Building with Intelligence: The Role of AI in Construction Management

Welcome back to the construction leaders podcast, the voice of leadership in the construction management industry brought to you by CMAA. Today we're jumping back into one of the most talked about topics, artificial intelligence. Although AI has been a topic of discussion for years, we've brought in a few members from CMAA's Technology Committee onto the podcast and we did this last year to really start exploring its impact on construction management and I've asked them to come back today to provide an update on what remains the hop topic in the entire world and that's AI. And if you didn't catch last year's podcast shame on you, you still have time to go back and check out that August episode and get caught up. But as a quick recap on the discussion, you heard how AI is beginning to transform everything from scheduling to safety monitoring. But since the conversation aired, the industry has moved quickly as the technology continues to quickly evolve. So in this episode, we're going to follow up and we're going to check and go a little deeper on what's working now, what's the hype versus the real return on investment, and how our construction managers actually using these tools in the field of day. We're happy to have three esteemed guests who happened to also be three of my favorite people, Jarvis, always from HNTV, also a CCM Marty Turner from Turner Townsend and Heary, also a CCM and Vinnie Testa, PE, CCM from Procrone Consulting. Thank you all for coming back and going through this with me one more time. Great to be here Nick, good to see you. So let's go right into this because AI's got so much we could probably spend all day talking about this. But since our last conversation, Jarvis, how have you seen AI adoption evolve on actual job sites? Nick, that's a great question. How have I seen AI adoption evolve on actual job sites? So you see a mix of things with safety, with coordination, with progress tracking. There's a whole host of ways that AI is being introduced. We've seen things where guys and guys in the field are wearing sensors now, either on their vests or on their hardhat or something that picks up like temperature or something like that. So you have AI monitors that'll pick up that a sensor is next to a very hot heat source and they'll send an alert to like that form in or to a project manager or to that actual employee and say, hey, look, you're close to something that's really hot. There's a high chance that you can get burned or something like that. So it's you have things monitoring safety as far as progress. We have different 3D 360 cameras that are picking up different things on site to make sure progress is recorded correctly and you got RFID sensors that can be read by the cameras to let you know if something is in the right place or if it's the right product or a whole host of ways that AI is just helping to do those checks and to monitor things in the background. Great. I also wanted to ask, as we consider these AI opportunities and construction, we're really seeing a lot of misconceptions, we're also seeing a lot of discussions of job loss. People are picturing robots. Things of that nature, Vinnie, what would you say about clearing up those misconceptions right away? Are there any things that topped mind for you? Yeah, AI means a lot of different things to a lot of different people right now, especially in the construction industry. We could talk about the procurement process. We could talk about a construction design. We could talk about commission and close out all the way to operations and maintenance. And where you have disparity in the language right now, it's really augmentation. It's not really there to replace workers. It's there more as a tool to enhance abilities. We were seeing labor efficiencies gained, making humans faster, smarter, safer on job sites with their day-to-day livelihoods. We're seeing tasks that used to take weeks to do being done in minutes and days now. And that efficiency is not replacing workers. It's making us focus on the nuts and bolts of construction and how much labor we have on site to be able to actually produce the product. We do some studies on robotics right now where we're using robot dogs and drones to do surveys and inspections. But again, monotonous task that somebody is checking that specs are being installed correctly out in the field is now being able to be done by somebody remotely because they have a robot on site and it's not replacing workers. It's making it more efficient. So I think that's one of the things that we have to clear up right away is that it's a complex term across the entire interface of the construction industry. So when you break it down by process, then it gets a little bit more detailed on that. I know Jarvis and Martin might have other aspects on it and can piggyback on that. So I'll let them weigh in real quick. Yeah, Vinnie, bring up a lot of good points there. One of the things I'm seeing obviously is how it's starting to make its way and get some traction in the administrative. And the things were such a data-driven business, right? Cost data, schedule data, quality data, safety data. It's that ability to organize and analyze that data that's really starting to really showcase its power and make it a useful tool to our industry, right? But I think to hedge off that concern about replacement of jobs, I've always seen it as a tool. And to Vinnie's point of augmentation, a supplement, I think it's really an opportunity for the industry to double down and focus on the basics, right? Really making sure people understand what good looks like, what a good schedule is, what a good budget is, what a good cash flow is, what is acceptable, safety data. The tool helps you move things faster, just a backhoe helps you dig quicker, right? But if you're digging in the wrong spot or you're doing the wrong things, bad things happen and I think that's the same holds true for AI or any tool or application. Yeah. And to jump on this too with Vinnie and Marty following you two up, that's made some great points. I think it's a great question. And I think just to pick you back on what Vinnie said, I think the fear of robots coming into replace people is stop short of saying an irrational one. But when you think about it, you have guys that are doing like, let's say confined space inspection work or stuff like that with hazardous materials. And now if we can send the robot dog in to do those inspections, we have a net gain on lowering our instances of having safety incidents on site or having something like that. And then there's not a loss of a job there. Instead of sending a guy in the tube to do the inspections in a confined space, now we train him to run the robot sitting there, we're not asking him to build it or program it. It's already in the sealed. It's ready to go. We just need you to know how to drive it, how to record the video. So instead of sending them in the tube, now he's an operator of that tool that could go in their form. And now we don't have the safety risk, but we still get the same result back with the inspection work. I think it's just a pivot in how we think about going about our day-to-day and doing a job with these new advanced tools at our disposal. And to the benefits to the project management team, right? It allows us to kind of offload some of the mundane administrative tasks and allows us to get back to doing the fun stuff like getting out in the field, using your eyes, putting eyes on the problems and solving them, collaboratively. So it's a huge time saver. And I'll add two examples there that I've seen more recently. So where it's accelerating abilities right now, I've seen a scheduler do a deep dive on a schedule looking at where logic ties were broken, looking at duration changes, looking at milestones that were inserted. And they pulled all the data, they did all the schedule review like they would. But then they used AI to create the report. And instead of taking it another week to draft a report from all of the things they identified, it was done in 15 minutes and they were putting a bow on it that afternoon. Things like that make it so that we're not losing the technical ability, but the detailed write-up information of sentences or actual factual logic is where it's enhancing the product to be put out. Another one that author out there, and this is more of a fun one, I've seen a really talented pull of inspection out there from dealing with usace and nafak and a number of other entities. And I've seen inspectors that are great technical people out in the field and they'll go out there and they'll do their thing, they'll do an inspection, they'll write deficiencies, they'll do observations. But then when they go to send an email, it's the end of the day, it's the end of their shift, and they're not very technical when it comes to writing a proficient, well-worded email that captures all of the data from the day, or let's just say their inspection report them the day, summarizing it all. So they can go, quote, "spec" and "verse" from a deficiency, but to put it into a sentence and make it formulate in a way that a contractor or a subcontractor or a design architect will be able to read it and understand and interpret, they're using AI and they can piece some of that stuff together. Usace has their Jackie system that they just implemented, but that's a big one. It takes basic knowledge from field use and policy and procedure, and then an inspector can type things into their system and say, "Hey, how do I do this?" And it gives them that verbatim from a knowledge pool that is way deeper than any one person or 100 people could have. So it's enhancing that quality product from technical people that may not have the full skill set of communication or other things, and it's making them have that full breadth that they're fingertips. Again, I think we're seeing a lot of efficiencies gain and not really replacing workers have a question originally. Yeah, and piling on to that a little bit, the idea is we're still a relationship driven business and our clients, they depend on us, you have to formulate that trust and that's earned over time. I think the tools help us get the data in line, and again, you still have to know a good look. like because you can't just trust the outputs like you wouldn't let an intern write your letters or whatever to the client. Similarly, you have to vet that. And I think it still requires people in positions to carry the day with that trust and relationship building. But the tool makes it go faster, more effective and efficiently. Y'all mentioned augmentation and increased efficiency and how some of these job loss fears are not super justified. But we're seeing outside of the construction industry, environmental and ethical concerns about AI. Has that permeated at all? Marty, I think we'll shoot you for this one first. Has that permeated at all into the construction industry? Absolutely. I think especially in working with the bigs, it's reputation and risk mitigation is always at the forefront. You have to protect that brand. I think if clients thought we were phoning it in and only reliant on just a tool-driven response, I think they would look for a different solution. But I think, again, it's pairing the tools ability to organize, manage, and collate and create metadata that gives you those tools to make better decisions. But again, it really comes down to the trained professional, the certified professional, if you will, to carry the day. They have to know what good looks like. And again, I go back to my earlier point. I think this is an opportunity for the industry to see the change in the tide with technology and really double down on making sure the folks coming in aren't just technically proficient, but they understand the foundational functional elements of what makes us go cost schedule risk, a safety. They need to understand those organically so that they can apply the tool, make it more effective. Evan, I'm going to throw a curveball at the group here, but there was a big debate last year that I was following on AI and creation of art and digital media. I'm sure you guys have heard a little bit about it. But who's the artist? Is the AI the artist or the person that's typing the prompt? So the question on ethical, there's a lot of times where you could see somebody just typing in a very well-worded prompt and then it puts out a product and then they're submitting that product. Is that author that creates a book and puts it out there that way any less different than the author that wrote the whole 180-page book, 300-page book, would AI have created that book without having an author do it? Same thing for music. Is somebody creating a melody that has AI written all over and the melody was created by AI, but they were telling it what to do at what time speed and what instruments to use all of those details. So I think we're running the same thing with what we're talking about for the construction industry and I'll go back to what I said earlier. I think it's really enhancing our abilities because you have to be able to put out a good product in the first place and know what a good product is to be able to drive AI to get you there. But anybody can put a prompt in and ask AI to create something, but you have to go in and tailor it regardless of what it gives you. And if you're not tailoring and editing it from there, everybody in the world can see its AI. An example, I'll throw out there and I'm sure we'll get some chuckles here, but I've been doing a lot of interviews recently for a bunch of positions. And you can tell when you get an AI resume because it reads exactly like your requisition that's out there. It's 100% just reads like it. So this is the perfect candidate. Oh my god, where is this person been hiding? How are they out there? So you can see where people were doing that, but you can see the artist really behind it when you start to see it get tailored to a final product. Jarras, I want to switch gears here a little bit. You've been with us for a while now and you've helped me in so many ways throughout my career here at CMAA on different things. You've always been technologically savvy, like you were talking about wearable technologies last time you were on this podcast. What are you, do you see the biggest barriers to construction teams are still facing when trying to implement the AI tools? And are there sectors that are more successful at adopting AI than others? It's the apprehension. And I will say our industry as a whole is a little bit more skeptical of introducing a new method or a new tool or a new technology when we've been delivering multi-million dollar, multi-billion dollar projects and programs for decades. When you introduce something new, especially in our industry, it's always met with I would say more skepticism than normal in other industries. We operate on the trust but verify. And I think everybody in this industry has heard that on every single project is we trust but we verify. And it's the same way with this. I'll go back to Marty's point is that you have to know what is right or what isn't right or what looks good and what doesn't look good. What I see a lot is you have two factions that you have some people that are like yeah let's bring it in let's see how it'll work and let's pilot it out and if it doesn't work out then we'll scrap it. And then you have the other side that's what would that mean to productivity? What would that mean to this? What does it mean to the bottom line? How much does it cost? Everybody that finds any reason to scrap it. You're swimming against the tide in that sense but what I find is that usually you can get a field office if you can get to one project engineer and say hey look you try this out. Pick one inspector, let them try it out too. The two of you work together and let's see what your outputs do compared to traditional way of doing things. And then we'll see where the benefits and where the shortfalls are with that's I've been in that works a lot better in our industry and that tends to resonate a lot more with a lot of the owners and clients too. Yeah Jarvis we all know on the committee that our industry is the laggard right we are incredibly slow adopters and a lot of that comes to skepticism and the return on investment convincing making the business case all those things right. But I think what's different about AI or maybe it's just kind of building on the bow way of created by some of the others. But I think people are really seeing the competitive value and how it does translate even those small gains can add up to big wins for the bottom line of a project or a company. And I think it's that competition and it's the risk of becoming irrelevant really that's driving this change. I think everybody's saying I got to give this a hard look. I can't afford not to give it a hard look. I think we're at that tipping point really where the bigs are starting to do it they're overcoming their fear of ethics and risk and those type of things because they just see the value add. And Nick to really run home the second half of your question is you see the adoption a little bit faster on the administrative side of things when you can load in an RFP and prompt a large language model to say look this is what I need you to do where we need to be responding to this thing and you can load in an RFP and then you can ask it questions and you can really dial in how you're going to respond to this thing and really and it can really assist you with that or like with technical emails like how many stated before when you've got multiple stakeholders let's say it's a progressive design bill so you've got multiple teams in here that are all trying to figure out a problem and you've got this big powerful AI tool that can also dig into the details of a lot of the documents that are coming through some of the field reports and they can make those linkages that you may miss and we're all professionals we've all done it before we've all been out there and delivering work but you know I think we can all admittedly say they are sometimes where we've missed things and come back that's why rework and that's why change what is our thing so you utilize these tools to help in the background and I see the adoption a little bit faster in the administrative side of things than the field side of things you brought up a really cool point Jarvis the tool itself is different than other tools it's not just helps you produce things or organize things you can use it to ask questions and boil things down it and respond back to you build up a technical case now explain it to me in layman's terms that's never been able to be done before that's that very robust interface application where you can ask produce and then okay pick holes in this tell me what the flaws in this all right and that's huge before you just could build something and you had to start from scratch to figure it out but then it definitely makes sense that it'd be more an administrative side up front as people get used to using it and moving forward Jarvis you've mentioned some apprehension and Marty mentioned some skepticism which makes sense with any new tool in any industry right but our construction companies or governmental agencies developing guidelines or policies to help limit AI use to specific use cases are you seeing that in one side or the other in is one part of the industry leading the charge in this or as private industry may be outpacing the government or vice versa let me jump into that one and I think we're seeing a multi-pronged approach you're seeing the government take initiative and do things on their end you're seeing RFPs come out with specific questions with AI using the development of your proposal we're seeing multiple entities on the architectural engineering side and on the construction management and GC sides also using and creating their own tools we're using AI and learning language models I think policy is being written as it's being seen what use cases are being developed for it so right now you have explicit use across multiple federal agencies where they're using it as a learning model for their employees and they're using it as a bank of data so almost educational platform to help level their teams where you're seeing regulation come into play is the office of AI strategy so you have the DOD's strategy going around this and they're starting to put policy into place they're starting to framework some of the things that are going on you have digital Transfer. information that's happening at the highest levels all the way down into the field. So you're seeing policy being put in place, some places, you're seeing it written into RFP's and proposals, you're seeing it written across the board, and then you have agencies like CMA that are putting framework around it and having these conversations to make sure that the industry knows it's coming and knows what's happening. So I don't think we have any policies in place, Nick, but I think we're putting best practices in play, we're discussing it at large, so we're talking about it to make sure that as the landscape changes, conversations like this are happening, we're paying attention to the news, we're reading about how it affects the industry, and it's going to affect other policy. Once it goes into play, you're going to see how other policy and procedures change across procurement, across delivery, across closeout, everything from the construction industry. And just to jump on that, Vinny is the best analogy I have is we're building the plan in the air with this one. As far as actual policy, actual things are going on, I've seen it be as broad as two sentences on something that says, "Hey, if you're using AI for this, you must get approval from the IT department." And that's literally the policy, so if you go in, depending on how you describe what you want to use it for, it could be a yes, it could be a no, but there's no hard lines that we have to stay with and there's no swim lanes that are set, at least that I've seen yet. And it's on a client-to-client company-to-company type basis, depending on what project or what environment you're working in, or what mindset, a particular firm or particular agency has around it, you may have a whole lot of leeway with it, you may have very little. I can say, as a firm, we've authorized co-pilot to integrate into a lot of stuff. That's one way where we're a lot more open to that. But when you get into some of the other ones, it's a still-in-progress. It all depends on what the mindset of the client is and how they're pretty much driving me. But the best analogy I can use is building the plane in the air. Yeah, Evan, to answer your question about who's leading the charge is it the public or private space? I think it goes back and forth, and I think it's driven by need and also by funding. Whoever has a particular need, if AI meets that, ask, or whatever, then people are looking at that as a solution. It also comes down to funding, are you willing to put your money where your mouth is? That's what I think drives it. Yeah, that brings me to my next question. How should construction managers evaluate whether an AI solution is worth the investment or not? And can you share specific examples where AI has delivered measured results on cost-saving scheduling improvements or reducing risks? I'll give a shameless plug for one of our close friends, Frank Lazaro, who we've presented with, but he wrote a book called Finding 12 Minutes. And I think it's those small gains when it's like Jarvis said, when you demonstrate on one-on-one case with that project manager and his inspector or whatever, it's winning the day at the grassroots level, but you also have to win it at the boardroom level, but you have to show how this is saving time, saving money, making things more efficient. Because I don't have to do this, I'm able to do this higher-leveraged activity. That's the value. Being able to capture that in a meaningful way to create that business case. I'll add there, Marty. I think, you know, his book on 12 minutes a day, shameless plug again. I think it hits the point. You just got to start using it and it's being integrated into all of our tools. It's in Microsoft everywhere. It's in Google everywhere. All those tools that everybody uses without thinking a second thought about it's being plugged into all of them. When you go to type a Gmail email now, there's a line that says polish and you can type in certain criteria and certain things and have it formalized your thoughts a little bit better. You can approach that in a word document. You can approach it in any document you're using right now. It's pretty much in everything we're doing. The finding 12 minutes is really around just start using it and start finding the efficiency so that you can get out of the mundane tasks into the more high-value, high-profile tasks that you need to do daily. Really suggest anybody read that if they're interested in getting into AI just because it really unlocks the beginning of what you need to be doing for your journey. For those folks that still haven't dipped their toe into the AI pool, we've had a lot of discussions as a committee. I think a lot of times it starts with going rogue a little bit, putting chat GPT, the free version on your cell phone, and trying it for a home application. Simple things like help me plan a trip or, hey, I'm in Washington, DC with my 12-year-old son. Give me places to eat and things to do. Yes, you could search that on Google, but it'll put it into a time-activated scenario that you can literally find super useful, right? Again, the applications, once you demonstrate its power and its capability on a personal note, you can start quickly find a million use applications and how it can propel your business, right? Start small, grow big. I did a presentation one time and I used it around just using digital tools in general, and necessarily AI, but it definitely applies to AI also, is if imagine yourself sitting in a meeting that you have regularly, let's say this is a weekly meeting, there's 20 people in the meeting, and let's say the average cost per hour for each individual is 200 bucks. So, let's say we have 20 people times 200 bucks an hour, and this meeting usually goes for two hours. We have it every week, right? Let's say I have identified that at least 20 to 25 minutes of that meeting every week, we spend with people fumbling through emails or fumbling through reports trying to find something to share up on the screen to talk about to say, look, this is this, oh no, this is that picture, hold, give me a minute, let me find this, or give me a second, and let me find that. Whereas, before the meeting, we used AI to identify, hey, look, pull out all of the instances where we've had RFIs for the past week on this project, and now, like Marty said, you type in the prompt, you say, hey, look, now, link all those RFIs to any change orders that's linked to this issue, or any potential change orders that have been sent in for this issue. We get to any project photos that we're taking within this geographical area, because we have metadata on the photos that we take when we use our smartphones, and we link all of that, and we bring that to this meeting now. So every week, we save that 20 minutes that we spend fumbling through everything. Now, you multiply that 20 minutes that we save weekly times 52 weeks a year, times three years for a project, how much money did we save? And when you put it like that's where the C-suite, and that's where, you know, clients and everybody understands, okay, this can really save us a lot of money, but it's those little things. It's the every inch, one of my favorite movies is any given Sunday, and it's like you call for every inch, you want to win that game, where if you take all of those inches and you line them up, we have saved a lot here. We've shown the efficiency, we've shown the cost savings, and I think that's the most important point to run home with a lot of the people that show skepticism around these things is that's where you truly find the additions to our skillsets in our industry, and that's where you truly find the efficiency is in situations like that. All right, I think we've covered a lot so far. How do you all see the next six to 12 months shaping up, and what do you think is going to make firms rethink their AI roadmap, Jarvis, we can start with you. So, I think what's in the next six to 12 months I see, and with all of us, we're all technology focused guys, and we all focus on digital adoption, so you see a lot of the tools that we use that in AI functionality to it. A lot of the PMIS platforms are adding an AI feature, a lot of the cost management platforms are adding an AI feature, so you see a lot of those digital tools that we use are seeing the value in adding an AI arm or AI functionality or something like that and ingrained into their offering, so you'll see it in the tools that you use every day, and then as a firm or as an agency, you'll say to yourself, our stance has always been ex, but now we use this tool every day to manage our financials, and now it has AI built into it. Look at how it has changed my reports, look at how it has changed our outputs, I'm finding things that are going back that we could really use in an audit situation because it's making leakages back to expenditures here or missed billing cycles by contractors that has thrown our numbers off and it's finding the things that we get deemed on for audits, and this is a tool we use every single day. So in the next six to 12 months, I really see the, I won't say the software vendors, but like the tools and the vendors that make the software that we use every day, really driving being a catalyst to more adoption of AI in our industry. Yeah, Jarvis, I think you hit the nail in the head, I think that's the next logical step for AI is it's driving tools and applications, being able to queue it up to engage a software that's where it's going, and in terms of Evan's question, what's going to drive that? I think again, it goes back to my point that the competitive edge, I think businesses have that awareness that this isn't like anything before everybody is starting to see the value, and it's that fear becoming irrelevant. If we don't figure this out, we're not going to be around. Our competition is going to knock us out of the game. So I'll add there to what is six to 12 months is like near term, right? What's going to happen tomorrow? I think you're seeing a lot of companies adopt and you're seeing a lot of smaller companies adapt a lot quicker because of the less red tape. I think you're going to start to see a leveling of small businesses up to large businesses based on their adoption of AI. and you're gonna see better material, better quality products coming out of smaller businesses that can then keep up with the better polished products of a larger company that maybe has design team and a proposal team and a quality assurance team, whereas a small business that has 50 employees maybe has a CEO that's wearing all of those hats. So you're gonna see better quality product coming out of smaller firms. I think larger firms are gonna start to adopt. Obviously they have more funding, they have more manpower, they have more intuitive growth in these areas. I think as they start to dive into them and embrace them more, you're gonna start to see better products that make them more efficient and make pricing go down on their quotes that they're sending in. So when they go to bid projects, you're gonna see a little bit more of an advantage innovation idea coming in. I think that in the longer term, if any of you watch stocks and follow who's trading what, you see some very interesting investments by Nvidia or other groups like that into AI tools. And so they're investing millions of dollars into some of these smaller AI startups that they're gonna have application for them within the next year. And we're gonna start to see that plan a large scale. So it'll be telling to see how the world evolves, not just the construction industry with those large investments being made and what efficiencies we're gaining as a whole. 'Cause construction's more about nuts and bolts, but when you think about it, you also have financing you have IT and procurement and everybody else involved from a holistic owner standpoint. So there's gonna be efficiencies and ties gained at those levels as well with those adoptions. So it's gonna be a huge impact. It's just a matter of when it's coming, how it's coming. - I can tell you, I've got at least three pages more of questions. So do not be surprised if you get a call from me later saying let's do part three of this and a couple of months and see how we are because this is changing so rapidly. But Marty, Beni and Jarvis, thank you for always representing CMAA well in the Technology Committee and it's always a pleasure to talk to you. Lastly, I wanna leave you with two more resources as a listener. One is that the Chartered Institute of Building Technology, CIOB, a partner of CMAA, has an AI handbook. They released in 2024. That has a lot of good information out there too if you're looking for more information and tools. And the second is to make sure you attend CMA conferences since the conferences have an entire technology track and those tracks are developed by forward thinkers just like Marty, Beni, and Jarvis. - In the next episode, guests from Jacobs and hisgenuity join us to explain how you can develop and deploy high performing teams on your projects. If you're looking to boost efficiency and build better, you won't wanna miss this podcast. To be sure you don't miss their next episode, please subscribe to the podcast and follow us on social media at CMAA_HQ. Don't forget to leave us a review of your thoughts on today's episode and let us know what you would like to hear next. On behalf of CMAA, thank you for listening.