Josh Whisenant: Yeah.
Luigi La Corte: Hello everyone, welcome to Construction Technology Podcast. I'm Luigi with Provision. We got Andrew with Join, Mike with Build Vision, and we got a special guest, Josh from Roger. What's up, gentlemen? How are you? Happy Friday. Josh, how are you?
Mike Powers: Happy Friday.
Josh Whisenant: Little sleep deprived, but that's always how it goes. How are you guys?
Luigi La Corte: Yeah. Yeah, good. What what time did you get in last night from Orlando? You flew last night, obviously.
Josh Whisenant: Yeah, I did. yeah, I I have no fun at the groundbreak brat bash, you know. I'm what's when did get home? I think I got home at like eleven thirty mountain time, so that would have been one thirty AM Eastern. So f you know, it's not so bad, but you know, I need my beauty sleep. Come on.
Luigi La Corte: Yeah, of course. That's what that's it's working for you,
Mike Powers: So
Luigi La Corte: dare I say. it was really fun. I mean, n not to not to rub it in, but they they rented out Islands of Adventure at Universal Studios. And it was amazing just to see like so many fifty year old men kind of giddily scream to be at the front of a roller coaster ride after ride and then get up there like we're we're going again.
Mike Powers: So is this the last hurrah before they crush every AI tool? Is that is that the pitch? They're like, Hey, we have this new model
Luigi La Corte: Yeah, interesting.
Mike Powers: and come have fun, you know, but like it's like the giving the dog a a hamburger before you take it to the vet. Is that is that we're at or not yet?
Luigi La Corte: Yeah, interesting. Maybe. Well, they did launch a model. Maybe we can do a bit of like a groundbreak retro.
Josh Whisenant: Yeah, sure.
Luigi La Corte: because Josh was there with us too, but they did release a model, Magpie, Magpie three point I wasn't aware that there were previous versions of Magpie, but
Josh Whisenant: Or was I?
Andrew Zukoski: start I there there might not have been.
Luigi La Corte: so Magpie Yeah that's right. That's true. Like if you just start if you just release a model with like, yeah, this is provision intelligence eight point guys, where have you been? You're like, well, fuck,
Josh Whisenant: Right.
Luigi La Corte: maybe maybe we've missed it. But anyway, MagPoint three point So Magpie, they it's frontier according to their benchmark. The problem with these benchmarks is like if it's not a third party benchmark, it's just so manipulatable. Like on this one very discrete task, you know, are we better? It doesn't mean you're frontier unless it's like this very small task. But anyway, so magpie three point it's a reasoning model to choose between their taxonomy of document types, what actions to take. I mean, there were so many words thrown around like model, ontology. They released Procore ontology, but you know, didn't really get into it. but for like I guess the like what a customer would care about is Magpie helps get tasks within Procore done better than GPT six Astra with API access. So it's better and that's good. I like you know, I don't know, that's it. We didn't really see much.
Mike Powers: They throttle what do you do you get to how does it work? Do you get tokens of the customer? And you're only gonna use 10 tokens a month. How does it work?
Luigi La Corte: I don't think they actually sh I don't know if they give you a token allowance. They probably do. But we could talk like very quickly. They do.
Josh Whisenant: No, they I think they do. I mean, I've heard I heard two things where I heard people saying, yeah, they're trying to get me to pay forty thousand dollars a year for Procore AI, but then they had this control tower thing where it was tracking all your token spend. So it seems like there's some combination of fixed end variable costs.
Luigi La Corte: I see. Is that just within Procore or is that like like a ramp type initiative where it's like all token costs?
Josh Whisenant: No, no, no. I think it was just within Procore.
Luigi La Corte: I see. Yeah. I I wonder why would they why they would I mean, yeah, like they have to limit their cost too, but I thought it was interesting how they're packaging it, Josh. Did you see how they were packaging Procore AI? So you can't buy an individual agent. You buy packs of agents. And I went to the innovation hub and yes, that's it. Digital coworker starter packs. You're buying starter packs. the starter pack is like
Josh Whisenant: Yeah yeah yeah.
Luigi La Corte: five agents, I believe. So deep search agent. You have a submittal agent
Mike Powers: What does that mean?
Luigi La Corte: deep search.
Mike Powers: No, what is a what I think what does an agent mean? Like why is there
Luigi La Corte: It's
Mike Powers: not just one agent?
Josh Whisenant: Search S
Mike Powers: It's Procore.
Josh Whisenant: Search doesn't feel like a very search feels like not an agent, right?
Andrew Zukoski: 'cause they they announced they announced that they announced that two two groundbreak keynotes ago with Procor.
Mike Powers: Here's agents. So how how are these not I mean, how are these not just skills?
Luigi La Corte: They they really are. I mean, a lot of like the the big model companies have really moved away from individual agents. It's abstracted away, right? Like if you prompt codecs, you will see the agents that are invoked, but they're created on the fly, right? Like an adversarial code agent or like a legibility agent. Honestly, the customer doesn't want to pick the agent. Like it's why?
Andrew Zukoski: Wait, I don't know if that's true. have you spent any time we've been rolling out our MCP server and so have been spending more time inside the Microsoft ecosystem. and there and you could argue this is necessarily behind the c like behind the times because I just said Microsoft, but there you package like whatever the underlying MCP is, like a skill into an agent, but then you whitelist tools and endpoints. At the agent level. Like packaging it reduces the scope in a way that lets you then like expand the permissions that people can access or edit it. Like you might want a lot of people to be able to review an RFI, but you might not want everyone to be able to submit a pay app to the owner or something like that. to execute
Luigi La Corte: Mm.
Andrew Zukoski: whatever the underlying capabilities are, even though they all rest on like the same model and the same skills and the same tool set. So I don't know if it's that
Luigi La Corte: Mm. Yeah.
Andrew Zukoski: like
Luigi La Corte: I mean that that makes sense from a permissions perspective. Like you cannot let's say like Josh, like you can't have access to you know pay apps because like we don't wanna we wanna gate that.
Andrew Zukoski: We we don't trust you.
Luigi La Corte: Yeah, we don't trust you. But I I I I wonder I'm I'm kind of bearish on the on the UX of a user having to invoke an agent to get a task done. I think
Andrew Zukoski: Yeah.
Luigi La Corte: knowing what the agent can do. And then having to choose it, I think is gonna be too much friction long term. It should just be present, it should do the work when it makes sense, or it should be invoked when the context justifies it, which is effectively just like how agents invoke skill. So that's why, like, yeah, like the agent skill thing to me is just like maybe it's a little bit pedantic. But Josh, what what did you think about the I don't know if you caught any of the keynotes, but what was your impression of the event at like at large about Procore and their AI?
Josh Whisenant: I mean yeah. I mean Propour is always such a it's just such a pleasure to go to that conference. because you really just see your people and you see all of these all of these contractors that you work with and it's really much more than like any other conference. I don't know. I I don't understand quite how they have all the all the people that we like to work with like just like walking in torrents down that hall, but like you see them the whole day and it's just super fun. I mean I have a lot of thoughts on pro core AI, not all of them are necessarily They're not all in a perfect flow, right? you know, as far as Procor's benchmark, I'm going to present a certain exhibit for consideration by the jury, which is the iconic you know, benchmark modeling
Andrew Zukoski: Ha ha ha.
Josh Whisenant: system. and you you can see scales on this where, you know, over here it's like sixty nine point one and sixty-nine point three or something, you know. and so it's kind of like you always see the benchmark and it's like, okay, here's the benchmark, show me the money, right? now I don't think that that is I think that saying that ProPore sucks at AI is not really a very useful thing because they have, you know, billions and billions of dollars of capital, I assume, and they've got a ton of money and they can throw on a huge conference. And so at some point they should figure out how to be competent. I think that there is an inherent incentive structure issue with ProPore right now where they have increased their costs for these contractors dramatically in the last five years. Right. sometimes as much as 50%, largely to show public market show growth to public market investors. And like everybody that walked up to my booth was like so freaking pissed at how expensive Procore is and how they feel like they're getting jerked around by them all the time. And to me, this package of all of this stuff you can go spend money on in software inside of Procore, it just kind of feels like it's serving Procor's need to increase average contract values rather than actually saying like, I guarantee you we're gonna make construction run better. Now that doesn't mean that their technology is bad. That doesn't mean that people won't get value out of these tools. But to me, like I'm picturing a project manager using this stuff in
Mike Powers: Yeah.
Josh Whisenant: in a trailer and like, is the project like a hundred times better? Maybe you know, I I I don't I I don't quite understand how this is going to change the world and their their strategy to me hasn't quite made sense. but you know, again, for abundance of clarity, I you know, I think that Procor has built a lot of software. They've built a huge business, a much bigger business than anything that certainly than anything I've ever built by like many orders of magnitude, right? And so, you know, the last thing that you want to do is be a hater because that'll you'll put your foot in your mouth really fast. last thing I'll say. Traversing through job site context to have a c to be ready to have a conversation with people is a huge pain in the ass for project managers. I remember doing that. Right? Like, I see this is here in this drawing and like, what's the spec? And what's the you know, nobody likes doing that. you know, and this is why I mean, this is kinda like what Trump Tools has tried to basically solve, is just like retrieval and this is what it seems like Procore is leaning into with bagpie tasks. Yeah, like they should absolutely solve that problem because it's like a real issue that everybody hates.
Andrew Zukoski: Aren't a lot of people chewing on that problem? I mean I guess Procor would be well well situated to go solve it.
Josh Whisenant: I mean yeah, i i okay, yeah, totally. And I mean you know our technology actually has been really good at that lately, which is a almost another story. it's kind of but i i th what what I'm illustrat what I'm the reason I said that is because I do think that the worst thing that you can do as a contractor right now is not is do anything that pr keeps you from benefiting from the unbelievable commoditization of these models and intelligence that is going on right now. And to me, you should grab as tightly as you can to the wave of the models getting better every two weeks and getting cheaper every two weeks and hold on to it. And the idea that I'm gonna pin my entire business on Procore frickin' magpie, that even if it is as good as they say it is, you know, you know who I don't want to compete with? Sam Altman and and Dario Amade. Like, you know, it's like I I i i it it actually felt like a real strategic misstep to me to try and compete with the foundation models and say that Procor was better than them.
Luigi La Corte: there's some I I think there's this there's something behind it. It's I think it's more commercial than it is functional. So AJ
Josh Whisenant: A hundred percent. Exactly, right? That's what I just said.
Luigi La Corte: Yeah I look I I agree with you like but I think the the reason behind it is okay Procore's price to earnings right now is is less than 10. AI stocks let's say you know broadly speaking trade at 25x PE so they are
Andrew Zukoski: But wait, which one? None of the foundation models have earnings.
Luigi La Corte: Well, I'm talking about Meta, Google, NVIDIA. I'm not talking about open AI and Anthropic.
Andrew Zukoski: Okay, fair. Yep.
Luigi La Corte: Obviously they're not publicly traded, right? But like roughly
Mike Powers: Mm.
Luigi La Corte: speaking, the the there's a premium to AI companies that's being assigned, right? and I think that's probably the bet 'cause like there was a lot of conversation in the keynote about becoming an AI company. And it's like so much so that he like Like I get but I think I think it's functional. Like that's why they said, Okay, we want to be an AI company, we're gonna hire technical talent. We are now assigning Nicholas Plinkington from Drone Deploy to be the principal architect of Frontier Models. They have an applied AI team that's building models. They want to benefit from the increased PE ratio. Like they want they want that that that premium. And then, you know, what what are they gonna do with it? Well, I don't know, maybe they maybe they're gonna be viewed as defensible. But I think my point, I was talking to someone from ProPro I was like, You guys just need to grow faster. Like the only reason you're really getting that premium is not because people view you as making models. Those models have to turn into growth. And like will it really matter to the customer? I'm not sure.
Mike Powers: I they need double revenue. Like they needed like they're gonna double revenue or P firms gonna buy them. Like that it like
Luigi La Corte: What do you think? Say more.
Mike Powers: there's their their growth, their the stock is getting crushed. Growth is not where it needs to be. They cannot squeeze another billion dollars of revenue out of general contractors. So they need to make a bet. This is their bet. Their bet is, hey
Luigi La Corte: So existing customer existing customer value extraction or at value add and taking
Mike Powers: I think if can do tasks on behalf of customers, they can get closer to outcome-based pricing. similar to what Intercom did with their switch to fin, then you get to outcome based pricing. And this rolls
Andrew Zukoski: Intercom.
Mike Powers: into very similar what Josh is doing. Josh and Roger is now right, we're we're selling outcomes or trying to sell outcomes. And if they're in the place
Josh Whisenant: Trying to.
Mike Powers: where they can have a platform tool that is a minimal cost. And they can sell outcomes on top of that and take on risk because they have specific proprietary data. There's a path there, there's a story there.
Josh Whisenant: That's what I I mean, obviously that is what I would do if I was Procor. That's kinda what I'm doing when I'm here.
Andrew Zukoski: It's interesting that that's what like without sorry we should transition to Roger because I'm interested in the new offering. The in a way that's what Procore Risk Advisors was a swing at. And like, you know, the past is not the future.
Mike Powers: They've hired their own team. They hired their own team.
Josh Whisenant: My understanding on that was that they were trying to not really take accountability on any outcome. They were just trying to like do billable hours, right? Am
Andrew Zukoski: Well
Josh Whisenant: I wrong about that?
Andrew Zukoski: on the risk advisors, I I thought it was like
Mike Powers: There's a risk advice. Yeah.
Josh Whisenant: Yeah.
Andrew Zukoski: I thought the play was and you guys it sounds like you know more about it than I do, but I thought the idea
Josh Whisenant: I don't.
Andrew Zukoski: was like, okay, we have all this data, we understand which companies perform better, if you use our system, we'll be able to underwrite or encourage people to underwrite at like a lower rate. Like
Mike Powers: It they wanted to build they wanted to build Shepard, they wanted to build Shepard.
Josh Whisenant: yeah, totally.
Andrew Zukoski: Yeah.
Mike Powers: and then they didn't hire the right team. And now Shepard's owning that market. This is going back to there's outside of general, like there's like 10 revenue streams for a general contractor. The so the the wedge they're in now is an IT budget. They need to get in profit share budgets. And we go back to where what Turner's
Andrew Zukoski: Pay. Procore pay could do that.
Mike Powers: what Turner's doing at scale, and how do you tap into each one of those levels? So Risk Advisors was SDI C SIP. they're starting to do equipment management tracking, which is what Turner does with with first equipment, where you actually rent lift scissor lifts. so they need to go into the additional revenue streams of these customers, which is harder.
Andrew Zukoski: So it's interesting in this way all these agents that are automating things that a person in a trailer or an office would do inside of Procore is trying to go deeper inside the same category of project management and paperwork. It's interesting.
Mike Powers: Correct. But a selling outcome. Hey, I don't have enough people on my job site to hire a pro core agent to do cemental reviews, whatever it is.
Andrew Zukoski: Well, I'm not I'm not gonna not have a person on my job site and I'm already grumpy about how much I'm paying Procor. It's gonna that's gonna be interesting. Hm. But Roger's selling outcomes
Luigi La Corte: So let's go let's go into let's go into Roger. Yeah, so Josh, so tell us about the launch. and bring bring us high level through what was launched, what's the what's different from what was before. And then I mean the website looks amazing. It'd be great if you like even if you
Josh Whisenant: Thank you.
Luigi La Corte: if if you can, like walking us through the messaging and like what people would expect would be great.
Josh Whisenant: I think that the launch We needed to do a few things with our launch. the first thing was generally have a website that wasn't built
Andrew Zukoski: Ha ha.
Josh Whisenant: with me vibe coding on a over a weekend in Replit. you know, we're like a we're like a good company and we should like not have a stupid website even if we're kinda lazy. I think the second thing This is more of a commentary on, I think, the nature of services than any particular problem. But we were often perceived as a pre-construction tool. which by the way, we never intended to get into pre-construction. That was not the intention when we started doing this, but we kind of got pushed there and we kind of let the customers pushing us there. Tri you know, we tried to figure it out there. and that didn't really work very well. That wasn't we didn't have the right fit. No, it worked sometimes, but it wasn't great. And So it was important to clarify that we are not a Team that is just doing a a widget where we just like process something, but we actually carry you to an outcome that's in operations. You know, the way that we bill was an interesting thing to add where we talk about how we're, you know, staking our fiana. We're not billing if w we don't give you outcomes. That's actually kind of You look thoughtful, Andrew.
Andrew Zukoski: What is that I would love to hear maybe you could pull up contract language, but I would love to hear like what the what the outcome is. Like I it's been a minute since I looked at the website, but like there's there's some part that's like okay, there's under there's specific tasks that are happening, and I can sort of understand what the finish line is. But then there's like, hey, if you're unhappy we'll refund you which is like if the outcome is you're happy with us we sell outcomes too, and I think you go further than that. Like
Josh Whisenant: Yeah, I mean the th th and this is a h this is a hard thing about like I mean now we're getting into the idiosyncrasies of outcomes of like how to bill for these things. It's very hard to like It's very hard to quantify quality, right? But it's kind of like, you know, you know it when you see it. And so we basically say, hey, if you don't feel like this is better than anything you've ever seen before, and it doesn't actually match what you want to do, then we just won't send you an invoice. So that's kind of the approach to quality. And then if there's anything that comes up later on down the road that was egregious that we all missed, then we'll be on the hook for that. if it's like, you know, if we just clearly missed something that was completely right there and obvious.
Andrew Zukoski: Do do you take contractual risk? Like if there's a scope gap when it's bought out and they've signed a G P based on the buyout that you did and there was a scope gap in between the G P and the subcontracts that you bought out, do you take on that risk?
Josh Whisenant: Up to the I mean, depending on the scope gap, yes, but only up to the total amount of our fee.
Andrew Zukoski: Okay.
Josh Whisenant: Which I would argue is
Andrew Zukoski: Okay.
Josh Whisenant: much more which is much more accountable than a project manager. Right? Like nobody's losing their salary when they have a scope gap. They might lose their bonus, I guess.
Andrew Zukoski: Yeah, I mean they
Luigi La Corte: Yeah, it'll be fine.
Andrew Zukoski: get fired. Or not. I mean I don't know. People there peop mistakes happen all the time, right? Like that that's
Josh Whisenant: Right, right, right.
Andrew Zukoski: part of it, but you like Okay, so you put the you put the fee at risk for scope gaps.
Josh Whisenant: I mean, what I can tell you is that I c I keep on going back to people over the last year of over the projects we've done the last couple of years and I keep saying, Hey, what have we missed? What have we screwed up? And nobody can tell me anything. I and and so it it it makes me feel like, you know, at some point we're gonna miss something and there's gonna I'm gonna have to have an unpleasant conversation. but that just We're just not even close to that. I it just doesn't seem like that's what's going on.
Luigi La Corte: Josh, do you typically I I haven't seen the scopes, but are you only providing the project specific scope? Are you also modifying their template for some kind of like broad language, like including but not limited to and as per plans and specs, but enumerated below? Like obviously that language was
Josh Whisenant: We do all of that stuff.
Luigi La Corte: you're doing all that stuff too.
Josh Whisenant: Yeah. And that's been that's been a big technology piece of work to figure that all out. Because it kinda depends on what everybody w you know, it has to be super flexible 'cause every every Bill Bob, Jim and Joe wants it a little different.
Luigi La Corte: Yeah. Yeah. Yeah. The r the reality the reality is is like irrespective of what's written there, it's like they need they need to understand it. Because it's not about the coverage. Cause like I think you
Josh Whisenant: That's right.
Luigi La Corte: had a you had a LinkedIn post. It was it was I I found it funny for this the scoping community. But hey, I found a new way to scope, just say as per plans and specs. We're done. And you like you can do that, your coverage, you're contractually right. but like the fact that you're gonna have a fight means now that you have to dispute it and have the argument.
Josh Whisenant: It's
Luigi La Corte: That doesn't go away.
Josh Whisenant: Yeah, that no, that's right. and actually
Luigi La Corte: You just gotta force them to read it. It'd be like, Look at this, read those thirty points.
Josh Whisenant: W I mean what we do, well, you know, I think it's very likely that we will get into I think we're gonna meet with the trades at some point here and talk to them about talk to the trades about our scopes along with their project manager in the room. and you know, so walking through the 30 points. But I think that a lot of, you know, and Luigi, this is me you know tipping my hand a little bit, but I feel like a lot of the value that we provide is in the deep conversations that we have prepared for our customers where we just go through all of these questions and discussion points to figure out the entire job. And I I feel like that's kind of, you know, where all of the nuance comes out is these debates that we have. And it's not, you know, we c it's not a half an hour debate. It's i i they're pretty serious conversations that we have for a long time.
Luigi La Corte: So what what do you think about let's say like the 'cause I think about that and I think that's great. Like that's that's a level of service like I couldn't provide. but what do you think about PMs and their like shouldn't they be able to have those conversations themselves and and do that work?
Josh Whisenant: Hundred percent. A hundred percent. and this I mean this brings you to a question of What is the problem that we really are solving and what is the future of the construction manager? Because Empirically, when you look at task completion outcomes, you're actually competing with employees. That's really the competition, right? Like if we're do if we're gonna go and buy this job out together, then I'm telling to the executive of the company, to the you know, VP ops or whatever, hey, I'm gonna go run this buyout process better than your PM can. I mean, that's really the the argument, and I don't, you know, I that that is the truth. I think if I think that's a bad so that like that's a really bad business model in pre-con in estimating. for a variety of reasons. It's really good for project managers for a variety of reasons. First, at the start of these jobs, there's an unbelievable amount of work to do all across the board, and there's a huge opportunity to like n fix things and not break things. Second, you have project managers that are like coming off of previous jobs, right? Third, a project manager does buyout. Once every three years, right? Often. and we do it, you know, three or four times a week. and so actually what you and and f and so and fourth like these project managers They understand the risk that they take on in these processes and so they're competent and present and paying attention to things. But it's not like I don't like anybody like loves writing contracts, whether with Claude or by themselves in a room or using a template or, you know, anything. I don't think that these this is a task that people are like, this is what I live to do. and so there's this moment here where in these particular tasks on jobs, labor shifting is actually acceptable to everybody in the ecosystem. Actually, more than acceptable on certain projects, it's very welcome. and so then the pitch that you have to make and the narrative that you have to articulate is if you're trying to clearly expand your business and you know we're moving into submittals and it, you know, clearly you the the The implication of that is that we do many things over time. What is the future of these people at these contractors? Right? are you just trying to are you the guy that's like gonna take all the jobs? I think we are not the people taking all the jobs, we are the people that are gonna change all the jobs. I think we can do that. Where as far as I can tell, there's like five times as much work that needs to be done on these jobs and there is people that actually do it. I lived this job, I did a terrible job because I was always so busy. And I was always so fragmented and I had all this crap everywhere. And so I think that if we succeed at what we're doing, the role of these operations teams is fundamentally different in the future. And
Luigi La Corte: Mm.
Josh Whisenant: it's a it's a different industry to work in. But not for a second do I believe that we are that anybody's gonna lose their jobs because of us. so that was a long conversation. I think you started with like, why what about a project manager doing this themselves? And I'm saying, yeah, like actually in some ways that's what we're competing with.
Luigi La Corte: Mm. It it makes
Andrew Zukoski: Are you a mort-
Luigi La Corte: sense. Like sorry, Andrew, go ahead, and then I'll go with my point.
Andrew Zukoski: I was gonna ask like like my you guys know more about biote and procurement than I do, but like some companies precon does it, some companies ops does it, some companies have like a a a procurement department that centralizes the skills and standardizes the process so that
Josh Whisenant: That's the way that's the way it should be, by the way. That's the smart way to do it. But yeah, keep going.
Andrew Zukoski: I having dealt with a number of procurement departments as a vendor, I hesitate to recommend it, but I will state that you definitely get a better price if you have people who are like good at it doing right, 'cause it like a a
Mike Powers: Okay.
Andrew Zukoski: PM does it once every three years or whatever, and this like department can do it every week. And so you're but like standing that up organizationally is a lift.
Josh Whisenant: It's really hard.
Andrew Zukoski: and so you're in a way in a way Roger is your outsource procurement department. Okay.
Josh Whisenant: I mean That I mean that was a that was actually a positioning that we considered taking. because and that I mean that actually opens up another interesting idea that I have been thinking about which is really I mean the the c the job of the software engineer is gone right it doesn't exist anymore like of as it was as it was constrained a year ago like nobody at Roger has handwritten a line of code since like February You know, I I I see you looking off into space, Andrew, with you know, so forgive I'm sorry that that's an appropriate response to me kind of throwing out a fireball just to be annoying, right? One of the interesting theses that I have about the space is that This technology makes it possible to actually industrialize the contractor. So rather than a project manager that does 20 different things on a project and is almost like craft labor with like less opportunities to automate. Now that you actually have tools that can automate knowledge work, you can actually start to treat projects as assembly lines and
Mike Powers: Mm.
Josh Whisenant: you can start to actually centralize services and you can start to get real efficiency out of that. This is not a new idea. The o the OG of this is like D PR with V Construct, where they have all of these services to support their teams. but that was non technology augmented offshore labor. and, you know, you could take it a lot
Andrew Zukoski: was technology augmented. Like the I don't I'm gonna Yeah,
Josh Whisenant: Not intelligence augmented.
Andrew Zukoski: but I don't know that intelligence is what unlocks the systematization. like Yeah, you there's more value, but it's like digitization. Yeah.
Josh Whisenant: it it creates a stronger incentive. How about there's more val exactly. Yeah, yeah, yeah, exactly. Yeah, yeah, yeah. Talk yeah, sorry. I mean but the truth is guys, like I mean the real story of Roger is we started doing this as a service because we had just pivoted and we had no engineers. I mean that's really what happened. And suddenly everything worked in a different way. Where we were in the trenches with the customers, they trusted us, they talked to us differently, willingness to pay was dramatically higher. We were not perceived as a software vendor. and I feel like my whole journey is like figuring out what that means as the technology changes. And so I have all these ideas that I can spew out, but the reality is it's just a model that seems to like significantly change your perspective and for me for the better. and i in some ways it's like one step into the dark, you know. I don't know if I would have said that if I had slept more.
Luigi La Corte: Yeah. I like it it does make a lot of sense. I think they don't get enough reps in and the iteration times are just so long, you just can't learn that much on the like on the feedback loops, right? Like if you're a PM and you've written a scope of work, great, now you gotta figure out if it was bad, but that's gonna take, you know, three years before you figure out there's a dispute. This is why like in people's processes, like they go talk to the site super, they talk to other PMs and say, like, what am I missing in this in this trade? And like the overwhelming minority
Mike Powers: So I guess the qu
Luigi La Corte: of G Cs have a procurement department that are writing the scopes. It's like usually PM. Or the recon like becomes
Mike Powers: If you're big though, any like top twenty five G C would have a procurement group, except for White Turner.
Luigi La Corte: Yeah, white intern is is the exception.
Josh Whisenant: What are talking about? There's
Andrew Zukoski: No.
Mike Powers: Kerner has a full
Josh Whisenant: a bunch of that don't.
Mike Powers: procurement team for writing contracts.
Josh Whisenant: A bunch of them don't.
Andrew Zukoski: They might have a team, but we don't I I do not believe that it if if they have a procurement
Mike Powers: Like McCarthy does.
Andrew Zukoski: if
Josh Whisenant: I I
Andrew Zukoski: they have a procurement team then the procurement team is responsible. Then then it is then it is utilized on a small fraction
Mike Powers: Like procurement of right subcontracts. So like an estimate so turn
Josh Whisenant: Well wait wait wait wait wait wait wait wait wait wait.
Andrew Zukoski: of yeah.
Mike Powers: So there's there's two different models, right? There's there's the Turner model, which is an estimator does an estimate, it passes it off to a procurement person that buys out the contract, and you pass it off to the team that runs a job. And then the wedding turner model is the estimator is the procure, project manager, turnover, and they they're the full life cycle. Like those are the two models that are in market.
Luigi La Corte: There's three. There's a third one too. So there's like the procurement department takes the information from pre-con, writes the contract. The PM writes the contract. Pre-con writes the contract. Those are the those are the three that I see. And pre
Josh Whisenant: And I would disagree w or sorry, keep going. You should you should finish your bit, sorry.
Luigi La Corte: No, I was just gonna say like I over like probably because of the people we sell into, we usually see pre-con takes is front and center in finalizing the contract, but they they solicit feedback from procurement and from PMs where there is a procurement department.
Andrew Zukoski: Wait, wait, this is funny. So Luigi sells into precon and sees that it's all precon. Mike is like
Josh Whisenant: This is
Andrew Zukoski: deep in procurement and so sees all
Josh Whisenant: Exactly.
Andrew Zukoski: procurement bars. Josh comes from an ops background and sees all ops. I'm sitting here, I'm like, well, it's like there's a lot of each.
Josh Whisenant: No, that's so let me let me explain this because y I mean this is like the metaphor of the elephant, right? Where you know one guy says, it's a you know, it's three blind men are touching the elephant, and one person says it's rough, and one person says it's it's smooth, and you know, one person says it's leathery when he's touching the ear of the elephant. I think on a job site, on e every project, two things have to happen. The first thing is you have to understand the job well enough to Price it and decide if it pencils. And you have to do and the the name of the game there is basically like efficiency and speed. And this isn't this is usually a task that is done by estimating, but it's like, look, we believe that this job is going to come in at let's say 60 million bucks. And you know, we're probably missing a few things, but we've got these budgets. We've seen on previous jobs where we have fee erosion, and so we've put these budgets in. And so we're pretty sure that this number is solid. And by the way, we were able to put this together with a couple people in a couple weeks. This task has to be done at some point on every project because you have to know that to win the work. This is almost always the purview of estimating. And then another task has to be done, which is I'm going to go coordinate every freaking detail. And as a general contractor, I'm going to make sure, well, and by the end, this doesn't have to be done. It often doesn't, but it should be done. And the good and the great teams do do it, which is I'm going to go understand. Every stinking thing in this project with a level of detail that kind of makes me six and make sure that I know everything that needs to be done and everybody that I'm working with knows what needs to be done. And those two tasks, in my mind, they look like the same thing, and you could call them both scoping. But I think that they are fundamentally different with different incentives. Where if you ask a pro an estimator or the person that's doing the first task to do the second thing, what they actually see is that you're going to drive up price on the project and you're going to they're going to lose work. and if you ask the second person to do the first thing, they're like, Well, I'm gonna be on site in six months with these guys and I need them to know and I need to know so that they actually go and do it. 'Cause I'm not worried about trying to figure out if this job is gonna pencil anymore because it's going. I signed the GMP. Now, these two tasks flip around a ton. So for example, the argument that estimating like rights, the scopes at or procurement at a big GC, it changes project to project as far as I can tell. Where if it's an early award project, you know, GMP project, early award on 30%, they've done a lot of pre-con on it. They know it's their job, it's institutional, the money's there. ICGCs bring a project manager in, and then the project manager tells the estimators what to do. Now, the estimators would never say that this is what happens, but you've got a project manager that's like getting the GMP executed. now, I'm I'm I'm probably speaking to a lot of Uncertainty here where there are more processes than what I'm illustrating. But as far as I can tell, the way that this happens fluctuates a
Mike Powers: I haven't talked about this.
Josh Whisenant: lot on every job, and every party thinks that they do it. I can tell you that we used to call it scoping. And like at last year's Procor, project managers were like, This is great. I think my estimators do it. I called it buyout this year. I mean, you know, we got our, you know, we had our we talked to our seventy-five customers or whatever the number was. Nobody said, this is an estimator thing. Everybody was like, yeah, I do this. And I was like, imagine that. Does that mean that we're a fit on every project? Does that mean that I understand this perfectly? Hell no. Not at all. absolutely not. I I still I still find the entire thing like profoundly confusing.
Luigi La Corte: Yeah, it's a it's a great process and it's kind of crazy. like the whole thing actually doesn't make sense if you if you zoom out. So we're gonna get like like let's talk about it this way. Let's get a PM who's not a trained lawyer to write a contractually binding exhibit that bakes in profit for your entire subcontract across the project. and let's make sure that as the project gets bigger. We don't give you any more resources to write these scopes. but as as this as the amount of subcontracts get bigger, or as you get more subcontracts, the amount of interfaces with other subcontracts increases quadratically. So like you have to write this perfect scope, and then you have to write 30 additional perfect scopes that perfectly mesh so there's no strand. The whole thing is just like almost impossible. It's like a fool's errand.
Josh Whisenant: I mean
Mike Powers: But isn't like the the first principal problem is the documents aren't right. Like so, like in in j in those examples, Josh, that you went through, right? Like if you buy out 30%, you better have the PM involved because there's 70% gaps at minimum, right?
Josh Whisenant: They're gonna well
Mike Powers: We know I mean, I know it's not equal, but like so that whole bet on a GMP is like, hey, based on our experience, I know I can build a stadium or a hospital for seven hundred million dollars with this idea of a project. And then if I go build it for six hundred million, I'm gonna split the pr split the savings with you because you have a seven hundred million dollar budget. So then you're throwing everything out the wind to try to reduce cost. You're using your procurement group to go talk to multiple subcontractors that calls multiple projects, saying, hey, like if we go buy these projects together, what does that look like? So it could take a competitive advantage in that sense. There's all these like pieces that are moving, but it comes down to like the documents aren't the accurate assessment of the project. And the delta between
Josh Whisenant: And they never will be.
Mike Powers: a $400 million project and a 60 million like night and day, to the point where Turner
Josh Whisenant: Yeah, we've seen that,
Mike Powers: has asset it's called spe special projects division which does sixty million dollar projects, which they actually do it the opposite way, where they they do the white and turner world where they do like that PM runs the entire project process as opposed to big turner that does the estimate to procurement to per purchasing to turnover to the PM.
Josh Whisenant: Right.
Mike Powers: It's messy. And their competitive advantage is like, hey, we're gonna go do the project at as lowest fee as possible. We're gonna sell our ten other services.
Josh Whisenant: You know yeah. and one of the things that I had found that I mean I won't say d th I won't speak to the details of this, but the way that the money flows In these buys is absolutely as much governed by relationships and power and the shadow of the future as it is particular cost certainty. Like I've seen like people do crazy make crazy decisions to like hire their buddy. and you know, or or you'll see people. Change the budget, change their budgets with these trades by hundreds of thousands of dollars or more, basically on a dime and flip it around and it's like everybody knows that there's this pile of money that we're all kind of fighting over to get the job to go. and it's like it's not like anybody really knows how much this is going to cost them to go and do the job. It's not like anybody really knows the exact price of all this stuff. They're all kind of, you know, throwing a board throwing a dart at a dartboard. And the value that we can provide as a services provider is at least we can make the dartboard really, really clear. Right? And make it so that everybody knows the dartboard. So wherever you throw the dart, you kinda know what you're gonna be fighting through in the future. but all of this is why I'm like so convinced that we're like not taking anybody's jobs because I just see people make these decisions and I'm just like, my gosh, are you kidding me? This is insane. like, dude, I like I'm like working with a contractor in a major city, major US city, who is changing how they buy electricians because they're pretty convinced that some of the electrical suppliers in this major city are like buying wire from the mafia, from like the Russian mafia. And so they're like, there are these certain prices that are going down, and we don't, we're trying to decide if we trust them. I'm literally, I'm completely serious about this. Like, this is literally a conversation that we've had. And it's like, how the hell do you systematize that? Like what are you y what are y you know, this is not a this is not a you know, we this is not a a a a fig jam flow chart. It's like it's just like total chaos.
Luigi La Corte: Well, I I do have a solution for that. You have a skill that aggregates choices made by the PM and you reflect that on a project by project basis. Cause usually PMs do like, from what we've seen, they like to mirror their decisions project by project, and they have these preferences. Like we've seen like
Josh Whisenant: Well they certainly yeah, for sure. yeah.
Luigi La Corte: one PM wants to give their carpenter literally as much scope as possible. I think this was an anecdote that you gave me
Josh Whisenant: Yeah, I was gonna say.
Luigi La Corte: the the other day. but I see some other examples where for for their for their Mason, they just have preferences for how they do it in this jurisdiction. But if you go into this asset class, it completely changes. Who's doing the the the
Josh Whisenant: That's right.
Luigi La Corte: the caulking or who's doing like it all changes? There's so many individual preferences, but these things like these
Josh Whisenant: it's exactly.
Luigi La Corte: these fuzzy things can be captured with with text. Like in this situation, this over time, right? Like it it's
Josh Whisenant: I
Luigi La Corte: hard to capture on day one, but it's I think you can capture it over time because You know, it's a lives in a human's brain.
Josh Whisenant: Luigi and yeah, no, I mean that's right. And you know, we we do do that. you know, we've we've done multiple iterations of trying to do that. We did we actually built like a very a ri a very rigid system to capture those decisions on projects that we ran on a few and it was just it was just too much work. It was it w you know, you need to just do it as a s as a skill file. and so that's kind of what we have started to do. There's sort of an interesting moment for us where it's like, all right, so we know how they did it on their last job. We still probably want to check again and make sure and talk to them about it, right? Just because they had some idiosyncratic way to give the carpenter everything, you know, rough carpentry everything, maybe they don't like that guy anymore. Maybe he sucked on their last project. Right. And so to me, to me, what it really is is you've got a a person who's gonna own the job, who's gonna manage this whole thing end to end. And this guy or girl is going to hate their life for many years because these jobs just suck. Always. And that person needs to be in the driver's seat and needs to understand this as well as possible. It needs to be able to make as authoritative and fast decisions as humanly can be done. And yeah, like if we can get them to that point, then we've we've broadly succeeded. And you know what you know what the highlight of Procor for me was? And this was really weird. we've done a we did a bad job getting feedback on some of our early jobs that we did like know, just o just under a year ago or something. we had five contractors come back to us at Procor and say, that job you did, that was like as good of a job as we've ever done. I I I I I didn't really bel I don't really I'm I I I I if I if you guys said that, I would be like, that's marketing copy. You know, screw that. Like these guys literally told me this and I don't I you know I think maybe they're trying to like justify their involvement with us. And so I don't know if I totally believe it, but You know, I do think that there's something to be said for just like getting a person that's usually really busy and stressed to make a bunch of decisions really fast and understand what's going on really fast and drive clarity as quickly as you humanly can. it's but still, like the hell
Mike Powers: Okay.
Josh Whisenant: if I know how jobs get bought. I just don't get it. I don't really understand.
Andrew Zukoski: On that note. Luigi, you're Luigi's muted.
Mike Powers: Luigi's on mute. Yeah, he's so typing.
Luigi La Corte: Sorry, go ahead. Go ahead. I've lost my opportunity.
Josh Whisenant: Now Andrew's on mute.
Andrew Zukoski: Yeah, sorry, my son's waking up in the background. There we go. On that note.
Luigi La Corte: Comedy bears.
Josh Whisenant: What? Are we done?
Andrew Zukoski: Yeah, I gotta go feed my son breakfast.
Luigi La Corte: Maybe if you drop off, Andrew, we'll we'll wrap up. if you gotta
Andrew Zukoski: See you guys.
Luigi La Corte: go.
Mike Powers: See Andrew.
Josh Whisenant: So he's gone.
Luigi La Corte: Josh, where like I think the services are I the services are great. I love the stories of people saying it's the best job. Like that's as good of an endorsement as you can get, right? 'Cause like those are the the opposite of that is what you obviously want to avoid, right? Like the subcontractor issues, no one knowing what they're doing, the coordination issues that hurt profit. Do you do you see a world where Rogers used on everything, on every project? And if not, where does it make the most sense? Like if if someone's listening to this, when should they say, Okay, I should I should work with Josh and Roger?
Josh Whisenant: We're definitely not gonna be I mean, you know, the twenty year vision of, you know, Procore has shipped like thousands and thousands of tools. I wanna ship, you know, these operating systems of the the technically correct way to run a construction project, and we'll have all of these different services in 20 years of how you could do this. And so I hope that in some capacity we'll be a fit on all these jobs. you know, we didn't even talk about our submittals service line, but you know, that has a very specific those have specific profile jobs where we are a fit, and we there are some where we're not a fit.
Luigi La Corte: Give give us the give us the strap
Josh Whisenant: although actu
Luigi La Corte: give us the strap line on on the submittal services 'cause this is part of the this is part of the release as well.
Josh Whisenant: Yeah, like Propor announces a submittal agent. The problem is you've got your most junior, most stressed, most unhappy people in construction running all of your submittals on jobs. And it's just like completely overwhelming, right? So it's like, hey, put us in your workflows. We're gonna review your submittals for you. We'll keep your project engineer on the loop. They still like, you know, they're actually gonna learn the job better because they can like look at our summaries and read everything. We do shop drawings, although we still have some iterations to do to really get that like perfect, but we're getting really pretty good at it. We've built a lot of technology there. And so yeah, we've seen cycle time in submittals be cut by like literally 30 days. Revise and resubmit rates cut by half. It's awesome. It's crazy good. and like these project engineers are just like, my gosh, you saved my life. Like I like my job all of a sudden. Don't don't ask how my team feels about their lives. Like, you know, we it's like actually like really painful, unpleasant work, but it's been really good. But the point is, you know, so submittals where it's been the best actually is data centers because they don't really coordinate all this stuff up front. They're trying to go so fast and they do a lot of coordination and figuring the job out in the submittal process. So you have this massive volume of submittals on these big projects. And we just like, you know, we just get hammered by it all day. And
Luigi La Corte: Mm.
Josh Whisenant: so for the buyout line, I think that we are overkill on jobs under 10 million bucks, for example. Almost always. Unless it's a really complex job and somebody really needs the help. You know, there kind of needs to be the i I also think that if you can really get a project manager. Into pre-con in a very involved way, which happens less than any GC would like to advertise that they do. If you can really get a PM that understands it in say like an IPD IFOA contract, or like some really nuanced iterative buyout scheme, or if it's industrial without a lot of complexity, I'm not convinced we're a fit on those jobs. to me, where we are a fit is when risk shifts quickly from the owner to the contractor, where it's like we're really building this job, there's a serious amount of complexity on the project. With lots of interfaces and coordination, vertical construction, and there's a schedule that you need to hit and you need to move quickly. And then we seem to we seem to basically bat a hundred out of a hundred on those
Luigi La Corte: Mm.
Josh Whisenant: projects. But on an industrial, like on a big industrial warehouse where you know you're starting to do buyout on 30% DD based on the last four years and it's not that price sensitive, I think we're like zero out of a hundred on that. Yeah, we just haven't been a fit. Now people love us for submittals there. So I don't know. I have a lot of humility about where we fit, and I do, you know, I I turn away I turn away a reasonable number of projects. I I be you know, 'cause 'cause we can't bill. I'm not gonna bill you if we suck. If it doesn't help you, I'm not gonna bill you. And so then we
Luigi La Corte: Mm.
Josh Whisenant: go and we lose all this money. Like, I don't know how much money I've lost on like doing no fit projects this year, but I've lost at least seventy thousand dollars of revenue.
Luigi La Corte: And is that because you don't think the quality's good, so you don't bill?
Josh Whisenant: No, the quality's great. It's just that the quality the the output it the d you know, we have a widget machine that produces a widget and runs a process and the process wasn't what that job needed.
Luigi La Corte: I see. I see.
Josh Whisenant: And so they don't use it. And we can see it in the usage. We can see it in the metrics. We can get the feedback from the customer. We can see it how they don't respond to our emails. You know, the whole vibe is there. And so you send them a note where you're like, this clearly wasn't a fit for you. we'd love to do another one with you. We learned some lessons here. Here's what we learned. you know, the invoice is voided.
Luigi La Corte: Right. Yeah, I'm sure that's appreciated. Josh has been great.
Josh Whisenant: Sh not as much as they should be. Nobody ever gives me any credit for it. They're always like, All right, thanks, see ya.
Luigi La Corte: That's crazy. Well, this is great.
Josh Whisenant: Yeah.
Luigi La Corte: I'm I'm bullish. I think there's a a definite need, obviously. And augmenting the team's capacity with a human is great. It's like people have a lot of tools. That is true. So how do you escape that? And I think this makes a lot of sense. Congratulations on the lot on the launch, Josh. Let us know how it goes. And like you're you're our first second time returner to the CTP.
Josh Whisenant: I'm s I'm super honored to be on the scene. You guys, I mean, I actually think this is pretty fun. I it's
Luigi La Corte: It is fun.
Mike Powers: Yeah.
Luigi La Corte: Hello everyone, welcome to Construction Technology Podcast. I'm Luigi with Provision. We got Andrew with Join, Mike with Build Vision, and we got a special guest, Josh from Roger. What's up, gentlemen? How are you? Happy Friday. Josh, how are you?
Mike Powers: Happy Friday.
Josh Whisenant: Little sleep deprived, but that's always how it goes. How are you guys?
Luigi La Corte: Yeah. Yeah, good. What what time did you get in last night from Orlando? You flew last night, obviously.
Josh Whisenant: Yeah, I did. yeah, I I have no fun at the groundbreak brat bash, you know. I'm what's when did get home? I think I got home at like eleven thirty mountain time, so that would have been one thirty AM Eastern. So f you know, it's not so bad, but you know, I need my beauty sleep. Come on.
Luigi La Corte: Yeah, of course. That's what that's it's working for you,
Mike Powers: So
Luigi La Corte: dare I say. it was really fun. I mean, n not to not to rub it in, but they they rented out Islands of Adventure at Universal Studios. And it was amazing just to see like so many fifty year old men kind of giddily scream to be at the front of a roller coaster ride after ride and then get up there like we're we're going again.
Mike Powers: So is this the last hurrah before they crush every AI tool? Is that is that the pitch? They're like, Hey, we have this new model
Luigi La Corte: Yeah, interesting.
Mike Powers: and come have fun, you know, but like it's like the giving the dog a a hamburger before you take it to the vet. Is that is that we're at or not yet?
Luigi La Corte: Yeah, interesting. Maybe. Well, they did launch a model. Maybe we can do a bit of like a groundbreak retro.
Josh Whisenant: Yeah, sure.
Luigi La Corte: because Josh was there with us too, but they did release a model, Magpie, Magpie three point I wasn't aware that there were previous versions of Magpie, but
Josh Whisenant: Or was I?
Andrew Zukoski: start I there there might not have been.
Luigi La Corte: so Magpie Yeah that's right. That's true. Like if you just start if you just release a model with like, yeah, this is provision intelligence eight point guys, where have you been? You're like, well, fuck,
Josh Whisenant: Right.
Luigi La Corte: maybe maybe we've missed it. But anyway, MagPoint three point So Magpie, they it's frontier according to their benchmark. The problem with these benchmarks is like if it's not a third party benchmark, it's just so manipulatable. Like on this one very discrete task, you know, are we better? It doesn't mean you're frontier unless it's like this very small task. But anyway, so magpie three point it's a reasoning model to choose between their taxonomy of document types, what actions to take. I mean, there were so many words thrown around like model, ontology. They released Procore ontology, but you know, didn't really get into it. but for like I guess the like what a customer would care about is Magpie helps get tasks within Procore done better than GPT six Astra with API access. So it's better and that's good. I like you know, I don't know, that's it. We didn't really see much.
Mike Powers: They throttle what do you do you get to how does it work? Do you get tokens of the customer? And you're only gonna use 10 tokens a month. How does it work?
Luigi La Corte: I don't think they actually sh I don't know if they give you a token allowance. They probably do. But we could talk like very quickly. They do.
Josh Whisenant: No, they I think they do. I mean, I've heard I heard two things where I heard people saying, yeah, they're trying to get me to pay forty thousand dollars a year for Procore AI, but then they had this control tower thing where it was tracking all your token spend. So it seems like there's some combination of fixed end variable costs.
Luigi La Corte: I see. Is that just within Procore or is that like like a ramp type initiative where it's like all token costs?
Josh Whisenant: No, no, no. I think it was just within Procore.
Luigi La Corte: I see. Yeah. I I wonder why would they why they would I mean, yeah, like they have to limit their cost too, but I thought it was interesting how they're packaging it, Josh. Did you see how they were packaging Procore AI? So you can't buy an individual agent. You buy packs of agents. And I went to the innovation hub and yes, that's it. Digital coworker starter packs. You're buying starter packs. the starter pack is like
Josh Whisenant: Yeah yeah yeah.
Luigi La Corte: five agents, I believe. So deep search agent. You have a submittal agent
Mike Powers: What does that mean?
Luigi La Corte: deep search.
Mike Powers: No, what is a what I think what does an agent mean? Like why is there
Luigi La Corte: It's
Mike Powers: not just one agent?
Josh Whisenant: Search S
Mike Powers: It's Procore.
Josh Whisenant: Search doesn't feel like a very search feels like not an agent, right?
Andrew Zukoski: 'cause they they announced they announced that they announced that two two groundbreak keynotes ago with Procor.
Mike Powers: Here's agents. So how how are these not I mean, how are these not just skills?
Luigi La Corte: They they really are. I mean, a lot of like the the big model companies have really moved away from individual agents. It's abstracted away, right? Like if you prompt codecs, you will see the agents that are invoked, but they're created on the fly, right? Like an adversarial code agent or like a legibility agent. Honestly, the customer doesn't want to pick the agent. Like it's why?
Andrew Zukoski: Wait, I don't know if that's true. have you spent any time we've been rolling out our MCP server and so have been spending more time inside the Microsoft ecosystem. and there and you could argue this is necessarily behind the c like behind the times because I just said Microsoft, but there you package like whatever the underlying MCP is, like a skill into an agent, but then you whitelist tools and endpoints. At the agent level. Like packaging it reduces the scope in a way that lets you then like expand the permissions that people can access or edit it. Like you might want a lot of people to be able to review an RFI, but you might not want everyone to be able to submit a pay app to the owner or something like that. to execute
Luigi La Corte: Mm.
Andrew Zukoski: whatever the underlying capabilities are, even though they all rest on like the same model and the same skills and the same tool set. So I don't know if it's that
Luigi La Corte: Mm. Yeah.
Andrew Zukoski: like
Luigi La Corte: I mean that that makes sense from a permissions perspective. Like you cannot let's say like Josh, like you can't have access to you know pay apps because like we don't wanna we wanna gate that.
Andrew Zukoski: We we don't trust you.
Luigi La Corte: Yeah, we don't trust you. But I I I I wonder I'm I'm kind of bearish on the on the UX of a user having to invoke an agent to get a task done. I think
Andrew Zukoski: Yeah.
Luigi La Corte: knowing what the agent can do. And then having to choose it, I think is gonna be too much friction long term. It should just be present, it should do the work when it makes sense, or it should be invoked when the context justifies it, which is effectively just like how agents invoke skill. So that's why, like, yeah, like the agent skill thing to me is just like maybe it's a little bit pedantic. But Josh, what what did you think about the I don't know if you caught any of the keynotes, but what was your impression of the event at like at large about Procore and their AI?
Josh Whisenant: I mean yeah. I mean Propour is always such a it's just such a pleasure to go to that conference. because you really just see your people and you see all of these all of these contractors that you work with and it's really much more than like any other conference. I don't know. I I don't understand quite how they have all the all the people that we like to work with like just like walking in torrents down that hall, but like you see them the whole day and it's just super fun. I mean I have a lot of thoughts on pro core AI, not all of them are necessarily They're not all in a perfect flow, right? you know, as far as Procor's benchmark, I'm going to present a certain exhibit for consideration by the jury, which is the iconic you know, benchmark modeling
Andrew Zukoski: Ha ha ha.
Josh Whisenant: system. and you you can see scales on this where, you know, over here it's like sixty nine point one and sixty-nine point three or something, you know. and so it's kind of like you always see the benchmark and it's like, okay, here's the benchmark, show me the money, right? now I don't think that that is I think that saying that ProPore sucks at AI is not really a very useful thing because they have, you know, billions and billions of dollars of capital, I assume, and they've got a ton of money and they can throw on a huge conference. And so at some point they should figure out how to be competent. I think that there is an inherent incentive structure issue with ProPore right now where they have increased their costs for these contractors dramatically in the last five years. Right. sometimes as much as 50%, largely to show public market show growth to public market investors. And like everybody that walked up to my booth was like so freaking pissed at how expensive Procore is and how they feel like they're getting jerked around by them all the time. And to me, this package of all of this stuff you can go spend money on in software inside of Procore, it just kind of feels like it's serving Procor's need to increase average contract values rather than actually saying like, I guarantee you we're gonna make construction run better. Now that doesn't mean that their technology is bad. That doesn't mean that people won't get value out of these tools. But to me, like I'm picturing a project manager using this stuff in
Mike Powers: Yeah.
Josh Whisenant: in a trailer and like, is the project like a hundred times better? Maybe you know, I I I don't I I don't quite understand how this is going to change the world and their their strategy to me hasn't quite made sense. but you know, again, for abundance of clarity, I you know, I think that Procor has built a lot of software. They've built a huge business, a much bigger business than anything that certainly than anything I've ever built by like many orders of magnitude, right? And so, you know, the last thing that you want to do is be a hater because that'll you'll put your foot in your mouth really fast. last thing I'll say. Traversing through job site context to have a c to be ready to have a conversation with people is a huge pain in the ass for project managers. I remember doing that. Right? Like, I see this is here in this drawing and like, what's the spec? And what's the you know, nobody likes doing that. you know, and this is why I mean, this is kinda like what Trump Tools has tried to basically solve, is just like retrieval and this is what it seems like Procore is leaning into with bagpie tasks. Yeah, like they should absolutely solve that problem because it's like a real issue that everybody hates.
Andrew Zukoski: Aren't a lot of people chewing on that problem? I mean I guess Procor would be well well situated to go solve it.
Josh Whisenant: I mean yeah, i i okay, yeah, totally. And I mean you know our technology actually has been really good at that lately, which is a almost another story. it's kind of but i i th what what I'm illustrat what I'm the reason I said that is because I do think that the worst thing that you can do as a contractor right now is not is do anything that pr keeps you from benefiting from the unbelievable commoditization of these models and intelligence that is going on right now. And to me, you should grab as tightly as you can to the wave of the models getting better every two weeks and getting cheaper every two weeks and hold on to it. And the idea that I'm gonna pin my entire business on Procore frickin' magpie, that even if it is as good as they say it is, you know, you know who I don't want to compete with? Sam Altman and and Dario Amade. Like, you know, it's like I I i i it it actually felt like a real strategic misstep to me to try and compete with the foundation models and say that Procor was better than them.
Luigi La Corte: there's some I I think there's this there's something behind it. It's I think it's more commercial than it is functional. So AJ
Josh Whisenant: A hundred percent. Exactly, right? That's what I just said.
Luigi La Corte: Yeah I look I I agree with you like but I think the the reason behind it is okay Procore's price to earnings right now is is less than 10. AI stocks let's say you know broadly speaking trade at 25x PE so they are
Andrew Zukoski: But wait, which one? None of the foundation models have earnings.
Luigi La Corte: Well, I'm talking about Meta, Google, NVIDIA. I'm not talking about open AI and Anthropic.
Andrew Zukoski: Okay, fair. Yep.
Luigi La Corte: Obviously they're not publicly traded, right? But like roughly
Mike Powers: Mm.
Luigi La Corte: speaking, the the there's a premium to AI companies that's being assigned, right? and I think that's probably the bet 'cause like there was a lot of conversation in the keynote about becoming an AI company. And it's like so much so that he like Like I get but I think I think it's functional. Like that's why they said, Okay, we want to be an AI company, we're gonna hire technical talent. We are now assigning Nicholas Plinkington from Drone Deploy to be the principal architect of Frontier Models. They have an applied AI team that's building models. They want to benefit from the increased PE ratio. Like they want they want that that that premium. And then, you know, what what are they gonna do with it? Well, I don't know, maybe they maybe they're gonna be viewed as defensible. But I think my point, I was talking to someone from ProPro I was like, You guys just need to grow faster. Like the only reason you're really getting that premium is not because people view you as making models. Those models have to turn into growth. And like will it really matter to the customer? I'm not sure.
Mike Powers: I they need double revenue. Like they needed like they're gonna double revenue or P firms gonna buy them. Like that it like
Luigi La Corte: What do you think? Say more.
Mike Powers: there's their their growth, their the stock is getting crushed. Growth is not where it needs to be. They cannot squeeze another billion dollars of revenue out of general contractors. So they need to make a bet. This is their bet. Their bet is, hey
Luigi La Corte: So existing customer existing customer value extraction or at value add and taking
Mike Powers: I think if can do tasks on behalf of customers, they can get closer to outcome-based pricing. similar to what Intercom did with their switch to fin, then you get to outcome based pricing. And this rolls
Andrew Zukoski: Intercom.
Mike Powers: into very similar what Josh is doing. Josh and Roger is now right, we're we're selling outcomes or trying to sell outcomes. And if they're in the place
Josh Whisenant: Trying to.
Mike Powers: where they can have a platform tool that is a minimal cost. And they can sell outcomes on top of that and take on risk because they have specific proprietary data. There's a path there, there's a story there.
Josh Whisenant: That's what I I mean, obviously that is what I would do if I was Procor. That's kinda what I'm doing when I'm here.
Andrew Zukoski: It's interesting that that's what like without sorry we should transition to Roger because I'm interested in the new offering. The in a way that's what Procore Risk Advisors was a swing at. And like, you know, the past is not the future.
Mike Powers: They've hired their own team. They hired their own team.
Josh Whisenant: My understanding on that was that they were trying to not really take accountability on any outcome. They were just trying to like do billable hours, right? Am
Andrew Zukoski: Well
Josh Whisenant: I wrong about that?
Andrew Zukoski: on the risk advisors, I I thought it was like
Mike Powers: There's a risk advice. Yeah.
Josh Whisenant: Yeah.
Andrew Zukoski: I thought the play was and you guys it sounds like you know more about it than I do, but I thought the idea
Josh Whisenant: I don't.
Andrew Zukoski: was like, okay, we have all this data, we understand which companies perform better, if you use our system, we'll be able to underwrite or encourage people to underwrite at like a lower rate. Like
Mike Powers: It they wanted to build they wanted to build Shepard, they wanted to build Shepard.
Josh Whisenant: yeah, totally.
Andrew Zukoski: Yeah.
Mike Powers: and then they didn't hire the right team. And now Shepard's owning that market. This is going back to there's outside of general, like there's like 10 revenue streams for a general contractor. The so the the wedge they're in now is an IT budget. They need to get in profit share budgets. And we go back to where what Turner's
Andrew Zukoski: Pay. Procore pay could do that.
Mike Powers: what Turner's doing at scale, and how do you tap into each one of those levels? So Risk Advisors was SDI C SIP. they're starting to do equipment management tracking, which is what Turner does with with first equipment, where you actually rent lift scissor lifts. so they need to go into the additional revenue streams of these customers, which is harder.
Andrew Zukoski: So it's interesting in this way all these agents that are automating things that a person in a trailer or an office would do inside of Procore is trying to go deeper inside the same category of project management and paperwork. It's interesting.
Mike Powers: Correct. But a selling outcome. Hey, I don't have enough people on my job site to hire a pro core agent to do cemental reviews, whatever it is.
Andrew Zukoski: Well, I'm not I'm not gonna not have a person on my job site and I'm already grumpy about how much I'm paying Procor. It's gonna that's gonna be interesting. Hm. But Roger's selling outcomes
Luigi La Corte: So let's go let's go into let's go into Roger. Yeah, so Josh, so tell us about the launch. and bring bring us high level through what was launched, what's the what's different from what was before. And then I mean the website looks amazing. It'd be great if you like even if you
Josh Whisenant: Thank you.
Luigi La Corte: if if you can, like walking us through the messaging and like what people would expect would be great.
Josh Whisenant: I think that the launch We needed to do a few things with our launch. the first thing was generally have a website that wasn't built
Andrew Zukoski: Ha ha.
Josh Whisenant: with me vibe coding on a over a weekend in Replit. you know, we're like a we're like a good company and we should like not have a stupid website even if we're kinda lazy. I think the second thing This is more of a commentary on, I think, the nature of services than any particular problem. But we were often perceived as a pre-construction tool. which by the way, we never intended to get into pre-construction. That was not the intention when we started doing this, but we kind of got pushed there and we kind of let the customers pushing us there. Tri you know, we tried to figure it out there. and that didn't really work very well. That wasn't we didn't have the right fit. No, it worked sometimes, but it wasn't great. And So it was important to clarify that we are not a Team that is just doing a a widget where we just like process something, but we actually carry you to an outcome that's in operations. You know, the way that we bill was an interesting thing to add where we talk about how we're, you know, staking our fiana. We're not billing if w we don't give you outcomes. That's actually kind of You look thoughtful, Andrew.
Andrew Zukoski: What is that I would love to hear maybe you could pull up contract language, but I would love to hear like what the what the outcome is. Like I it's been a minute since I looked at the website, but like there's there's some part that's like okay, there's under there's specific tasks that are happening, and I can sort of understand what the finish line is. But then there's like, hey, if you're unhappy we'll refund you which is like if the outcome is you're happy with us we sell outcomes too, and I think you go further than that. Like
Josh Whisenant: Yeah, I mean the th th and this is a h this is a hard thing about like I mean now we're getting into the idiosyncrasies of outcomes of like how to bill for these things. It's very hard to like It's very hard to quantify quality, right? But it's kind of like, you know, you know it when you see it. And so we basically say, hey, if you don't feel like this is better than anything you've ever seen before, and it doesn't actually match what you want to do, then we just won't send you an invoice. So that's kind of the approach to quality. And then if there's anything that comes up later on down the road that was egregious that we all missed, then we'll be on the hook for that. if it's like, you know, if we just clearly missed something that was completely right there and obvious.
Andrew Zukoski: Do do you take contractual risk? Like if there's a scope gap when it's bought out and they've signed a G P based on the buyout that you did and there was a scope gap in between the G P and the subcontracts that you bought out, do you take on that risk?
Josh Whisenant: Up to the I mean, depending on the scope gap, yes, but only up to the total amount of our fee.
Andrew Zukoski: Okay.
Josh Whisenant: Which I would argue is
Andrew Zukoski: Okay.
Josh Whisenant: much more which is much more accountable than a project manager. Right? Like nobody's losing their salary when they have a scope gap. They might lose their bonus, I guess.
Andrew Zukoski: Yeah, I mean they
Luigi La Corte: Yeah, it'll be fine.
Andrew Zukoski: get fired. Or not. I mean I don't know. People there peop mistakes happen all the time, right? Like that that's
Josh Whisenant: Right, right, right.
Andrew Zukoski: part of it, but you like Okay, so you put the you put the fee at risk for scope gaps.
Josh Whisenant: I mean, what I can tell you is that I c I keep on going back to people over the last year of over the projects we've done the last couple of years and I keep saying, Hey, what have we missed? What have we screwed up? And nobody can tell me anything. I and and so it it it makes me feel like, you know, at some point we're gonna miss something and there's gonna I'm gonna have to have an unpleasant conversation. but that just We're just not even close to that. I it just doesn't seem like that's what's going on.
Luigi La Corte: Josh, do you typically I I haven't seen the scopes, but are you only providing the project specific scope? Are you also modifying their template for some kind of like broad language, like including but not limited to and as per plans and specs, but enumerated below? Like obviously that language was
Josh Whisenant: We do all of that stuff.
Luigi La Corte: you're doing all that stuff too.
Josh Whisenant: Yeah. And that's been that's been a big technology piece of work to figure that all out. Because it kinda depends on what everybody w you know, it has to be super flexible 'cause every every Bill Bob, Jim and Joe wants it a little different.
Luigi La Corte: Yeah. Yeah. Yeah. The r the reality the reality is is like irrespective of what's written there, it's like they need they need to understand it. Because it's not about the coverage. Cause like I think you
Josh Whisenant: That's right.
Luigi La Corte: had a you had a LinkedIn post. It was it was I I found it funny for this the scoping community. But hey, I found a new way to scope, just say as per plans and specs. We're done. And you like you can do that, your coverage, you're contractually right. but like the fact that you're gonna have a fight means now that you have to dispute it and have the argument.
Josh Whisenant: It's
Luigi La Corte: That doesn't go away.
Josh Whisenant: Yeah, that no, that's right. and actually
Luigi La Corte: You just gotta force them to read it. It'd be like, Look at this, read those thirty points.
Josh Whisenant: W I mean what we do, well, you know, I think it's very likely that we will get into I think we're gonna meet with the trades at some point here and talk to them about talk to the trades about our scopes along with their project manager in the room. and you know, so walking through the 30 points. But I think that a lot of, you know, and Luigi, this is me you know tipping my hand a little bit, but I feel like a lot of the value that we provide is in the deep conversations that we have prepared for our customers where we just go through all of these questions and discussion points to figure out the entire job. And I I feel like that's kind of, you know, where all of the nuance comes out is these debates that we have. And it's not, you know, we c it's not a half an hour debate. It's i i they're pretty serious conversations that we have for a long time.
Luigi La Corte: So what what do you think about let's say like the 'cause I think about that and I think that's great. Like that's that's a level of service like I couldn't provide. but what do you think about PMs and their like shouldn't they be able to have those conversations themselves and and do that work?
Josh Whisenant: Hundred percent. A hundred percent. and this I mean this brings you to a question of What is the problem that we really are solving and what is the future of the construction manager? Because Empirically, when you look at task completion outcomes, you're actually competing with employees. That's really the competition, right? Like if we're do if we're gonna go and buy this job out together, then I'm telling to the executive of the company, to the you know, VP ops or whatever, hey, I'm gonna go run this buyout process better than your PM can. I mean, that's really the the argument, and I don't, you know, I that that is the truth. I think if I think that's a bad so that like that's a really bad business model in pre-con in estimating. for a variety of reasons. It's really good for project managers for a variety of reasons. First, at the start of these jobs, there's an unbelievable amount of work to do all across the board, and there's a huge opportunity to like n fix things and not break things. Second, you have project managers that are like coming off of previous jobs, right? Third, a project manager does buyout. Once every three years, right? Often. and we do it, you know, three or four times a week. and so actually what you and and f and so and fourth like these project managers They understand the risk that they take on in these processes and so they're competent and present and paying attention to things. But it's not like I don't like anybody like loves writing contracts, whether with Claude or by themselves in a room or using a template or, you know, anything. I don't think that these this is a task that people are like, this is what I live to do. and so there's this moment here where in these particular tasks on jobs, labor shifting is actually acceptable to everybody in the ecosystem. Actually, more than acceptable on certain projects, it's very welcome. and so then the pitch that you have to make and the narrative that you have to articulate is if you're trying to clearly expand your business and you know we're moving into submittals and it, you know, clearly you the the The implication of that is that we do many things over time. What is the future of these people at these contractors? Right? are you just trying to are you the guy that's like gonna take all the jobs? I think we are not the people taking all the jobs, we are the people that are gonna change all the jobs. I think we can do that. Where as far as I can tell, there's like five times as much work that needs to be done on these jobs and there is people that actually do it. I lived this job, I did a terrible job because I was always so busy. And I was always so fragmented and I had all this crap everywhere. And so I think that if we succeed at what we're doing, the role of these operations teams is fundamentally different in the future. And
Luigi La Corte: Mm.
Josh Whisenant: it's a it's a different industry to work in. But not for a second do I believe that we are that anybody's gonna lose their jobs because of us. so that was a long conversation. I think you started with like, why what about a project manager doing this themselves? And I'm saying, yeah, like actually in some ways that's what we're competing with.
Luigi La Corte: Mm. It it makes
Andrew Zukoski: Are you a mort-
Luigi La Corte: sense. Like sorry, Andrew, go ahead, and then I'll go with my point.
Andrew Zukoski: I was gonna ask like like my you guys know more about biote and procurement than I do, but like some companies precon does it, some companies ops does it, some companies have like a a a procurement department that centralizes the skills and standardizes the process so that
Josh Whisenant: That's the way that's the way it should be, by the way. That's the smart way to do it. But yeah, keep going.
Andrew Zukoski: I having dealt with a number of procurement departments as a vendor, I hesitate to recommend it, but I will state that you definitely get a better price if you have people who are like good at it doing right, 'cause it like a a
Mike Powers: Okay.
Andrew Zukoski: PM does it once every three years or whatever, and this like department can do it every week. And so you're but like standing that up organizationally is a lift.
Josh Whisenant: It's really hard.
Andrew Zukoski: and so you're in a way in a way Roger is your outsource procurement department. Okay.
Josh Whisenant: I mean That I mean that was a that was actually a positioning that we considered taking. because and that I mean that actually opens up another interesting idea that I have been thinking about which is really I mean the the c the job of the software engineer is gone right it doesn't exist anymore like of as it was as it was constrained a year ago like nobody at Roger has handwritten a line of code since like February You know, I I I see you looking off into space, Andrew, with you know, so forgive I'm sorry that that's an appropriate response to me kind of throwing out a fireball just to be annoying, right? One of the interesting theses that I have about the space is that This technology makes it possible to actually industrialize the contractor. So rather than a project manager that does 20 different things on a project and is almost like craft labor with like less opportunities to automate. Now that you actually have tools that can automate knowledge work, you can actually start to treat projects as assembly lines and
Mike Powers: Mm.
Josh Whisenant: you can start to actually centralize services and you can start to get real efficiency out of that. This is not a new idea. The o the OG of this is like D PR with V Construct, where they have all of these services to support their teams. but that was non technology augmented offshore labor. and, you know, you could take it a lot
Andrew Zukoski: was technology augmented. Like the I don't I'm gonna Yeah,
Josh Whisenant: Not intelligence augmented.
Andrew Zukoski: but I don't know that intelligence is what unlocks the systematization. like Yeah, you there's more value, but it's like digitization. Yeah.
Josh Whisenant: it it creates a stronger incentive. How about there's more val exactly. Yeah, yeah, yeah, exactly. Yeah, yeah, yeah. Talk yeah, sorry. I mean but the truth is guys, like I mean the real story of Roger is we started doing this as a service because we had just pivoted and we had no engineers. I mean that's really what happened. And suddenly everything worked in a different way. Where we were in the trenches with the customers, they trusted us, they talked to us differently, willingness to pay was dramatically higher. We were not perceived as a software vendor. and I feel like my whole journey is like figuring out what that means as the technology changes. And so I have all these ideas that I can spew out, but the reality is it's just a model that seems to like significantly change your perspective and for me for the better. and i in some ways it's like one step into the dark, you know. I don't know if I would have said that if I had slept more.
Luigi La Corte: Yeah. I like it it does make a lot of sense. I think they don't get enough reps in and the iteration times are just so long, you just can't learn that much on the like on the feedback loops, right? Like if you're a PM and you've written a scope of work, great, now you gotta figure out if it was bad, but that's gonna take, you know, three years before you figure out there's a dispute. This is why like in people's processes, like they go talk to the site super, they talk to other PMs and say, like, what am I missing in this in this trade? And like the overwhelming minority
Mike Powers: So I guess the qu
Luigi La Corte: of G Cs have a procurement department that are writing the scopes. It's like usually PM. Or the recon like becomes
Mike Powers: If you're big though, any like top twenty five G C would have a procurement group, except for White Turner.
Luigi La Corte: Yeah, white intern is is the exception.
Josh Whisenant: What are talking about? There's
Andrew Zukoski: No.
Mike Powers: Kerner has a full
Josh Whisenant: a bunch of that don't.
Mike Powers: procurement team for writing contracts.
Josh Whisenant: A bunch of them don't.
Andrew Zukoski: They might have a team, but we don't I I do not believe that it if if they have a procurement
Mike Powers: Like McCarthy does.
Andrew Zukoski: if
Josh Whisenant: I I
Andrew Zukoski: they have a procurement team then the procurement team is responsible. Then then it is then it is utilized on a small fraction
Mike Powers: Like procurement of right subcontracts. So like an estimate so turn
Josh Whisenant: Well wait wait wait wait wait wait wait wait wait wait.
Andrew Zukoski: of yeah.
Mike Powers: So there's there's two different models, right? There's there's the Turner model, which is an estimator does an estimate, it passes it off to a procurement person that buys out the contract, and you pass it off to the team that runs a job. And then the wedding turner model is the estimator is the procure, project manager, turnover, and they they're the full life cycle. Like those are the two models that are in market.
Luigi La Corte: There's three. There's a third one too. So there's like the procurement department takes the information from pre-con, writes the contract. The PM writes the contract. Pre-con writes the contract. Those are the those are the three that I see. And pre
Josh Whisenant: And I would disagree w or sorry, keep going. You should you should finish your bit, sorry.
Luigi La Corte: No, I was just gonna say like I over like probably because of the people we sell into, we usually see pre-con takes is front and center in finalizing the contract, but they they solicit feedback from procurement and from PMs where there is a procurement department.
Andrew Zukoski: Wait, wait, this is funny. So Luigi sells into precon and sees that it's all precon. Mike is like
Josh Whisenant: This is
Andrew Zukoski: deep in procurement and so sees all
Josh Whisenant: Exactly.
Andrew Zukoski: procurement bars. Josh comes from an ops background and sees all ops. I'm sitting here, I'm like, well, it's like there's a lot of each.
Josh Whisenant: No, that's so let me let me explain this because y I mean this is like the metaphor of the elephant, right? Where you know one guy says, it's a you know, it's three blind men are touching the elephant, and one person says it's rough, and one person says it's it's smooth, and you know, one person says it's leathery when he's touching the ear of the elephant. I think on a job site, on e every project, two things have to happen. The first thing is you have to understand the job well enough to Price it and decide if it pencils. And you have to do and the the name of the game there is basically like efficiency and speed. And this isn't this is usually a task that is done by estimating, but it's like, look, we believe that this job is going to come in at let's say 60 million bucks. And you know, we're probably missing a few things, but we've got these budgets. We've seen on previous jobs where we have fee erosion, and so we've put these budgets in. And so we're pretty sure that this number is solid. And by the way, we were able to put this together with a couple people in a couple weeks. This task has to be done at some point on every project because you have to know that to win the work. This is almost always the purview of estimating. And then another task has to be done, which is I'm going to go coordinate every freaking detail. And as a general contractor, I'm going to make sure, well, and by the end, this doesn't have to be done. It often doesn't, but it should be done. And the good and the great teams do do it, which is I'm going to go understand. Every stinking thing in this project with a level of detail that kind of makes me six and make sure that I know everything that needs to be done and everybody that I'm working with knows what needs to be done. And those two tasks, in my mind, they look like the same thing, and you could call them both scoping. But I think that they are fundamentally different with different incentives. Where if you ask a pro an estimator or the person that's doing the first task to do the second thing, what they actually see is that you're going to drive up price on the project and you're going to they're going to lose work. and if you ask the second person to do the first thing, they're like, Well, I'm gonna be on site in six months with these guys and I need them to know and I need to know so that they actually go and do it. 'Cause I'm not worried about trying to figure out if this job is gonna pencil anymore because it's going. I signed the GMP. Now, these two tasks flip around a ton. So for example, the argument that estimating like rights, the scopes at or procurement at a big GC, it changes project to project as far as I can tell. Where if it's an early award project, you know, GMP project, early award on 30%, they've done a lot of pre-con on it. They know it's their job, it's institutional, the money's there. ICGCs bring a project manager in, and then the project manager tells the estimators what to do. Now, the estimators would never say that this is what happens, but you've got a project manager that's like getting the GMP executed. now, I'm I'm I'm probably speaking to a lot of Uncertainty here where there are more processes than what I'm illustrating. But as far as I can tell, the way that this happens fluctuates a
Mike Powers: I haven't talked about this.
Josh Whisenant: lot on every job, and every party thinks that they do it. I can tell you that we used to call it scoping. And like at last year's Procor, project managers were like, This is great. I think my estimators do it. I called it buyout this year. I mean, you know, we got our, you know, we had our we talked to our seventy-five customers or whatever the number was. Nobody said, this is an estimator thing. Everybody was like, yeah, I do this. And I was like, imagine that. Does that mean that we're a fit on every project? Does that mean that I understand this perfectly? Hell no. Not at all. absolutely not. I I still I still find the entire thing like profoundly confusing.
Luigi La Corte: Yeah, it's a it's a great process and it's kind of crazy. like the whole thing actually doesn't make sense if you if you zoom out. So we're gonna get like like let's talk about it this way. Let's get a PM who's not a trained lawyer to write a contractually binding exhibit that bakes in profit for your entire subcontract across the project. and let's make sure that as the project gets bigger. We don't give you any more resources to write these scopes. but as as this as the amount of subcontracts get bigger, or as you get more subcontracts, the amount of interfaces with other subcontracts increases quadratically. So like you have to write this perfect scope, and then you have to write 30 additional perfect scopes that perfectly mesh so there's no strand. The whole thing is just like almost impossible. It's like a fool's errand.
Josh Whisenant: I mean
Mike Powers: But isn't like the the first principal problem is the documents aren't right. Like so, like in in j in those examples, Josh, that you went through, right? Like if you buy out 30%, you better have the PM involved because there's 70% gaps at minimum, right?
Josh Whisenant: They're gonna well
Mike Powers: We know I mean, I know it's not equal, but like so that whole bet on a GMP is like, hey, based on our experience, I know I can build a stadium or a hospital for seven hundred million dollars with this idea of a project. And then if I go build it for six hundred million, I'm gonna split the pr split the savings with you because you have a seven hundred million dollar budget. So then you're throwing everything out the wind to try to reduce cost. You're using your procurement group to go talk to multiple subcontractors that calls multiple projects, saying, hey, like if we go buy these projects together, what does that look like? So it could take a competitive advantage in that sense. There's all these like pieces that are moving, but it comes down to like the documents aren't the accurate assessment of the project. And the delta between
Josh Whisenant: And they never will be.
Mike Powers: a $400 million project and a 60 million like night and day, to the point where Turner
Josh Whisenant: Yeah, we've seen that,
Mike Powers: has asset it's called spe special projects division which does sixty million dollar projects, which they actually do it the opposite way, where they they do the white and turner world where they do like that PM runs the entire project process as opposed to big turner that does the estimate to procurement to per purchasing to turnover to the PM.
Josh Whisenant: Right.
Mike Powers: It's messy. And their competitive advantage is like, hey, we're gonna go do the project at as lowest fee as possible. We're gonna sell our ten other services.
Josh Whisenant: You know yeah. and one of the things that I had found that I mean I won't say d th I won't speak to the details of this, but the way that the money flows In these buys is absolutely as much governed by relationships and power and the shadow of the future as it is particular cost certainty. Like I've seen like people do crazy make crazy decisions to like hire their buddy. and you know, or or you'll see people. Change the budget, change their budgets with these trades by hundreds of thousands of dollars or more, basically on a dime and flip it around and it's like everybody knows that there's this pile of money that we're all kind of fighting over to get the job to go. and it's like it's not like anybody really knows how much this is going to cost them to go and do the job. It's not like anybody really knows the exact price of all this stuff. They're all kind of, you know, throwing a board throwing a dart at a dartboard. And the value that we can provide as a services provider is at least we can make the dartboard really, really clear. Right? And make it so that everybody knows the dartboard. So wherever you throw the dart, you kinda know what you're gonna be fighting through in the future. but all of this is why I'm like so convinced that we're like not taking anybody's jobs because I just see people make these decisions and I'm just like, my gosh, are you kidding me? This is insane. like, dude, I like I'm like working with a contractor in a major city, major US city, who is changing how they buy electricians because they're pretty convinced that some of the electrical suppliers in this major city are like buying wire from the mafia, from like the Russian mafia. And so they're like, there are these certain prices that are going down, and we don't, we're trying to decide if we trust them. I'm literally, I'm completely serious about this. Like, this is literally a conversation that we've had. And it's like, how the hell do you systematize that? Like what are you y what are y you know, this is not a this is not a you know, we this is not a a a a fig jam flow chart. It's like it's just like total chaos.
Luigi La Corte: Well, I I do have a solution for that. You have a skill that aggregates choices made by the PM and you reflect that on a project by project basis. Cause usually PMs do like, from what we've seen, they like to mirror their decisions project by project, and they have these preferences. Like we've seen like
Josh Whisenant: Well they certainly yeah, for sure. yeah.
Luigi La Corte: one PM wants to give their carpenter literally as much scope as possible. I think this was an anecdote that you gave me
Josh Whisenant: Yeah, I was gonna say.
Luigi La Corte: the the other day. but I see some other examples where for for their for their Mason, they just have preferences for how they do it in this jurisdiction. But if you go into this asset class, it completely changes. Who's doing the the the
Josh Whisenant: That's right.
Luigi La Corte: the caulking or who's doing like it all changes? There's so many individual preferences, but these things like these
Josh Whisenant: it's exactly.
Luigi La Corte: these fuzzy things can be captured with with text. Like in this situation, this over time, right? Like it it's
Josh Whisenant: I
Luigi La Corte: hard to capture on day one, but it's I think you can capture it over time because You know, it's a lives in a human's brain.
Josh Whisenant: Luigi and yeah, no, I mean that's right. And you know, we we do do that. you know, we've we've done multiple iterations of trying to do that. We did we actually built like a very a ri a very rigid system to capture those decisions on projects that we ran on a few and it was just it was just too much work. It was it w you know, you need to just do it as a s as a skill file. and so that's kind of what we have started to do. There's sort of an interesting moment for us where it's like, all right, so we know how they did it on their last job. We still probably want to check again and make sure and talk to them about it, right? Just because they had some idiosyncratic way to give the carpenter everything, you know, rough carpentry everything, maybe they don't like that guy anymore. Maybe he sucked on their last project. Right. And so to me, to me, what it really is is you've got a a person who's gonna own the job, who's gonna manage this whole thing end to end. And this guy or girl is going to hate their life for many years because these jobs just suck. Always. And that person needs to be in the driver's seat and needs to understand this as well as possible. It needs to be able to make as authoritative and fast decisions as humanly can be done. And yeah, like if we can get them to that point, then we've we've broadly succeeded. And you know what you know what the highlight of Procor for me was? And this was really weird. we've done a we did a bad job getting feedback on some of our early jobs that we did like know, just o just under a year ago or something. we had five contractors come back to us at Procor and say, that job you did, that was like as good of a job as we've ever done. I I I I I didn't really bel I don't really I'm I I I I if I if you guys said that, I would be like, that's marketing copy. You know, screw that. Like these guys literally told me this and I don't I you know I think maybe they're trying to like justify their involvement with us. And so I don't know if I totally believe it, but You know, I do think that there's something to be said for just like getting a person that's usually really busy and stressed to make a bunch of decisions really fast and understand what's going on really fast and drive clarity as quickly as you humanly can. it's but still, like the hell
Mike Powers: Okay.
Josh Whisenant: if I know how jobs get bought. I just don't get it. I don't really understand.
Andrew Zukoski: On that note. Luigi, you're Luigi's muted.
Mike Powers: Luigi's on mute. Yeah, he's so typing.
Luigi La Corte: Sorry, go ahead. Go ahead. I've lost my opportunity.
Josh Whisenant: Now Andrew's on mute.
Andrew Zukoski: Yeah, sorry, my son's waking up in the background. There we go. On that note.
Luigi La Corte: Comedy bears.
Josh Whisenant: What? Are we done?
Andrew Zukoski: Yeah, I gotta go feed my son breakfast.
Luigi La Corte: Maybe if you drop off, Andrew, we'll we'll wrap up. if you gotta
Andrew Zukoski: See you guys.
Luigi La Corte: go.
Mike Powers: See Andrew.
Josh Whisenant: So he's gone.
Luigi La Corte: Josh, where like I think the services are I the services are great. I love the stories of people saying it's the best job. Like that's as good of an endorsement as you can get, right? 'Cause like those are the the opposite of that is what you obviously want to avoid, right? Like the subcontractor issues, no one knowing what they're doing, the coordination issues that hurt profit. Do you do you see a world where Rogers used on everything, on every project? And if not, where does it make the most sense? Like if if someone's listening to this, when should they say, Okay, I should I should work with Josh and Roger?
Josh Whisenant: We're definitely not gonna be I mean, you know, the twenty year vision of, you know, Procore has shipped like thousands and thousands of tools. I wanna ship, you know, these operating systems of the the technically correct way to run a construction project, and we'll have all of these different services in 20 years of how you could do this. And so I hope that in some capacity we'll be a fit on all these jobs. you know, we didn't even talk about our submittals service line, but you know, that has a very specific those have specific profile jobs where we are a fit, and we there are some where we're not a fit.
Luigi La Corte: Give give us the give us the strap
Josh Whisenant: although actu
Luigi La Corte: give us the strap line on on the submittal services 'cause this is part of the this is part of the release as well.
Josh Whisenant: Yeah, like Propor announces a submittal agent. The problem is you've got your most junior, most stressed, most unhappy people in construction running all of your submittals on jobs. And it's just like completely overwhelming, right? So it's like, hey, put us in your workflows. We're gonna review your submittals for you. We'll keep your project engineer on the loop. They still like, you know, they're actually gonna learn the job better because they can like look at our summaries and read everything. We do shop drawings, although we still have some iterations to do to really get that like perfect, but we're getting really pretty good at it. We've built a lot of technology there. And so yeah, we've seen cycle time in submittals be cut by like literally 30 days. Revise and resubmit rates cut by half. It's awesome. It's crazy good. and like these project engineers are just like, my gosh, you saved my life. Like I like my job all of a sudden. Don't don't ask how my team feels about their lives. Like, you know, we it's like actually like really painful, unpleasant work, but it's been really good. But the point is, you know, so submittals where it's been the best actually is data centers because they don't really coordinate all this stuff up front. They're trying to go so fast and they do a lot of coordination and figuring the job out in the submittal process. So you have this massive volume of submittals on these big projects. And we just like, you know, we just get hammered by it all day. And
Luigi La Corte: Mm.
Josh Whisenant: so for the buyout line, I think that we are overkill on jobs under 10 million bucks, for example. Almost always. Unless it's a really complex job and somebody really needs the help. You know, there kind of needs to be the i I also think that if you can really get a project manager. Into pre-con in a very involved way, which happens less than any GC would like to advertise that they do. If you can really get a PM that understands it in say like an IPD IFOA contract, or like some really nuanced iterative buyout scheme, or if it's industrial without a lot of complexity, I'm not convinced we're a fit on those jobs. to me, where we are a fit is when risk shifts quickly from the owner to the contractor, where it's like we're really building this job, there's a serious amount of complexity on the project. With lots of interfaces and coordination, vertical construction, and there's a schedule that you need to hit and you need to move quickly. And then we seem to we seem to basically bat a hundred out of a hundred on those
Luigi La Corte: Mm.
Josh Whisenant: projects. But on an industrial, like on a big industrial warehouse where you know you're starting to do buyout on 30% DD based on the last four years and it's not that price sensitive, I think we're like zero out of a hundred on that. Yeah, we just haven't been a fit. Now people love us for submittals there. So I don't know. I have a lot of humility about where we fit, and I do, you know, I I turn away I turn away a reasonable number of projects. I I be you know, 'cause 'cause we can't bill. I'm not gonna bill you if we suck. If it doesn't help you, I'm not gonna bill you. And so then we
Luigi La Corte: Mm.
Josh Whisenant: go and we lose all this money. Like, I don't know how much money I've lost on like doing no fit projects this year, but I've lost at least seventy thousand dollars of revenue.
Luigi La Corte: And is that because you don't think the quality's good, so you don't bill?
Josh Whisenant: No, the quality's great. It's just that the quality the the output it the d you know, we have a widget machine that produces a widget and runs a process and the process wasn't what that job needed.
Luigi La Corte: I see. I see.
Josh Whisenant: And so they don't use it. And we can see it in the usage. We can see it in the metrics. We can get the feedback from the customer. We can see it how they don't respond to our emails. You know, the whole vibe is there. And so you send them a note where you're like, this clearly wasn't a fit for you. we'd love to do another one with you. We learned some lessons here. Here's what we learned. you know, the invoice is voided.
Luigi La Corte: Right. Yeah, I'm sure that's appreciated. Josh has been great.
Josh Whisenant: Sh not as much as they should be. Nobody ever gives me any credit for it. They're always like, All right, thanks, see ya.
Luigi La Corte: That's crazy. Well, this is great.
Josh Whisenant: Yeah.
Luigi La Corte: I'm I'm bullish. I think there's a a definite need, obviously. And augmenting the team's capacity with a human is great. It's like people have a lot of tools. That is true. So how do you escape that? And I think this makes a lot of sense. Congratulations on the lot on the launch, Josh. Let us know how it goes. And like you're you're our first second time returner to the CTP.
Josh Whisenant: I'm s I'm super honored to be on the scene. You guys, I mean, I actually think this is pretty fun. I it's
Luigi La Corte: It is fun.
Mike Powers: Yeah.
