.

Stephen Brown: Okay. Well, we could go on all day. You wanna end on that?

Wade Carpenter: Yeah, no, I think we've hit a lot today and hopefully we hit all the ones you wanted to talk about today. I t's been eye opening for me to really look into thinking about what we can do [00:26:00] on the back office. I'm constantly searching, but on the job site, that's where the backend systems as well as job site systems, it's coming and it's there. Some of these things, maybe our listeners should be thinking about and looking into.

Stephen Brown: Okay.

Wade Carpenter: And I'll tell you this too, I am a geek on this stuff, but I have taken some pro accounts, the higher end AI accounts and getting some deep research and those kind of things, to me it's well worth it if you think about it.

Okay, well one particular plan cost me $200 a month. That $200 you think about one employee would cost you. And if I could save any kind of amount of hours, that money is well spent. Especially if I use it and not just another subscription you're gonna spend money on like cable channels and you got 200 cable channels. You look at three.

Stephen Brown: It makes perfect sense.

Wade Carpenter: We appreciate everybody kicking this around with us. Stephen, this is great information. Appreciate you talking through it with me. If you enjoyed this episode, please [00:27:00] put your comments below. We'd love to hear them.

Please share, subscribe. It always helps us out. We do this every single week. And we'll look forward to seeing you on the next show.