So, at the end of a project that would typically take three years to create all the details of engineering, that could get compressed down into three weeks. So, when you’re looking at those time frames and three years is becoming three weeks, that opens a whole host of possibilities, right? All you have—every detail of your project is planned and prepared in advance, all the RFIs are answered before construction, you have a full construction package, all of the constructability reviews, and commissioning ability reviews—everything is answered up front, and there’s zero ambiguity because everything is defined upfront. All of your risks are eliminated before you even put a shovel in the ground because you have a full definition of your project before you even start. So, this gives project owners unprecedented confidence in their project before they even have to make any financial commitment. If you can do three weeks of work and develop your project and prove it out to know that it’s financially viable before even making a financial commitment, before awarding any construction contracts, that revolutionizes the industry when you can take advantage of the information systems that are out there to rapidly design your projects. So, the key is an information approach here—an AI-driven design right at the outset of projects. AI systems, while they’re not currently capable of doing that right now, it’s hard to say when they will be—if that’ll be in six months or in six years—but the scenario seems very likely where AI will be able to generate all your preliminary designs, all your detailed designs, being orchestrated by project leadership.
You’ll be able to do a full simulation of your entire project—think of it as a digital rehearsal—where you can not only design and construct your project, but you can also see full functionality of the project before we even do any testing in the field. Everything will be tested in your AI model that you created in three weeks, and that’s pretty powerful. When you can create a system like this in three weeks, that means you can do it once, you can do it 10 times, you can create 10 different versions of your project and compare scenarios to find out which one has the most optimum performance, which one is most financially viable, and do those comparisons amongst 10 projects before you even decide which one you want to build. You can validate everything up front when you’re using information in this way to design your projects. So, you can review all the operability, all the commissionability, all of your PLC logic files are already created and ready to go before you even start construction, and it’s a pretty powerful world that will be coming here shortly. Like I mentioned, you could create several versions of your same project; you could have a digital twin with a generative design, an iterative design approach. You produce a design maybe in three weeks, and you can tweak it, optimize it, and do a real-time simulation to find the optimization throughout—not just the project, the entire project lifecycle.
You can have dynamic digital twins that can represent every aspect of your project for full functionality, and with this generative design approach, then AI—you can use it to rapidly explore multiple configurations and determine the best approach before you even put a shovel in the ground. So, what this means then is that the construction phase is going to be exclusively focused on execution. With every detailed piece of information available up front, there’s no ambiguity; construction teams know exactly what they need to build, and the contracts can be written so they can focus strictly on execution—so installation in the field. Because everything is precise with AI, everything’s defined up front, every RFI is answered, there’s no change notices because that’s all been done up front through the interactive generative AI design process, and construction focuses exclusively on execution—no design changes that are late-stage in the project. This takes away all the ability for rework or claims because everything is defined up front; everything is fully defined, and there’s no ambiguity, so there’s no excuse for smoother execution during construction. What this means then is, with this generative AI design approach up front and all of the design—every detailed engineering piece of information defined up front—then there’s no need for EPC contracts. AI completely eliminates the need for the EPC contract model, and it’s likely that we return back to a design-bid-build with the full design that’s complete. Now, in the past, it did make sense to have EPC contracts—to have the construction and the engineering part of the same group so they can collaboratively work together over years. It was risky to try and complete all of the design and then bid that for construction, but with this generative AI upfront approach where every detail of the project is defined in advance, then we can go back to the design-bid-build model because the design is 110% defined up front, and you can bid that work for strictly installation.
So, with construction-only contracts, there’s a lot lower risk on projects to the owner, a lot lower risk to the contractors because they know exactly what they’re going to build, and there’s no changes that are later in the project. So, the contractors that win are the contractors that are able to focus on executing, and they’re not the contractors that are filling up their pockets by submitting claims on the project and using that ambiguity of fine details to their advantage to submit claims. So, this completely eliminates the claims scheme; the focus shifts entirely from that claims game that often takes place on projects and it focuses on quality and precision. So, the poor-performing contractors—they won’t be able to hide anymore. If they’ve got business models that are built on submitting claims at the end of the project, they’ll quickly disappear from the market because it’s the high performers, the excellent contractors that are out there that can install to quality and precision requirements and meet the contract requirements—they’re going to thrive in this transparent, efficient system. And the claims-focused business model just won’t be a viable business model anymore; it becomes obsolete, and those contractors that are following those business models will quickly disappear on the market, which is ultimately what needs to happen, right? We need high-quality groups that are focused on excellent budgets. So, this one may be a bit further out there—you see things like Pi passwords after this robot. Maybe in the future, AI gets really good, and we’ve got humanized robots to perform some of the repetitive tasks like tying rebar or pouring concrete. Now, I’ll leave this for maybe a future discussion. If we focus on the information aspects of AI, then it’s quite plausible to see that the engineering design aspects and the information aspects are significantly impacted with AI, and then this would be the next step where genuine robots are embodying that AI and actually doing work in the field for construction. That could be a future step after that to create an industrial workforce, but we’ll leave that for another discussion another day. So, the equipment procurement could still be an issue, right?
It’s only going to help if we have construction groups that are focused on quality installation if we can get material to site when it’s needed to prevent those delays as well. So, AI can revolutionize the equipment procurement processes as well and the supply chain that construction projects need. AI can manage vendor networks and actually do real-time analysis to see which vendors are in a good position to be able to supply equipment versus others that have a backlog versus others that are underperforming or are historically not able to meet schedule dates. This entire supply chain, this logistics working orchestration, allows for smarter, faster, and autonomous supply to projects. Other industries are already using this for warehousing and supply chain management, so it’s only inevitable the construction industry uses these technologies as well so that we can optimize our delivery to site and make sure that we actually have equipment to site when it’s needed for installation. Take it even further, and you can implement blockchain technologies for smart contracts so that you can manage your supply chain in more dynamic ways, leveraging some of these technologies. So, with an upfront design aspect using AI agents and an installation based on 100% of the design being developed up front, then commissioning becomes the cornerstone. Commissioning must make sure that this is all working right—that we’ve got solid designs, that everything is being installed in the field, and all of these AI designs and installations still need to be verified with commissioning—real boots in the field to verify installation and accuracy of all of these systems to make sure we’re getting quality products in the end. We can significantly reduce the risk of commissioning by having a digital rehearsal at the beginning of projects that essentially commissioning is the real-world validation of those digital rehearsals.
Commissioning being in the field to make sure that what was rehearsed earlier actually is being implemented in the field. So then, all of that asset management system that comes out of this—a lot of that information is created up front in our AI agentic design process, transitioned through construction. Commissioning is then processed to ensure all of that asset management information is gathered and archived in the proper systems for use during the life of the facilities. So, commissioning is really the digital thread that’s going to connect our agentic designs up front to our asset management systems at the end of the project for use for many decades. So, really, AI is kind of producing this commissioning sandwich where we’ve got our AI-driven design up front, compressing three years of work down into, say, three weeks or a much shorter time frame. We’ve got construction execution in the middle here that’s still required to hold equipment and pull wires and connect all of our systems together, and then a rigorous commissioning process at the end. So, we create this data sandwich with our AI-driven process at the beginning and our rigorous commissioning process at the end to ensure that data is linked from start to finish, and we’ve got systems being installed correctly in the middle between these two engineering functions. So, by having this model, it creates much higher accuracy because everything is defined upfront; there’s nothing that’s pushed later into the project as typically has been done to kick the can down the road to commissioning. This is a much lower-risk approach because everything is defined much earlier; everybody has access to all the same information, and there’s no ambiguity, no hiding behind claims or risks. We get much more predictable outcomes on projects and much stronger asset reliability when the agentic transformation that we’re in takes place. So, it’s really a question of watching these trends and seeing when the systems get good enough to be able to use on our projects—or there are others that are going to just resist this and say, “Well, this isn’t going to happen.” It could be possible to believe it’s a long way out, but like I said, just look at the software engineering industry—it was transformed within less than 12 months. So, act now and start watching these trends and seeing when these systems get capable enough to be able to implement. I know there are companies that are already working on this—from the front end of projects through construction to commissioning.
The systems are already being developed, and is it going to be six months or six years? That’s the trend that we need to watch, but one thing is clear: with any inaction, if you’re going to ignore any of these trends that are out there, then project teams will become obsolete pretty quick if they’re not keeping up with these trends. So, the best time to start watching these trends was yesterday, and of course, the second-best time is today. Definitely watch some of these trends because I think these changes are going to happen a lot quicker than we’re anticipating, and this is likely the change that the construction industry needs to drag some of the dinosaurs in the industry, kicking and screaming, into the new world. You just won’t be able to refuse and deny, right? Because the companies that are leveraging these technologies and advancing their engineering design from three years to three weeks, of course, will be the market leaders, right? And if you just refuse and still insist on doing three years of design, well, you’re not going to get too many projects with that trait. So, the question always is: “Is AI a bubble, right? Is this all a bunch of hype?” You can see in real-world situations that it is real—that real things are being transformed overnight. Just a few days ago, last week, ChatGPT released their new image generation model, and it’s phenomenal; it’s a step-change better than anything they had before. So, there are still lots of gains to be made in the AI industry, and it is transforming real things in the real world. In the software engineering industry, for example, like I’ve mentioned—there are real changes.
The money that’s being plowed into a lot of the infrastructure by some of the big companies—this is real, and it certainly does not seem like AI is a bubble. It is making some huge impacts in the world, and the construction industry must evolve. I know that the construction industry does not want to and has refused, but it will be forced to change with these sorts of transformative changes that are coming our way. The winners will be the companies that can embrace this agentic transformation, and they’re the ones that will thrive in this industry, but the losers will clearly be the ones that are clinging to the past and refuse to change or adapt. Alright, so I’ve got a few resources to help you out here as well. We are releasing a new line of courses; you can check them out and see the different modules that we’ve created. So, there are modules in there for commissioning phases and handover module, commissioning safety strategy and planning, digitalization—like we’re talking about here—as we watch some of those trends, construction completions, on-site testing, project handover, in-service lessons learned. If you need any help with any of your stages of commissioning, there’s a module there to help you, and you can check those out at ixa.net/courses. So, another resource—if you don’t already have a copy of the ICA Global Commissioning Standard, you can get a copy at ixa.net/join. Get your hands on that document; it’s a nine-series set of documents, and they are a pretty good guide to help you through commissioning each of the phases and give you that standardized, proven process to complete your projects. This will be essentially what our AI agent transformation will be based on—this commissioning process starting right at the beginning of projects right to the end. Now, in some cases, when you’re having discussions on the standard, not everybody wants to see a full nine-set of documents. A good tool we’ve created here to help you have some of those discussions with your project managers or others is a one-page brochure to tell them what the standard is about. You can check that out at ixa.net/brochure. Let’s get into some questions here; I’m sure there should be some interesting discussion with this topic that we’ve discussed today because it will be quite transformative, and I’m sure you can see that there are going to be changes in the construction industry, whether some people like it or not.
Q: So, AI is a good way to start your checklists? It gives you an excellent insight into what needs to happen; however, you need to tailor it to your project specifically?
A: Yep, that’s absolutely true. The technologies are already being used right now, like you’re saying there, right? You can go into ChatGPT and you can say, “Give me a checklist for a three-phase centrifugal pump system,” and it’ll give you a pretty good list, right? But, like you said, it’s maybe only 80% complete, and you have to go tweak the last 20% of information there. But if you can get a head start by getting 80% of what you need already and then just adjusting the last 20%, that’s a huge time-saving. So, yeah, these systems exist right now, and we need to be using them on our projects because it saves significant time. You can get that 80% head start and then tailor it specifically to your project.
Q: AI-driven design streamlines commissioning by simulating operations early, allowing teams to validate PLC logic, troubleshoot issues, and optimize workflows before site execution begins—sounds exciting?
A: Yes, for sure, and yes, AI will change a lot of the information aspects of projects. So, the upfront engineering and design, the detailed engineering and design, are probably the first aspects—the information aspects—of projects to be transformed. Like I mentioned, the physical aspects of projects—so installing stuff in the field—that may come secondary if we get some of our robotic systems developed, but the first thing, clearly, with the transformation in the industry will be the information aspects of projects—being the knowledge workers, the engineering design aspects.
Q: I’m going through this challenge now trying to systematize P&IDs and single-line diagrams, and they are not matching up with the tag level, so I have to send it back to engineering, which delays completion of proper tag list development and cross-referencing?
A: Yep, so with these AI agentic systems, really, we don’t even create P&IDs anymore, right? The reason we create a P&ID drawing is we split it up into, say, 20 or 30 different pieces only so that it can fit on a piece of paper, right? That’s kind of—when you think of a system, an entire system—it doesn’t make a whole lot of sense to split it up just so that it can fit on a piece of paper. So, our process and instrumentation diagram will be one complete system modeled within an AI software system. So, there’s no need then to chunk it out into a bunch of paper pieces; it’s within the system, and you would view it through your 3D modeling software to see your entire P&ID and track it through there. So then, when we’re trying to systematize, we would be doing that within software too—or have AI take a first cut at it, and maybe it gets, like we’re talking about, 80% of systematization, and then we would go in and do the last 20% to make it exact for what we need for startup on site. So, the systems will take out a lot of the legacy paper processes where we’ve stuck things on drawings just because that’s what we needed to do to fit them on an A2D drawing. Everything will be modeled in software right from the beginning and may not even ever end up on a piece of paper or split into these document segments. We’re entirely managing the project through our information set created at the beginning from our AI agents, managing construction right from that, and then managing commissioning right from that. So then, there’s no need to be loading commissioning software because this system that’s built right at the beginning of projects is your data management system and is your commissioning tracking system. When we split up all of our P&IDs into the various systems right in the software, then that’s how we’re also tracking completion through our cutting commissioning workflows. It’s quite exciting where we can skip the middle part where things end up on paper, and then we have to get them back into systems. We start from a digital system right at the beginning and carry right through to the end.
Q: Ohh, yes, procurement—always a challenge?
A: Yes, for sure, and we need to address that challenge as well because we can have the best digital systems on-site, but if the supply chain feeding into that isn’t on board with this and isn’t able to meet deadlines for deliveries to site, then that’s not going to work, right? So, an AI transformation of the supply chain would happen as well, where a lot of the supply chain challenges are still due to human error, right, or unknown issues. But when your AI agents can perform that work more precisely and more accurately, then it will essentially be managers that are managing those teams of 10,000 AI agents for procurement to make sure that systems and equipment are being delivered on time.
Q: Now, the lessons learned are a really important moment of projects, and I believe that this already reached the maturity necessary—the concepts certainly have. I see—I would say most projects are gathering lessons learned, but I would also say that most project teams or organizations are pretty terrible at actually applying those lessons learned. I’ve been through lessons learned on every one of the projects that I’ve been on, but then you go to the next project and you start it up, and it’s all the same challenges that you’re encountering that nobody seems to actually learn from these lessons learned. So, yeah, you’re right—the systems have the lessons learned, I think, as part of the system, but applying the lessons learned still has a lot of work ahead of us. We’re good at taking the lessons; still, some work is required on learning them.
A: Yes, absolutely, Darren—yeah, that’s my point as well. It’s great to gather lessons learned, but then if you put those on the shelf or save your file in that digital archive and never look at them again, then yeah, you’re banging your head against the wall on the next project because these lessons were already learned. That was the intention with the ICA Global Commissioning Standard that we put together—these are solved problems. This is not new; these commissioning processes, these best practices—these are solved problems. So, everything that’s in this standard is decades of lessons learned that people have had on projects, and there’s no need to learn these lessons again. These are solved problems, and all of the processes that need to be followed for commissioning—they’re all defined in the ICA Global Commissioning Standard. So, don’t try and go invent your own process because there’s a lot of knowledge, expertise, and decades of experience that’s gone into that standard. People often think that maybe they know better, but these are solved problems that you’ve got to leverage—the collective wisdom of the industry and the collective wisdom of people who have already gone through these challenges and have produced these lessons learned and these best practices. There are reasons that they’re defined in the standard that way because those are the best practices to reduce risk and allow projects to be completed as efficiently as possible. So, even though people may think they know better and that this doesn’t apply and that they’re smarter than the group of people that put that standard together, I assure you that these lessons have already been learned—these are solved problems. So, definitely stick with the standards because that was our intention—to give you these lessons learned and help you learn them so that you don’t have to make the same mistakes on your projects.
Q: What is an AI model to drive during FEED project scheduling, change management, and punch list—still consistent to the content of the project contract?
A: AI models aren’t currently this sophisticated, right? If you look at that Meter report, it’s still basing the doubling of AI progress based on a 50% success rate. So, a 50% success rate obviously isn’t sufficient enough for projects, but that will increase, right? So, when it increases to 80%, ninety percent, 99%—at some point, it will cross and be better than human capabilities to perform FEED activities for project scheduling, to eliminate change management altogether, and for some of our on-site test processes as well. But for the upfront engineering aspects—the knowledge, the information aspects of projects—it’s only a matter of time before AI agents can do that better, quicker, and more efficiently than humans. So, the AI models aren’t there yet, but it will happen, right? We’re on an exponential curve here, so at some point—is that in six months or six years?—we don’t know, but it’s on that trend for it to happen for sure. Another point there is “still consistent to the content of the project contract”—the point is that all of this agentic transformation and upfront information means that we can write better contracts. So, it’s not that we want to conform our commissioning processes to the contract; it’s that we want to get even ahead of that process before we even write that contract so that we can write better contracts in the first place. A lot of the commissioning problems we see on sites are due to poorly written contracts. The contract obviously precedes commissioning, so if we’re stuck with a bad contract, then it’s very tough for us to perform proper commissioning functions. The point of all of this is to get all of the information upfront with this agentic transformation and allow us to write better contracts. Contracts, unfortunately, out there are often written as a tool to start the project and kind of miss the point of being used as a tool to complete the project. But when we have all this information upfront, this digital AI model is essentially your contract that says, “Build this,” and that’s the intent here—to write better contracts rather than be stuck with some of the legacy, poorly written contracts that are out there—so that we can have better-finished projects.
Q: Which groups are working on developing these engineering AI agents so far?
A: AI agents are definitely being worked on across the entire world, right? Agentic systems are being released from ChatGPT, some of the other models as well. They aren’t significantly being applied within any software as of this moment, but I know that there are software systems overseas that are just on the cusp of implementing AI agents at the beginning. So, it’s still in its early development—will it take six months to develop, maybe a little bit more, will it be 18 months? I could definitely see it, yeah, where these systems are capable enough to not only manage the middle part of construction and not only have our typical commissioning completion systems at the end but to completely link all three of those aspects of projects together to have an end-to-end AI solution for all aspects of projects. So, definitely stay tuned as we follow some of these groups and see who’s the one that’s going to step out and solve these problems.
Q: Data collection and data logging are critical in the setup of a construction management system. Could AI be used to extract this data from engineering design systems to prepopulate the ITRs with the required technical details? There are already information management systems like Aveva—what are your thoughts?
A: We’re thinking even one step deeper than that—we’re not looking at taking one system and then pre-populating the next system. This system that we’re talking about with this AI agent design system is the system for everything, right? No more paper—this is a digital model, a 3D model with all aspects of the project, and everything is managed and controlled with this AI system. There’s no pre-population—even if we get to, say, our commissioning software systems at the end, right now, in all the systems that I’m aware of, you have to go through the process of wrangling with Excel to get all that information loaded into your commissioning software. So, that doesn’t happen anymore—that’s what we’re talking about in this presentation. This data management system right at the beginning—your 3D modeling software and every aspect about it—is created in a three-week time frame, in a very short period of time, and then that system is used to manage every aspect of the project. So, do you need an ITR? It’s coming out of that. Do you need a work package for construction installation? It’s generated from that system. Do you need to do a design review? It’s generated from that system. And that data management system lives for the life of the project, from start to finish—it is essentially the project. It defines every aspect—it defines the contract, it defines what’s required for construction, and it defines your completions in your staged process for commissioning. Everything is driven out of this system. So, when this system is fully populated and fully designed with every piece of information in a three-week period of time, everything is known upfront, and there’s zero ambiguity. So, some of the questions that you’re asking there aren’t even questions anymore because it’s all fully defined upfront.
Q: What are some of the tasks where AI can be used for completion and commissioning management?
A: So, yeah, it can be used today—create a checklist, review, and document—all those good things. But we’re taking that even a step further so that AI is your team of engineers on the projects, managed by a human. That’s the evolution that we’re talking about—to take this even further, to do all of the knowledge information work on your project up front in a three-week period of time and iterate through that process before making any financial commitment.
Q: What could be the first challenge to have a stable approach?
A: Yes, for sure, there will be challenges, right? This isn’t an overnight success. The accuracy and success rate of these AI systems certainly do need to improve—like you mentioned, if we’re looking at a 50% success rate of AI systems, and that was the metric that was used on the chart that I showed earlier, we need that to improve. So, what is the tipping point? I don’t know—what is the assumed human error rate in design? Probably higher than we think—we’re all human, and we all make mistakes, and that’s natural, right? So, are we making mistakes 5% of the time, 2% of the time, 1% of the time? I don’t know that specific statistic, but once AI crosses that threshold and can reduce the human error rate and produce systems more reliably than humans currently, that’s probably the first challenge, right?—to prove the reliability of the information that they generate from these systems, to have confidence that we do have a stable approach and a stable design here, right? But that iterative process can certainly happen pretty quickly to get to that stable approach. So, I think that’s the trend that we’re watching—and when do those two points cross so that we can leverage some of these AI systems?
Q: Curious if we are able to discretize the model, as it’s time-consuming going from drawing to drawing?
A: Absolutely—that’s the vision that we’re talking about here, right? We would no longer actually even be going drawing to drawing, right? Because there are no drawings—everything is in this system. So, the entire system is modeled in 3D modeling software with full functionality, and it’s more than just a pretty picture. 3D modeling exists now, and you can walk through the system, and you can see all the walls and the roofs and the cable tray and the HVAC ducts, but it’s more than just that pretty picture—it’s a functioning digital representation of the project to allow that digital rehearsal. So, you can actually, in that model, be defining your systems and even going through the staged digital rehearsal of your entire commissioning process. This is actually done—not with these sophisticated levels of systems—in the aerospace industry. They won’t go through the detailed step-by-step processes of commissioning before they even launch their satellites. I have a good presentation on the James Webb Space Telescope and how they did that. So, stepping through that—essentially, the term that I heard is “plan for the best work on the best but plan for a bad day.” So, with these digital systems, then you can do that complete digital rehearsal and see all of the potential failure scenarios or anything that could go wrong in software before you even get to testing in the field. So, systematization will be baked into that process as we’re going through that digital rehearsal, and then when we get to commissioning in the field, it’s just simply executing what we’ve already done tens or many times already—just doing that exact same thing in the field. It’s pretty exciting to be able to adapt or evolve to some of the processes that the aerospace industry has been doing for a while but then also use these digital systems to allow that to happen a little bit easier for any company.
Q: The project has started to apply this in some aspects?
A: Yes, the software that I’m thinking of that is currently working on this to evolve to this level of sophistication is being actively used on some pretty big projects—multi-billion-dollar projects. The next step for them to evolve to that is then to build in this AI agentic aspect here to leverage the information to still reduce that engineering phase down from three years to three weeks. It’s definitely being worked on, and some of the companies that are already using these software systems, I’m sure, will be very keen to incrementally add some of these capabilities to their next projects. Those are the companies and those are the software systems that are going to dominate the industry and revolutionize everything because if you’re not using these software systems and you’re going to insist on three years of design, that’ll be your last project, I guarantee you.
Q: How can AI help prepare the contractual aspects of the commissioning?
A: So, again, the AI model created at the beginning is your contract. I know we’re very much thinking about paper processes here, right?—as a P&ID drawing or a written contract for somebody to sign. There’s no paper in what we’re envisioning here—your AI systems are your contract, and the P&ID models in your system are every aspect of your project. So, everything that would be information-related, such as a contract, is in this data management model. There would be no need to print anything out because everything is defined digitally in this system. To get information in and out of this system, you’re using a digital device, so there really would be no reason to print out a contract—everything would be digitally signed and digitally executed with a 100% full definition of your project before even starting. This model is the definition of your project and is your contract and every aspect of your project—everything comes in and out of this model. Yes, that’s right, Peter—it’s a lot of the problems, a lot of the challenges on projects right now are information, right? And some groups use that to their leverage and use that to their advantage because others don’t necessarily have that information. That’s why this information age that we’ve been in for the last 20-30 years is bridging that divide, right?—making information more easily accessible to everybody. So, when you can have 100% of the information on your project defined up front, it completely closes that gap so that everybody has the same information. Everybody can visually see, visually function—fully go through a digital rehearsal of your project—and everybody has full transparency and access to all the information. No longer can you hide or withhold information to your advantage and use that as a discrepancy or something that wasn’t provided between engineering and construction—everything’s fully defined upfront to remove all those problems, remove all those information gaps, and make the only focus that’s possible on projects to focus on quality and precise execution.
Q: Sorry, Sir, did I make a mistake? I haven’t reached the maturity necessary.
A: Not quite yet, no, but the trend would show that we’re on our way there now. Maybe you’re referring to the lessons learned comments earlier—yeah, that could be as well. Definitely some more work to do there, but when we have these digital systems, then it’s no longer a human process that’s relied upon to apply these lessons learned. The systems just evolve, and with access to all this information, they improve for the next time. The systems are the lessons learned—the systems can track and manage productivity and see what worked and what didn’t, and this is all just automatically built into the next iteration of the project for improved performance.
Q: Will it be possible to catch up with the transcription afterwards?
A: Yes, absolutely. This recording is on our YouTube channel—you can check it out there, you can download it, get a copy, and share it with others. Definitely check it out because I think the more people see and understand this transformation that’s going to fall upon us, the better we can prepare and take advantage of some of these systems.
Q: You’re seriously considering implementing AI for project delivery and commissioning—do you know of companies working on building AI models to disrupt industrial construction and commissioning industries?
A: Yes, there are lots of software systems out there. I can see that several software systems are still focused on kind of the current or legacy aspect of completion—some of the systems have been around for 20 years—but there are some cutting-edge software systems out there that are working to the level that we’re talking about here, where there is no paper, there is no contract aspect, there is no need to load a software system with information or wrangle with Excel to feed into it. There are these digital twin systems, these AI models for projects that form every aspect that we’re talking about here. There are companies that are currently working on this—it’s probably not any of the big companies that you’ve heard of because some of the big companies are slow to move, right? So, there definitely are project systems that are out there that probably aren’t too far away from implementing everything that we’re talking about here.
Q: How could we deploy the AI technology in the commissioning industry?
A: So, not only the commissioning industry—this is an entire project-wide, end-to-end solution for AI, right? We don’t want to operate in silos and just speak of the commissioning industry—this is an upfront overhaul of the engineering design process and really compressing three years of engineering design down into three weeks for improved accuracy. When you have that fully defined upfront, then that is your system—that is your system for design, that is your system for construction, and that is your system for commissioning. When you implement a system from project end-to-end, linking all information, then that’s how you deploy this AI technology in the commissioning industry. So, you can’t quite do that today, but definitely watch because I wouldn’t be surprised if it’s not too far away.
Q: But therein lies 99% of the problems we currently face.
A: Yes, I agree completely, and lots of people like those problems, and that’s why the construction industry doesn’t want to change. But with these types of systems that come along and this transformative change that is essentially going to be forced on the industry, right?—because if you’re going to ignore these systems, then you’re out of business at the end of your current project because the next one’s going to be much different. The stats from the Meter report there—that the time AI takes to complete a task is doubling every seven months, and it looks to be increasing—so let’s say that becomes like four months or five months for the doubling of AI capabilities. By the time a company reports on their quarterly progress, when they do that quarterly report the next time, these systems have doubled in capability. So, when you’re faced with that kind of exponential change from quarter to quarter, then over the life of a project that could be three to five years, the technology has significantly increased. So, the current project you’re working on right now may be the last one that is done with these old legacy systems, and by the time the next project that you work on—say, in five years from now—comes along, it’ll be implementing systems like this. I can see that happening definitely in a shorter time frame than we’re expecting.
Q: It actually bridge the chasm?
A: That’s absolutely right—because when there are gaps or there’s information that’s being withheld, then that gives people leverage, right? When you take away that leverage and put everybody on an equal playing field, then the only option is precise project execution and quality performance because there’s no other way to use that chasm to your advantage, right? This is going to significantly transform the construction industry, and I think I used the example in our Industrial Commissioning Association newsletter this week that there’s going to be people that kick and scream—just like a little child that just wants to eat their Cheerios—they’re going to fight this as hard as they can because they don’t want it to change; they like the way that things are. But the companies that do adopt and implement these systems will thrive and quickly put those others in a tough situation—they’ll either have to adapt to these new ways or find a new line of work.
Q: In the future, the AI model will be finalized and perfect for design and commissioning, so this will decrease/reduce the total number of engineers to handle the project?
A: Well, potentially, but that may actually be a blessing for the industry, right? Because there are fewer and fewer young people coming into the industry—it’s a brutally tough industry, and it’s not very fair, and it’s hard for new people to get this knowledge and information. There’s definitely a workforce shortage in the industry for engineers who want to work on projects, for construction folks who want to work on projects, and for people who are going into the commissioning industry. Our commissioning experts are all retiring, taking that knowledge with them, right? So, there’s definitely a labor shortage out there. The labour pool that exists, I think, is still going to be gainfully employed because they’re the ones that are going to be required to step up as the leaders and manage our AI systems—manage our 10,000 AI agent systems to complete projects. So, it’s going to significantly increase productivity and enable the current workforce to leverage these systems to complete projects. I don’t see a point in time where there’s going to be a need for fewer engineers because there’s already a huge, significant shortage right now. These are going to enable the current workforce to actually complete their work a lot quicker and build projects a lot quicker to gain that efficiency—so that’s some of my thoughts on that one.
Cool—lots of great questions. This is some revolutionary discussion that we’ve had here; this is pretty exciting to see that there’s this kind of engagement and this kind of excitement on these types of systems. I wasn’t entirely sure if there would be some negativity or people resisting this or saying this isn’t a good thing, but I’m very pleased to see that there are lots of positive comments and people that are interested in pursuing these types of systems because these will be the groups that excel and win and deliver successful projects rather than the ones that want to resist this change and make this transformation more difficult. Alright, so I definitely do appreciate everyone’s comments and questions and the fact that you took some time to join me here. I think these are some exciting times that we’re in, and I’m definitely going to be watching some of these trends to see when the technology crosses these limits in success rates and efficiencies so that we can start leveraging them on these projects. Definitely watch because I’m watching some of the systems that are out there to see what’s evolving and what’s developing, and when I find the systems that are the ones that are going to be the standout leaders in the market and for developing some of these, I’ll be sure to let you know so that we can all start making this transition as quick as we can as we’re on this exponential transformation. So, we’ll leave it there for now—thanks for joining, and have a great day.
Recent Comments