Opening the Palantir Black Box - Cavtera

Robert: Welcome, welcome, welcome everyone to the Con Tech Exec Podcast. My name is Robert Zimmerman and today I'm here with Ryan and Joe from CavTech. s Cav actually Cavterra. Sorry guys. Apologize for that.

Ryan: All good.

Robert: May have to edit that one out. But anyway. We're here chatting that you guys just got off the sales call. Folks, today we we've got something special. and I've been waiting for a long time to to talk with you guys about Palantir and Foundry. And I had I had discussed trying to work with you know, Kavanaugh Industries a long time ago. And then suddenly we you know, I'm no longer at this place and you're no longer at one place and now we're kinda all here what together at And talking about Palantir, which seems like a bit of a black box for some folks. Everybody just seems a bit of a mystery. so I've got the experts here today, folks. And so Ryan, Joe, I'm gonna just ask I d whoever wants to take this one first, but what is Foundry? What is Palantir Foundry?

Joe Patrois: Yeah, we think of it like an ecosystem where you can now, you know, organize or assimilate all your your data, which, you know, data sounds kind of like a fancy word or everybody gets confused by data, but it's, you organize the people, the places, the things. And then the verbs, your business, the actions, the things you do every day in one place, and then build applications on top of them. And so, you know, everybody asks what it is, but it's actually just a tool that allows you to do whatever you need to do. And so everyone says, is it an ERP or is it a this or is it a that? And it's like the answer is actually it's it can be all of those things if you want it to be.

Robert: Okay. So it's the it's the genie in the bottle? Is that you know we kinda ask it to give it three wishes? Ryan, what do you think?

Joe Patrois: Well, we kind of explain it like it's like a Lego set, right? You can build whatever you need to build. It's got all the tools

Robert: All right.

Ryan: Yeah.

Joe Patrois: and things inside of it, yeah.

Ryan: Yeah, it's you know, it it really is it's a it's an ecosystem. You know, it allows you to you know, to ingest, to integrate and and to really build something that, you know, especially in our industry, you know, something that we really haven't been able to imagine before. You know, it's It's on the other end of the spectrum, you know, from this kind of siloed ecosystem where you have individual solutions solving individual problems. So this is a cohesive way. It's the glue, if you will, to bring everything together. And you can do that to the, you know, to as narrow or as broadly as as as really you choose. And that's one one of the really, really big powers or one of the really big testaments to the to the platform.

Robert: Okay. So it's it's an ecosystem that you have. There's a word that gets thrown around and he was just on the other day. I can't remember what show he was on, but he was talking about ontology.

Ryan: Yeah.

Robert: so again, this kind of stuff to you know, regular construction guys, when you say a word like ontology, you feel like you're at the hospital or something. It's like

Ryan: Yeah.

Joe Patrois: Right. I kind of touched on it in my you know, I jumped the gun a bit on the earlier example, but you know, Palantir explains the ontology as the nouns and verbs of your business. And so,

Robert: Okay.

Joe Patrois: you know, that sounds a little abstract, but they're actually right on, you know, so nouns being the person, place or thing, and then the verbs being actions, right? And so The nouns and verbs of your business are really just the people, places, and things, and those become objects digitized in your ontology. And so though it sounds like data or abstract, it's like you gotta manage your people, your equipment, your materials, the work, and then you do actions on top of those. So schedule people, schedule work, you know, move equipment back and forth. Those are the actions of your business. So really ontology is is just the nouns and verbs of your business. But again, in construction, that's the work and the resources.

Robert: So it ontology basically just says, This is what you do. Here's the people that do it and these are the things that they do. Okay. So that that's all ontology is saying is like we're just it's a kind of a big word to say, This is your business. It's sh it's it's kind of this is your business in in one word. It's you have an ontology of some sort. Okay. All right.

Joe Patrois: Yep. It's the digital representation of your business. Again, sounds abstract, but it's you know, what is a business? You do work with

Robert: Yeah.

Joe Patrois: resources.

Robert: Well, okay, so now here's one of my pet peeve guys, and I gotta tell ya, it's one thing that has bugged me about technology and construction for so long is that technology uses these big words and it scares the hell out of most construction companies, scares the hell out of me because sometimes when I don't understand it I'm just like, Okay, wait a minute. Why did we use that word? Why are we tr I mean, we're supposed to be, you know, kind of connecting. And then when you throw out something like that, if you're a technologist and you're a you know, you're trying to sell some software and you throw out the word ontology, it just sounds like you're trying to sound smart.

Ryan: I you know what, it's I'm glad you brought that up and there is this black box effect. And I think it's really important, you know, especially for for your viewers to understand, you know Joe and I are construction guys. We're not we're not computer scientists. We're not we're not technology guys. Like my my my whole career has been in mining and heavy civil. And you know, the the biggest barrier is I guess just getting the courage to open up that box. And once you get over the terminology and this kind of black box effect, and when you start to really practically see what it is, It's it allows you to just f visually like visualize your business in a way that you haven't, you know, been able to before. So they have this view that you actually can look at your ontology and look at what you've built in the platform and I can see when a time card goes in and when it goes to the supervisor application and it gets approved and that starts feeding job costing, I can physically, you know, hit hit a hit some hotkeys and look at that. You know, it's not code, it's something I can look at. And then early on in our journey, that was something that was really important. for me to, you know, as as someone that isn't a computer scientist, to be able to visualize what the developers were doing and how they were describing and tackling these problems and I could participate in a way that just hasn't been possible. You know, looking at 10 pages of code wasn't wasn't going to be something that was going to be useful to me.

Joe Patrois: think one thing we do differently too is, you know, we don't start with data or technology or any of that stuff. We start with what does your business do? It's right, it's like you do work with resources or you build houses or you produce asphalt or you manufacture things. It's what do you do and what are you trying to do? And then we build tools around the best way to work. Palantir allows us to solve, you know, the business problems unencumbered by software is what we like to say. It's it's not about the software implementation, it's about solving the problems we need to do with the best tool.

Robert: See y you brought up data. you know, l lots of talk in the industry. There's the industry talk of you need to have clean data. I've

Joe Patrois: Okay.

Robert: said it myself, you know, garbage in kind of garbage out. You can't be using that there's that ninety percent number that gets thrown around of data that goes unused, that kind of thing. When I talked to you guys, we we kinda touched on the idea of like data entered once, right? We talked

Joe Patrois: Yes.

Robert: about and I've, you know, I've talked to several companies about this same thing. They're all looking for that same path. They're like, I don't see why I get it from here and then I've got some manual translation I've got to do and some Excel file and then I've got to get some data over here. Is is that what we're talking about and when we start talking about ontology and digitizing everything your business does is kind of creating that atmosphere?

Joe Patrois: You're you're totally right. And you know, we see it as like the saying we like to use is everybody's thought data was the product, but the way we view it is data is the byproduct of your operations. And so historically, if you wanted to understand your data or, you know, analyze your data, somebody had to enter it in. s you finished your day and then you type something in. But the way we approach it is just by doing the things you need to do, which is submit a time card, order a truck, schedule work. The data is the byproduct of doing that. So when you build in the ontology, and that's actually why Foundry is a big differentiator from the way we used to do it is now just by doing what you need to do day to day, the data is a byproduct. And then you've got all the data you know, and I hate the word data, frankly. You've got all the the transactions of your day in one place that you can now analyze any way you want. And nobody had to do any extra work to generate it.

Robert: Is data just one of those words now that's become overused and everybody just we kind of now we just glance over it because it's you know, it's like when people say racism

Ryan: Well well it almost is, yeah.

Joe Patrois: It's the right word, but it's been, you know, overused and so it's lost its meaning. Like data is the right word, but you, you know, it it's meaningless to just look at data. Like the way I think of data is nobody needs more data on how to lose weight, right? You actually just have to do the thing, right? Like nobody needs to read another book on how to get six packs abs. It's just like move more, eat less. Everyone knows that. And so data is the same thing. We have tons of data. We generate data all day long, you know, we just have to generated in a way we can use it.

Robert: I see.

Ryan: Yeah, and I think I I I forget there w it was maybe during it was I forget if it was during a sales call or or what, but someone came up with a good analogy and this is this is great, you know, being up here in Canada. It's called it's like passing the puck. You don't wanna have to recreate the puck. You wanna whether it's your your your dispatch order that you made a request for or your timesheet you put in You want that information to flow seamlessly, not be recreated at every single interaction. You know, our an average, I think, heavy civil PM or heavy civil project manager is going into 14 different platforms, whether it's their their spreadsheet or their ERP or their time tracking or their safety. And a big piece of what they're doing is just reproducing information that already exists in another platform. So you're tying up these highly skilled and experienced individuals' time. essentially doing administration when they could be put to a higher best use. So it's that's kind of where the the pass the puck analogy comes in. Like that thing exists. You want it to flow seamlessly between your workflows throughout the day.

Joe Patrois: We recognize this, you know, from operations. It always killed me when we were out in the field that, you know, at the end of the day you'd have to enter in how many trucks you used, but it's like you already ordered the trucks. We know that. And so the whole idea of passing that puck up the ice and you know generating the data from the things you do day to day came from, you know, living in that world and the frustration that comes with, I just ordered the trucks, you're telling me I have to enter that all back in at the end of the day. It's like you know most of that. All I really want to enter is what change.

Robert: Yeah. That was the that's always the biggest pushback you get, right? When you get the the guys out in the field, especially 'cause they I think that's kind of that baseline data, you know. again, to to your point, Joe, it it is a by product, but nobody's really writing if it's not getting written down, you don't get the by product, right? Well, the problem is, and this is what everybody runs into, is like, well, this stuff is supposed to be making my life easier, right? But it doesn't. This is the biggest complaint about current technology is hey, this is supposed to be making my life easier, but it seems to me you're just adding more stuff onto my plate.

Joe Patrois: That's why, you know, we're so passionate about putting your entire business in foundry because you know, historically every new problem you're solving, so if you have a scheduling problem, you gotta get a new schedule software. If you have a time card problem, you need a time card application. Historically you've been adding all these softwares up. So every solution required one more. This is the only point in time now where the solution you can now start eliminating them and make it simpler. This is now the last software you can have only one at this point now. So it's that's the thing that's changed.

Robert: all right, so probably gonna get a lot of pushback on that. And I and this is an interesting topic because you've got the ProCores out there, you've got the Revit's out there, you've got all this stuff. You've you've bought massive amount I mean millions of dollars, right, in tech stack. and now, you know, Palantir comes along and says Yeah, get rid of that. Now, is that true? Or does foundries sit above it all? I ha how does that work? I mean, because that seems like it's a big, big mountain to climb.

Ryan: Yeah, I mean I I think it's it's gonna be a combination. Like I th I don't think anyone's under the impression that the the whole industry is is able to change overnight. because there's you know, parts of your business are gonna have to change with you know, with as you tr as you start to generate some of these solutions. So I think the one really important element is you can begin by tackling, you know, a really specific pain point. You know, like we see a lot of customers going after their labor force management or equipment management and addr and address a real business problem. but then you can build off of those solutions and you can actually deprecate other standalone solutions, but you don't have to start by doing everything at once. So it can be a combination of integration integrating with the solutions that integrate well. You know, like maybe ProCore is an awesome solution and and and that customer is really happy with it. So maybe you're integrating with it and building new features or capabilities that are standalone. So you can do a combination of, you know, assimilate your ecosystem over time. It doesn't have to be all at once because there's also an important piece of of kind of change management. Because at some point you bought a solution and you've you've now bent your business to fit how that particular SAS solution works. So there's there's some work to undo that. It I don't want we don't want to present it as you know, it's all it's all rainbows and rosy. There's some there's some real hard the real work is is is is actually the change management. It's it's you now you're unencumbered, you don't have any excuses. How do you actually want to run your business?

Robert: So that's always been the thing, right? plenty of tech solutions out there, but a lot of people are just skeptical of the change management process. some number out there that I had read, you know, been a couple of years ago now, but you know, some seventy percent of implementations fail. is because they hand it to the wrong person. Maybe it just it just works for a couple of weeks and then they're like, it's just it's just not doing it. Right. let's let's talk a little bit about that and then what you know, how how does Foundry approach that as as far as in how did they get past that? I mean, change management is a huge I mean, again, it's kinda like the word data. y it's a it's been used a lot, but what does it really mean? What it I mean, you can say change management and it sounds easy, but it is not easy at all. Like you can say data but it doesn't mean that data's hard. Data's difficult.

Joe Patrois: Yeah, I think people confuse, you know, buying a software with that organizational change and it's you know, in the case of Foundry, it allows you to solve these problems if if you are up to the challenge, right? Like just hitting the buy button doesn't magically solve all the problems. It now allows you to. and it's you know, it's it's It's a total game changer, but there's still that element of change management and Ryan can speak to, you know, what we did to make the thing stick in the field and, you know, how we approach that change management.

Ryan: Yeah, and I think you know, I think what you're referring to, Robert, there with the with the stat that you know, a lot of a lot of tech initiatives do fail. And and often it's not because the the technology itself didn't perform the function it was supposed to, but you know, generating the the the right top-down kind of buy-in, you know, does the arg organization understand, you know, do they understand the task? when you start surfacing this this this new information or this newfound information, you're integrating things, you actually have better visibility. When does my when are people starting and finishing? When are they taking breaks? How often is my equipment idling? What kind of utilization I getting? Do you have people in actually in in position to act on that and make a different business decision? You know, historically, if this stuff has been obscured, there might not be someone there. You know, you're you're you're you've dialed the number and no one's on the other side of the phone to say, hey, you know what, this is a problem. We actually need to operate differently. You know, and sometimes those decisions aren't aren't easy to make. It's When things become un become unobscured, sometimes there's there's there's some tough decisions to make. You know, maybe maybe some people that shouldn't be there within your business don't need to be there anymore. or may you maybe you need a different type of talent that's that didn't exist. so it's you know it's it's a it's a full pronged approach. I think and it and it v and it varies. Once again, it's just because you can go so narrow and solve something specific. But as you get broad, as as we done at Thomas Kavanaugh Construction, which is our anchor client now, there's there's a lot of complexity and there's a lot of hard work to do and and I would say a big piece of it lies outside of of of the technology entirely.

Robert: Yeah,

Joe Patrois: I think a tangible example could help too, which is

Robert: well let's let's g yeah, let's get it get it let's get into that, Joe. because I really again, this is something that everybody knows there's gonna be tough questions to ask, but th this is something you guys have gone through and there were some hard conversations that had to be had. Let's let's discuss that a little bit.

Joe Patrois: Yeah, and I think a simple example is, you know, we have Foundry now, we built out a time carding application, a field time, you know, review or a field time capture application for those time cards to allow now land in. We got to the point where we're like, okay, now we want to automate some of that manual entry that the foreman and the folks filling out the tools are running up against. It's like, okay, let's automate the payroll policy. Okay, night shift overtime, all that fun stuff. Well, what we found out quickly is like do we actually have a an agreed upon payroll policy? So it just forces you to say, Okay, well we gotta get the whole group together, agree on what a night shift is. Not that anybody ever really disagreed, we just never, you know, sat down and really hammered out 'cause in order to code it, we have to know With it without a shadow of a doubt what a night shift is, right? And so and it forces you to kind of go through this simplification process too, because the more complex it is as a policy, the more complex the code is. And so you get this cool forcing function where you get to really evaluate how simple can we make this thing and still get all the benefit, s but you know, speed up the process.

Robert: Right. Well, I always found when you get into that kind of solutions engineering stage, you know, I had th punk plenty of people there throwing in all the edge cases. We used to call Yeah. Yeah, it's like, well, what about this? And what about this? And it might happen two percent of the time or might happen ten percent of the time. What do you throw in? What do you leave out? is the question. Is there a workaround on that?

Joe Patrois: way we've approached it is you solve for the eighty percent and then we'll handle the edge cases. But you definitely don't want to make the exception the rule because that's like that's usually the reason why you don't do something. And that's usually the reason why you end up in the place you end up is because you've allowed those exceptions to be the rule for all these years.

Robert: Yeah. I well and I've heard it said before, you know, eighty percent is better th eighty percent done is better than zero percent perfect. Something like that. I was thinking

Ryan: Yeah.

Joe Patrois: Well, that's a good one. Well, and funny enough, you know, even a lot of those exceptions, as you pull the string or the thread far enough, you f you realize that's actually not the problem and some other issues, the thing that's actually blocking it, you know. And so I don't know. I I won't go into detail on that one. But there's just you you uncover organizational stuff that

Robert: Wait a minute, I think we touched a nerve. We touched a nerve over

Ryan: Yeah.

Joe Patrois: Yeah, well there's

Robert: there.

Joe Patrois: you know, it's not that this tool won't work for one part of the organization. It's that that one port of the part of the organization needs to actually fix how it operates a little bit. So you uncover a lot of that stuff.

Robert: There you have it. Yeah, there you have it. And Ryan, I I think you understand this just as well. I think we've ran into this as as many times in in our past conversation. We talked about this is that A lot of times and I feel like the age of AI, the ability to utilize these new tools is gonna make us discover more about what we need to change in our operations than what technology needs to do to fi try to fix you. You need to fix yourself.

Joe Patrois: Yeah, and our actual you know, our value proposition as Cabtera is we say we help, you know, businesses untangle the legacy systems and processes and then build the tools around the best way to work using Palantir Foundry.

Ryan: And I think it's you know, y you brought up AI and like that's that's the holy grail, that's what everyone's talking about, you know, that's the you know, the stuff that's floating around in kind of the macro economy. right now, I think a lot of businesses have realized this, and this isn't even limited to construction. You're you're you're moving in one of two directions. You're you're increasing complexity through, you know, natural growth and expansion with your business or acquisitions, and you're you're either adding more pieces of technology into your stack or Or you've made the conscious decision that you want to go the other way and that's something that Foundry allows you to do is you gotta start unwinding it back. I mean, large, large the kind of multinational, you know, builders. It's we know several and you know it's not uncommon for them to have three or four hundred pieces of technology in their tech stack, you know, because they bought company A B C in this jurisdiction and they ended up with their time carding platform and their dispatch platform and scheduling and they never they never adopted whatever the the parent company was doing. So you you can imagine what it's like to try and now integrate all of that w for some future where you want to actually make I operational. Well if AI AI can't see something if it's not integrated. You know whatever LLM you're going use, the better they get, great. If the data's in a silo somewhere, it's actually it's actually use it's actually useless. It doesn't matter how great AI gets until we until you start creating the single source of truth. And you gotta start winding it back. And you know that you can do that through one win at a time and you can justify the you know the spend of kind of one win at a time, but this this bigger picture, this North Star, becomes available to you. and I think that's that's the that's one of the real powers of of Foundry.

Robert: So y this gets back a little bit to what I was talking about, you know, sitting above like everything, like your pro cores, your your your kind of tech stack. you need to you need to tackle it kind of one at a time, but I wanna talk about 'cause you mentioned you know, legacy technology and it God knows that, you know, ERPs across the board are pretty much legacy tech. You know, at one point ERPs were supposed to be the single source of truth for everything, and then BIM was supposed to be the single source of truth for everything and now we're gonna introduce, you know, large language models or whatever we're gonna intro you know, whatever we're gonna do here. this I It it seems like a lot to juggle. i specifically on the ERP level. That gets very, very scary for folks. I mean, you know, how does how does Palantir kind of appropriate how does Foundry kind of approach that? You w do you wanna recreate your ERP? That seems like a lot.

Joe Patrois: We think of it as we're deoperationalizing our ERP system because I think that's one of the you know the things that we've done over the years, which is you know the inverse is operationalizing our ERP, meaning we've tried to force the ERP into operations and we all know, you know, ERPs aren't great at scheduling work or planning resources or doing any of that stuff operationally in the field, in the tools, you know. Same with scheduling software, it's like nobody's ever successfully run a job off of P6 schedule day to day, there's all these shadow tools, Excel sheets and things like that. And so what you end up is not only did you not have the source of truth, you now have all these shadow tools serving all the functions that those classic systems weren't able to do. And so you end up in a worse spot. Yeah, operationalized ERP system has l led to where we are.

Robert: Well I th and always remember those shadow tools are being run by a C student who didn't pass algebra.

Ryan: That's right. That's right. And left the company ten years ago. And nobody knows how that how nobody knows how their their sheet works, but

Robert: Yeah. Yeah. How they how are they w how were they working that? I'm not sure.

Joe Patrois: That's a true story. Yeah.

Ryan: Well and it and it's funny, you know, we're working with a group right now and it's we're at this point where you know, some of the the late you know, our industry is has of has long had the the label of being laggards in adoption. Now, you know, you're actually seeing some of them now come to the table and they're actually unencumbered by either not like having having very little tech in their tech stack, like they're still paper based. And then for for those companies to take a leap and now go in all in with something like Foundry, in many cases, they actually have an easier time because you're simply digitizing directly.

Robert: You're not unwinding a bunch of stuff.

Ryan: You're not unwinding, you know, some some you know you're not unwinding their stack. They're actually just going directly now into you know this this all-solving solution. so in many cases they're able to kind of leapfrog over maybe some of the other, you know, early adopting companies, which is Which is something interesting. I wasn't me it wasn't really it wasn't really on my radar when we kinda took when we went to market with our solutions, but you know, you're s you're seeing them kind of skip ahead, skip ahead of the line in their in their progress.

Joe Patrois: It's actually an advantage for some of those companies and the analogy we use is that, you know, they didn't bury all the phone lines and have all the sunk costs in the infrastructure. They just went straight to cell phones. So you can make that leap.

Ryan: Yeah.

Robert: Still need that fiber though. We gotta have that fiber. Let's

Ryan: That's right.

Robert: let's be clear. We gotta have that fiber there. I you know, I find it interesting to talk to you guys about this because you guys have made the transition. You Thomas Kavanaugh decided that this was the way we were gonna go. Like I say, I had talked to you a couple of years ago, you know, when I was still working at another company and talked to you about, you know, like we were talking about time tracking solutions and and things like that. And it was you know, there's a lot of frustration around all these kind of peripheral you know, technologies that you have to have that to even integrate into your ERP and it seemed like there's a and so you guys just you guys just walk the walk. And now you're you know, you're not just talking the talk, you're you're walking the walk. You did the whole thing. So I kinda wanna c dive into that just a little bit to give these to give us a little bit of background. It's like, what was it? Was it I know it wasn't one thing, but what was the straw that just said we've got to s stop the madness here and go with this? Like it can you kinda give

Ryan: I

Robert: me an idea?

Ryan: Yeah. Go ahead, Bill.

Joe Patrois: Yeah, it was it was actually organic with every you know, 'cause originally we did start the same way everybody else starts with all those integrations. So we you know, we started with an initial use case with Palantir and it was real time costing. And so we started stitching together some of our traditional tools, our ERP system with our field time capture. But we really we realized really quickly that You know, though we could ingest all the data, you're beholden to the existing software schema if you want to read and write. And so what you find yourself doing is we ended up we called it breaking these tools or we were running them at 110% capacity, meaning, okay, well, if we want costing on our truck utilization, well, we need a column or a mechanism to capture that data. And so when you when you talk to these other softwares, you say, Can I have a column to enter in my trucks? And they say, No, we can't do that. And so

Robert: Yeah. And that's gonna cost you.

Joe Patrois: Yeah, and so it's either three to six months later or just a no in general. Yeah. And so we ended up doing things like you'd

Ryan: cost you or not possible, you know, or there isn't a price, yeah.

Joe Patrois: be you'd be tucking your truck numbers in the notes section and then using the description section for your notes and then you what are we doing? This you know, we kinda came to this realization that, well, what if we just built it in foundry? And so after doing

Ryan: Yeah.

Joe Patrois: that three or four times, we said, Yeah, what if we just built it all in foundry? And so that was, you know, part of our proposal to Kavanaugh was, hey, let's just do this all in foundry and it'll be easier.

Ryan: And I think you know and to be fair, you know, we've got we've got a lot of stuff across the line, you know, at at the at at Thomas Kavanaugh construction, but we're very much still active. You know, some some rocks are easier to overturn than others. And you know, we're definitely feeling some pain in some areas where we, you know, we had lots of success in others. but I think the You I think the way you need to think about this is you need to find some some pain points where you can derive some really clear value. But ultimately this needs to be something that you know the organization truly views as a North Star. Like they have a strong vision around this. It's communicated, it's it's bought in, you know, ownership believes in it. You know, you get the right stakeholders believing it. It's not something that's gonna be done lightly, and it's you know, and those are the companies you're seeing step forward, and there's a lot of them saying Hey, we're ready, we're ready to take charge of this. We're not gonna passively let AI decide or let you know a particular solution, you know, choose our destiny for us. You gotta be very deliberate in deciding to do that and supporting that, you know, at all levels. It's not just a technology or an HR or you change management. You you kind of have to take this holistic view because it's you know it's it's it's a big challenge. But the the gains are huge when when you can when you can cross that threshold. You know, the the the the treasure is sweet on the other side if you can get there.

Robert: Yeah. And I I you know, again, it's a big risk, when you talk about doing this because, you know, there is the potential of failure, right? It's

Joe Patrois: All change is risk. Yeah, change is expensive and risky.

Robert: Yeah, change is expensive in racing. This is what most people are looking at. Here's when I you know talk to clients and talk to folks about this kind of stuff, my o my question is always this, Well w what do you want it to look like five years from now? Do you want it to be in the same space that you are today? And, you know, a lot of them are they I don't sometimes they don't they they just go, well you know, they just nod their head or th but they don't really give you an answer because they can't really answer the question. They can't really say what and the way that I'm seeing things, the way that you know, that from what I'm reading and from what the information I'm gathering, it's like there's a lot that's going to be changing here in the next five years that we really in the construction industry, in any business whatever, we're really gonna have to start thinking about different ways of approaching these we're gonna have to sometimes we're gonna have to face our demons, right? And I that gets to the the whole talk about no matter what you do, and it wouldn't matter if if, you know you're out there, you decide to go with Caf Terry and you decide to go with Palantir Fallonter. It doesn't matter if you're gonna use Procore's new solution or try to use their A whatever you're going to try to do. Folks, I want you to understand and and these guys are probably here to testify about it, you are gonna have to face those demons that are in front of you. You're gonna have to face those bottlenecks. The things that are causing you problems today, doesn't matter what you choose, they're still gonna face you tomorrow if you don't get in front of

Ryan: No, and and building off of that, you know, there's there's a c there's a cost to doing there's a cost to inaction. You know, there's there's risk to doing nothing. You know, the you know, the the you know, the wait and see crowd, you know, the the risk is what if if your competitor figures it out, you know, first, you know, that's you know you you know, the cost is maybe just less visible or it's spread out within your organization, you know, all of these fractional losses or these fractional inefficiencies. And it's easier to point out, you know, when you go to spend money, you know, that's a singular thing you can look at and say it's gonna this is what it's gonna cost me. But there's a there's a real cost to just to sitting on the sidelines and just y seeing how this whole thing is gonna pan out. You know, so yeah, it's it's

Robert: I think and I think that's a lot of times I think we all know it. I know even my little biddy, you know, my little construction business, you know, I spend the money somewhere on, you know, the things that it it's spent somewhere. It might be s in soft costs, time here, time there. The money gets spent somehow. It's just how long do you are you willing to sp spend it? You know, or would you

Ryan: Yeah.

Robert: rather, you know, bite the bullet for a little bit, you know, really attack what you need to go and do and and and get it done. So we're not a we're not eliminating any bottlenecks here, folks. What we're doing is we're just trying to find you a better way to get through them.

Ryan: Yeah, and I think you know, it's not a perfect analogy. but one like one thing I like to think about is, you know, what what's what's the ROI of having the internet? You know, when the when the internet was new and it was coming about, or you know, point your finger at the one place where you where you where you got the return on that. Like if we can all we can't imagine a world where that's you know, where the internet doesn't exist and your business isn't doing you know, you're not doing business off you know, offline. You know, but it's it's It's probably that transformational, if not more transformational, which also, you know, it creates a this whole other set of challenges because they're it's so inter it becomes you know, a fabric or interwoven within within your business. You know, so you that's why I was saying you kind of have to pick off the things that are obvious, but there's gonna be this, you know, the sum's gonna be greater than than the parts. And that that part of it's gonna be harder to measure. But it's you know, it's what does your what what price do you put on being able to operate your business the way you want to operate it? You know, or operate it more efficiently. That's that's

Robert: Yeah.

Ryan: what you're having to be aiming for.

Robert: S Purdue University just came out today and said that every student will have to take some sort of AI course. And they're they're the first university. I'm a boilermaker so I just had to throw that out there. still a shout

Ryan: Yeah.

Robert: out to Purdue University for you know, making this happen. I believe that this is an important step. I work with AI every day. I do everything I can to try to learn more about it and get better with it. Joe, I want your thoughts on that. We're we're going to in the construction industry, we're gonna have to start learning this, we're gonna have start using it. Your thoughts?

Joe Patrois: Yeah, I think I have a very construction worker take on it, which is, you know, lucky for us we don't actually have to build the L L Ms. We don't technically have to train the models. What we need to do is actually organize, you know Our businesses in a way that we can take advantage of it. And the simplest analogy I have is: could you imagine if your entire life ran through ChatGPT, right? Like most of us have logged in and asked the questions and how to do things for us, but could you imagine if every soccer practice and oil change and bank transaction and funeral and everything you can think of in your life was organized through Chat GPT, how great would it be at giving you insights, like, hey, you better leave in five minutes because you're gonna be late because traffic's heavy? You know, everybody can kind of grasp that analogy, and that's actually. what we're building here. Everybody likes to get caught up in you know in the AI itself or how do we do AI and it's like no no you you layer AI on top of the context of your business or your life and when you build it that way you're able to take full advantage of these LLMs.

Robert: Brian?

Ryan: Yeah, and I think you know, at at the field level, you know, one one thing I like to think about is you you would just ultimately you wanna be able to ask plain language questions, you know, about your business. You know, who's the subcontractor? How am I getting paid for this work? You know, what what quarry do I need to be pulling material from? Who's the supplier for this? when's my truck gonna be here that I ordered? You know, it's we it need until you run your business, you know, in in with a s within a single source of truth and you're operating in a in a unified way, it's really hard to answer those questions or, you know, it's and that's what the field needs those plain language, that plain language capability. They just need the answer. And these are answers that they're, you know, calling the office for, they're going into the back of the truck and flipping through the field binder for. And, you know, it's it's how changes don't get you know, made and updated in the, you know, in the in the root folder so they didn't see that change order or that change to the drawings. You know, so when you when you operate, you know, in in one cohesive platform, foundry or otherwise, that's kind of to me, that's one of the, you know, major milestones for our industry is arming the field with the kind of this this plain language interaction. And you know, and and then it that that kind of evolves into taking action like, hey, I need some more pipe. You know, send me more pipe. I just want I just want to see say that into my phone and I want the right supplier at the right unit price with the right delivery method, show up to the right place. And you know, right now that that's there's a lot of coordination just to facilitate that simple request and that you start thinking about what is what has Amazon done for us and what has Uber done for us. You know, how do we how do we turn those tools into into things that operate that way for construction.

Joe Patrois: I th Ryan just nailed it, which is and I'll I'll just add to it selfishly, is you know, we c we all know what it looks like when you have an application or a tool integrate seamlessly with reality. Sounds abstract again, but you know, when I think about we'd spend all this time planning and scheduling our construction work and then you'd walk out on site and it felt like a totally different world of reality. All the work you just did there wasn't actually what you're facing in the field, right? Like There's there's the realities of the site. It felt disconnected. I'm having a hard time explaining what I mean, but I think most construction workers will get it. Where the inverse of that is like you order an Uber, you watch the dot on the screen blink towards you, and then you look out the window and there he is. Right. And that's

Robert: Mm-hmm.

Joe Patrois: where you have a tool integrated seamlessly with reality. Whereas construction, we don't have that right now. And so what we're talking about is

Robert: Yeah.

Joe Patrois: building those tools that integrate, you know, the the digital side of our life with the physical side, and we haven't yet done that in construction. So Ryan nailed it.

Robert: Yeah, that's true. That it is true. I mean i I was at a meeting n not so long ago and you know, one of the supervisors was like, you know, Rob, can you just can you just make it so that it's easy? You know, can you can can

Ryan: Mm-hmm.

Robert: we just walk up, you know, it was really about kind of like a check in, check out kind of thing. It's like can we just w walk into the job site? I mean, and and and get checked in, checked out and we don't have to do anything, right? Because his analogy was well, he went to a baseball game and he had bought the ticket online and I don't know how it all worked, but I come come to find out it'd like a it was like a some high dollar system but you know, walked into this main stadium and th it just biofacialed you right there. It it di you just walked through. They you know, they did the scan on him and they he didn't even have to show a ticket, didn't have to do anything, it biofacialed him because of the way he bought his ticket. And it was like, can't why don't we just do that? And it was like, Well, you know, okay. I get where you but this is what, kinda to your point, like, we watch the dot and the car shows up. You know what I mean? It's reality based. This this works. Why can't we do it for construction? Why why does it seem so difficult?

Joe Patrois: We don't have a technology shortage. We have a you know an actual tool shortage. We don't have the tools in the field to do that. We have the technology.

Robert: Yeah, we do. I you know, a lot of times that the biggest complaint and I y you guys tell me about it, it's connectivity. It's like I don't get internet on my site, you know, or it's spotty, or I don't get it on my phone, like they're using this, you know, iPhone three, you know, whatever, and they it's not there for them, or or something like that. So I I feel like that's changing.

Joe Patrois: Mm-hmm. Starlink's been a game changer and that's actually that's a good example of like solving a different problem where people say, Well, you know, we don't have site connectivity, therefore we can't do pound tier. It's like, well, actually, let's just go solve the site connectivity problem instead of trying to make pound tier work offline or do something like that. It's like, why don't we just throw up a Starlink at every site or have, you know, a a pelican case with a, you know, a router in it that you can take in your truck or wherever? You know, it's like solve that problem. but so often we download all of the the friction around it to the you know, the tool itself you're trying to solve the problem with it. It's like, no, no, no, just pull the thread all the way

Robert: Right. Yeah.

Joe Patrois: back to solve the connectivity problem.

Robert: Well, I think it gets back to kind of this first principles s you know, kind of thing. We don't solve for we we look at it and it's like up till now we've had to have stuff that worked offline because we've had really no opportunity for connectivity. But Starlink, to your point, has kinda changed the game on that because now I can open that thing up and, you know, point it up at the sun or wherever I point it to and bam,

Ryan: Yeah.

Robert: you've got connectivity and it's I you know, I don't know what that is a month or you know what it would cost. doesn't matter because now what you've done is you've totally gotten rid of the whole excuse of no connectivity and you've gotten rid of the excuse of like real time, you know, analytics from from the from the job site.

Ryan: Yeah, and I think you know, our industry need like we need that flexibility because you're gonna have some companies that, you know, connectivity is gonna be the bottleneck. Or, hey, we don't have users that you know, only half our workforce uses a smartphone or we're we're we're slightly you know, our vertic our vertical integration is our business. We actually we only buy material, we don't supply ourselves. So there needs to be this element of of customization and, you know, this the solutions have to be bespoke for the business. And

Robert: Yeah.

Ryan: and that's why traditionally there hasn't been a one SAS to rule them all for construction, because construction is such a broad term and we always talk about it as one thing. But there's, you hundreds of different iterations of construction companies.

Robert: Mm-hmm. Yeah.

Ryan: And they all have different challenges and they and you know, they've got better at solving different companies have got better at solving, you know, certain parts of their business c compared to others. So they they may have different they may start their journey in a different way. And that and that's okay. Having that flexibility is important.

Robert: Yeah, and scale yeah. Scale is a big deal. folks, you know, it really is. I I run a very, very small carpentry company, you know, and and that's totally fine by me because I enjoy what I do. I don't necessarily need Starlink, you know? But

Ryan: Yeah.

Robert: when I'm if I'm working on a civil job or something like that and I've got twenty five miles to cover, and I need to be able to do this, that and the other. And somewhere in that twenty five miles I may not get connectivity or something like that, I'm gonna need it. Right? And so again Scale is everything. Scale multiplies all of the issues by, you know, whatever the the the size change is. It's e maybe it's a hundred X, maybe it's ten X, whatever it is, but scale makes a difference. And you gotta get back to the whole, hey, is this really our problem? I think that was kind of your your point, Joe, is like, is the technology really your problem or is it the fact that you just are too cheap to buy Starlink and get yourself hooked up?

Ryan: Yeah, yeah.

Joe Patrois: That's right. Yeah, usually, you know, a lot of the time it is. You gotta ask the five whys and the you know, really get down to the what are you trying to solve? 'Cause more often than not, you're treating symptoms and not the problem. And, you know, even ourselves, as we went all in on Foundry, you know, we'd get asked, Well, we just want cost reporting. What are you doing over there in dispatch? And so, you know, as you pull the thread, it's like, Well, if you want truck costs, the best way to get is to order your trucks through the system and so you end up over in places you didn't think as we we always just call it pull the thread all the way, but you know, It's funny, sometimes the way to solve the problem isn't the way you thought, but you gotta be open to pulling that thread all the way.

Robert: Yeah. Well it comes full circle, back to, you know, really understanding what your ontology is, right? Understanding what are the true nouns and verbs of your whole system. And to your point, you know, which string are you pulling when you talk about said noun or said verb? and where does that come from? Where does it really tie back to And it's it's been a fantastic journey here just today. we're gonna do a couple more of these episodes, folks. I'm so Joe and Ryan, thank you for your forty five minutes here today, just talking about kind of illuminating a a little bit about, you know, what really is Palantir Foundry. I'm hoping we shed some light on it. but in the next couple we're gonna dive a little bit deeper into the product and so I'm looking forward to it.

Joe Patrois: Yep, same.

Ryan: Yeah, thanks so much, Robert. Same here. It's it's always a pleasure and I've never been more hopeful, you know, for our for our industry and you know, we really wanna move the needle and we know we're not alone in that.

Robert: Yeah. Well I think you guys are doing a great job and I appreciate you guys being here with me today. But until next time, folks, we're gonna be talking about we're gonna be talking about Palantir some more. We're gonna dive a little bit deeper. I'm not gonna give you all the all the clues, but we're gonna do about three of these and and and really kind of dive a little bit deeper. But thanks guys, appreciate you

Joe Patrois: Thank you.

Robert: and we'll see y'all next time.

Ryan: Thank you.