How to pay construction crews: hourly vs piece rate

0:00 As always, this content is for educational and informational purposes, only it is not legal, financial, or professional advice. Welcome to Scaling Legends. Glad to be here. We have a really fascinating one today. Yeah, we're jumping right in with zero warm -up today. So, two of construction's most tech -forward contractors just made massive equity investments in an AI hiring startup. Which is huge news. Right. We were talking about DPR Construction's Venture Arm, which is WND Ventures, and Suffolk Technologies. Both of these heavyweights just backed a company called Skillet. And Skillet, for those who don't know, operates what it calls the largest network of vetted for -hire trades workers in the country. Exactly. And I want to emphasize right up front, you know, this is not a pilot program. No, not at all. They didn't

0:46 just buy a few software licenses to see if their project managers, like, enjoy the interface. They took equity. They bought into the company. Yeah, they have a real stake in it now. Right. So our mission for this deep dive is to unpack exactly what this bet means for the industry. Because, I mean, when the giants start buying up the actual plumbing of the labor market, the ground shifts for everybody else. It really does. And the scope of this move becomes a lot clearer when we look at the financial footprint. Tell us about the numbers. Well, according to the latest market intelligence report from Smart Business Automator, Skillet has raised approximately $13 .6 million in total funding. Wow. $13 .6 million. Yeah, it's a serious number. Now, the exact equity amounts put in by WND Venture and

1:32 Suffolk Technologies, those are completely undisclosed. Sure, they usually keep that quiet. Right. But honestly, the actual dollar figure is secondary to the strategic signal here. Large general contractors are actively trying to own the infrastructure of how craft labor is sourced. Which means, you know, if you're dealing with the massive administrative overhead of crew sourcing, like if you were spending your weekends trying to figure out if the drywallers are actually going to show up on Tuesday, this deep dive is for you. Oh, absolutely. It maps out exactly where the hiring process is heading. But before we dissect the whole venture capital strategy, I feel like we have to clarify the technology itself, right? Yeah, we definitely should. Because when we say AI and labor in construction, a lot of people immediately picture like a robotic

2:17 dog carrying cinder blocks on a site. Right, or some kind of fully autonomous dozer grading a site by itself. Exactly. But that is not what this is. No, not at all. The technology we are looking at operates entirely in the back office. It sits strictly in the realms of matching, sourcing, and verification. Okay, so break that down for us. What does that actually look like operationally? separate Skillet or any AI -vetted network from a standard job board is automated curation. Curation. Right. It's doing the heavy lifting of verifying skills, safety certifications, OSHA cards, work history, and it's doing all of that while drastically optimizing the matching speed and the quality of that match. Let me stop you right there, though, because parsing a PDF of an OSHA card or, you

3:04 know, checking an expiration date that isn't really revolutionary, is it? Well, it depends on how you look at it. I mean, we've had optical character recognition for over a decade. Is this AI just a... slicker, slightly faster database? Like, does it just reduce the workforce coordination headaches we always complain about? Or is there a genuine mechanism here that changes the game? It's definitely the latter. And the distinction lies in how the system handles messy data. Messy data. Yeah. The massive volume of unstructured, chaotic data that basically defines construction labor. A standard database requires someone to manually input neat, categorized information. Which never happens in the real world. Exactly. In the real world, a trades worker might upload a blurry photo of a safety certification, right?

3:52 Or a handwritten dispatch slip from a union hall. Or a resume that's literally just a text message listing their past five projects. I've seen that so many times. We all have. And that's where the AI comes in. The engine uses natural algorithmic processing to read and interpret that total mess. Oh, okay. It extracts the relevant data. cross -references it with national safety databases to ensure it isn't forged, checks the expiration dates, and automatically builds this standardized, highly structured competency matrix for that specific individual. So it completely removes the friction for the worker. Yeah. I mean, we aren't expecting a 30 -year veteran pipe fitter to sit down at a laptop and optimize a digital profile for an hour. No, you'd lose them immediately. Right. They can just submit whatever fragmented proof they have, and the AI translates it into

4:39 a language the hiring man can actually use. Exactly. And then the AI takes it a step further through predictive availability. Wait, what does that mean? Predictive availability. So instead of just waiting for an electrician to log on and say, hey, I need a job, the system analyzes historical project data. It looks at regional weather patterns, the typical life cycle of specific builds. So it's anticipating when people finish up. Exactly. It can predict when a crew of specialized workers is likely wrapping up a phase on a current project and will soon be available. That is wild. It is proactively matching the supply of labor to your demand before that labor even hits the open market. That completely flips the traditional sourcing model. I mean, if you use a traditional

5:24 general job board, it feels a bit like shouting into a crowded room and just hoping a qualified available electrician shouts back. It's a perfect analogy. You know, you get 20 people who say they're electricians, but when you actually dig into it, maybe three of them have the right commercial experience. And two of those have expired certifications. Right. And the last one is already booked for the next six months. So you have HR teams and field superintendents playing phone tag, manually calling previous foreman. Which is a massive time sink. It is. That administrative drag is precisely the kind of bottleneck smart business automator tracks when they analyze crew sourcing inefficiencies. Yeah, the time spent on manual verification is just staggering. But when the AI assumes that burden, it reduces the friction

6:12 of hiring down to near zero. So it's learning over time, too. Absolutely. It learns which combinations of skills, past projects, and regional experience lead to the most successful placements on specific types of builds. It is actively connecting the dots faster than a human team ever could. But look, if this engine is successfully pulling off that level of predictive matching, I have to ask, how much do these contractors actually trust it on day one? That's the million dollar question. Because letting an algorithm staff a multi -million dollar job site seems incredibly risky. It does. And I think that tension brings us to the deployment strategies. Because we have two massive... highly sophisticated contractors investing in this exact same platform, but they

6:58 are injecting it into their operations in completely different ways. Right. And the contrast here reveals exactly how large organizations segment technology adoption to manage their unique risk profiles. Okay, so let's break them down. Let's start with strategy one, which is DPR construction. They are deploying the platform nationally right out of the gate, and they are specifically targeting their mission -critical work. Wait, really? They are bypassing a soft launch and going straight into national high -stakes projects. Yep, straight into the deep end. Because mission critical usually means data centers, advanced manufacturing facilities, life science labs. The technical requirements on those bills are unforgiving. Completely unforgiving. I mean, if a specialized copper puller makes a mistake on a data center, the cost of that delay is astronomical. Why would DPR trust a

7:45 relatively new AI platform with their absolute highest liability projects nationwide? It sounds counterintuitive, but the severity of the pain point dictates the aggression of the solution. Okay, explain that. On a massive data center build, the single biggest variable that can derail the schedule is a lack of highly specialized vetted tradespeople. True. The risk of a delayed project far outweighs the risk of trusting a new technology. By deploying it where the demand for verified elite talent is the absolute highest, DPR is testing the platform's ability to perform under maximum pressure. They need volume and they need it yesterday. So they are basically throwing the engine straight into a Formula One race because they need the speed to win. That's a great way to put it. But then we look at strategy two. Suffolk is taking a totally different path

8:31 here. Very different. They are keeping this technology heavily contained. They are deploying it purely through Liberty first. And Liberty is Suffolk's self -performed subsidiary, right? Correct. They are not rolling it out to their general contracting business where they manage external subcontractors. They are only using it for the division where they directly employ the tradespeople pouring the concrete or doing the carpentry. Interesting. Why isolate it like that? Suffolk's approach is deeply focused on machine learning and algorithm refinement. By incubating the technology inside a self -perform arm, they control every single variable in the environment. Wow, I see. AI requires a massive amount of ground truth data to actually learn and get better. If Suffolk rolled this out to their subcontractors, they would get the

9:17 initial match data, but they would lose visibility into what happens next. Right. The feedback loop would be broken. They wouldn't know if the carpenter the AI matched actually showed up on time, performed well. Or quit after two weeks. Precisely. Because Liberty is the direct employer, Suffolk can monitor the entire life cycle of that worker. So the AI makes the match. Liberty does the onboarding. The superintendent evaluates the field performance in -house. And HR tracks the retention. Yeah. Exactly. All of that downstream data is fed back into the algorithm. So it's constantly correcting itself based on real outcomes. Yeah. If the AI recommends 10 concrete finishers and two of them struggle with a specific pouring technique, the system analyzes what those two individuals had

10:03 in common in their work history and adjusts its future recommendations. Wow. Suffolk is basically using their self -performed division as a closed -loop laboratory to perfect the platform's accuracy. Which makes the next phase of Suffolk's plan even more significant. Yeah. Because they're not planning to stop at Kraft Labor, are they? No, they're not. They've explicitly stated they want to expand this AI matching upward into white -collar roles as the platform grows. We are talking about project managers, estimators, superintendents. Yep. Because the architectural foundation of the AI is agnostic to the role. Meaning what exactly? Meaning if the system can ingest messy data to verify a welder's certifications, match them to a project's risk profile, and predict their availability. That exact same underlying

10:50 mechanism can evaluate an estimator. So it could, like, ingest an estimator's track record of bid accuracy. Exactly. It could cross -reference that with the historical cost overruns of similar regional projects and match them to the project type where they have the highest statistical probability of success. That is just, you know, it's a total paradigm shift in how human capital is evaluated. It really is. We have one giant testing the breadth of the platform by pushing it nationally on complex builds, and another testing the depth by preparing to push it into salaried management roles. Both are seeing immense value. But this forces a strategic question, doesn't it? If you love the software, just buy a commercial license, pay your annual recurring revenue, log in, and staff your project. Right. Why take the extra step? Exactly. Why take the

11:38 highly complicated step? of involving your venture capital arm, W &E Ventures and Suffolk Technologies, to actually buy equity in the startup. Because the distinction between being a commercial customer and being an equity owner is about controlling the future ecosystem. Okay, unpack that for me. When a contractor signs a commercial agreement, they are renting a tool. They get the immediate efficiency gains, sure, but they are subject to the roadmap dictated by the software company. They are just a passenger. By taking an equity stake, these venture capital arms are attempting to shape and own the infrastructure of the future labor market. So it's the difference between a shipping company buying a larger fleet of trucks to beat the competition versus buying the actual highway system everyone has to drive on. That

12:23 highway analogy perfectly illustrates the defensive and offensive benefits of the investment. How so? Offensively, as Skillet grows and potentially becomes the industry standard for vetting trades workers, DPR and Suffolk benefit financially from the appreciation of their equity. Right. They make money in the startup succeeding. But defensively, they secure early access to beta features. They get a seat at the table to steer the product development, ensuring the AI is trained to solve their specific edge cases and integrate flawlessly into their own proprietary workflows. So they are building a technological moat. that a competitor relying on manual processes simply cannot cross. Exactly. A massive moat. And the speed at which these tech -enabled giants will be able to staff up a massive project will make

13:11 traditional methods look like they are moving in slow motion. Without a doubt. If you are listening to this and you want to understand where your own operational bottlenecks are before you have to compete against these AI pipelines, you need visibility into your data. You have to know your baseline. Right. Which is why we continually point to tools like Smart Business Automator, because they specialize in identifying those exact administrative drains in workforce coordination. Visibility is the first step toward remaining competitive. Because look, the heavyweights are placing their bets, but the vast majority of the construction industry operates in the mid -market. Yeah, if you are a $50 million or $100 million contractor, you do not have a venture capital arm. You cannot buy equity in the next big AI startup to build a moat. So the critical question is how the mid -market responds to this

13:57 shift. Right. When does a vetted network actually transition from being an expensive toy to an absolute operational necessity for a mid -sized general contractor? Because the math changes entirely when you don't have a bottomless innovation budget. It does. But the inflection point for the mid -market contractor is purely mathematical. It happens the moment your speed to hire and the hidden costs of skill verification outpace the platform's fees. To visualize the mechanism here, we have to look at the anatomy of a bad hire on a critical path. Walk us through that scenario. What does that math look like in the real world? Okay, imagine you win a mid -sized commercial project. Let's say a medical office building. Okay, medical office. The schedule is incredibly tight, and you realize you need

14:43 10 highly skilled carpenters on site by next Wednesday. Which is a scramble. Total scramble. Using traditional methods, your human resources team, or worse, your highly paid field superintendents are going to spend... roughly 40 to 60 hours trying to source those workers. Because they're writing job postings, sifting through resumes, calling references who never answer, and verifying safety certifications. Exactly. So you immediately have the hard cost of those internal salaries being burned on administrative tasks. Plus the opportunity cost. I mean, what those supers should be doing is managing the site logistics, focusing on safety, driving the schedule. Right. But that is just the baseline cost. The real financial damage occurs when the manual verification process

15:30 failed. Oh, boy. So what happens? Let's say you finally get 10 carpenters on site, but two of them exaggerated their experience. They don't actually know how to read the complex blueprints for this specific medical build. Oh, that's a nightmare. So they install the framing incorrectly. You now have the cost of the wasted materials. You have the cost of the labor hours spent doing it wrong. You have the cost of the labor hours required to tear it down and rework it. And most importantly... Exactly. For a mid -market contractor operating on tight margins, a single bad hire on a critical phase can entirely erase the profit margin for that portion of the job. It just cascades. Yeah. And then you are back to square one, spending another 20 hours trying to replace those two carpenters. Compare that cascading failure to

16:16 the deployment of an AI -vetted network. Right. How does the AI handle that same medical office building? Well, yes, you are going to pay a commercial licensing fee or a per placement fee. to use the platform. But if that predictive matching engine can deliver 10 verified, perfectly matched carpenters in 48 hours with near zero internal administrative time required from your superintendents. The math heavily favors the technology. Heavily. You are weighing a known software fee against the thousands of dollars you would have inevitably lost in delays, rework, and administrative overhead. But here's the catch. You have to know your own baseline to make that calculation. You cannot determine if a platform fee is worth it. If you have absolutely no idea how much money you are currently bleeding through slow hiring and rework.

17:04 That's exactly right. Which circles right back to tracking that administrative drain. Yeah. If you want to get a handle on those hidden costs, you really need to get involved with the platforms actively solving this. It's the only way to see the true cost. Yeah. And you can sign up to be a beta tester right now at smartbusinessautomator .com. That is smartbusinessautomator .com. Getting first access to those new features gives you the diagnostic tools you need to understand your own speed to hire metrics. And you have to establish that baseline before you start competing for labor against the AI -enabled giants. Because they already know their numbers. They absolutely do. So the ultimate takeaway for the mid -market is that... You don't need to invent the algorithms and you don't need to buy the startup. But you must adopt the technology as a commercial user the very second it proves it can protect your

17:51 margins better than your manual processes. The adoption curve in construction technology is just steepening so rapidly right now. It is. The gap between the tech -enabled contractor and the analog contractor is no longer just a conversation about efficiency. In a market where skilled labor is the most constrained resource, it is a conversation about basic survival. Survival is the exact right word for it. Wow. We have covered a massive amount of ground in this deep dive. I really have. I mean, we broke down how the AI actually ingests unstructured, messy data to build standardized profiles, how predictive matching anticipates labor availability, and why massive players like DPR and Suffolk use radically different deployment strategies to test the tech. Plus the strategic advantage of

18:37 buying equity in the labor pipeline. Right. And we ran the real -world math on why mid -market contractors have to adopt these tools to protect their margins from the catastrophic costs of a bad hire. It's a lot to process, but it's essential. It is. Before we wrap up, we do have a request for you. Send this deep dive to any contractor you know who is spending significant time on craft labor sourcing. Or... anyone who is currently struggling to find reliable tradespeople quickly. We all know someone whose weekends are completely ruined by workforce coordination. Send this directly to them. And take a second to follow Scaling Legends on whatever app you are using right now. It is the best way to ensure you don't miss our next analysis of where this industry is heading. And look, looking at the broader implications

19:24 of this technology, there is a lingering dynamic we haven't fully resolved. And it is something you should deeply consider as you look at your own hiring practices moving forward. What's that? Well, we just analyzed how the largest, most well -funded contractors in the world are actively buying equity stakes in the algorithms that identify, verify, and curate the absolute best, most reliable trades workers. Right. They are intentionally building a high -speed, frictionless pipeline straight to the top talent. Exactly. They are essentially setting up a massive filter to catch all the premium talent before it ever hits the open market. That's a scary thought for everyone else. Which leads to a critical realization. As these AI platforms increasingly siphon off the top 20 or 30 percent of the skilled workforce,

20:10 exclusively for the contractors who own and utilize the technology, what happens to the quality of the labor pool left over for everyone else still relying on traditional job boards?