Why clean data beats cool AI tools with Constellation Homebuilder Systems

HousingWire

On this episode of Power House, Zeb Lowe sits down with Paolo Benzan, VP of Data Strategy at Constellation HomeBuilder Systems and Rich Swier, founder of Raia, to discuss why the future of home building won't be determined by who adopts AI first, but by who builds the strongest data foundation.   The conversation explores the hidden work behind successful AI initiatives, from data quality and governance to cost control and security. Benzan and Swier explain why clean, connected data is the real competitive advantage and how builders can use AI to move beyond static reports and make faster, smarter decisions. Their message is clear: AI isn't replacing expertise — it amplifies organizations that have done the hard work to prepare for it.   Related to the episode:   ⁠Zeb Lowe’s LinkedIn⁠  https://www.linkedin.com/in/zebulon-lowe-a02353a4/  Paolo Benzan's LinkedIn  https://www.linkedin.com/in/paolobenzan/  Rich Swier's LinkedIn  https://www.linkedin.com/in/swier/  Constellation HomeBuilder Systems  https://www.constellationhb.com/   Want more from Zeb? Don’t forget to subscribe to LendingLife.  https://www.housingwire.com/newsletter/   The Power House podcast brings the biggest names in housing to answer hard-hitting questions about industry trends, operational and growth strategy, and leadership. Join HousingWire’s Zeb Lowe every Thursday morning for candid conversations with industry leaders to learn how they’re differentiating themselves from the competition. Hosted and produced by the HousingWire Content Studio.

Transcript

00:00Welcome back to Powerhouse. Before I got into mortgages, I built homes. Framing, trim, cabinets. So I can tell you

00:08the most important parts of a house are the ones you never see. The foundation, the plumbing, the wiring. And

00:14today's conversation is all about why AI works exactly the same way.

00:20I'm joined by two guests from Constellation Home Builder Systems, the largest provider of home building software in North America.

00:27Paolo Benzin is the Vice President of Data Strategy at Constellation, and Rich Sweer is the founder of Raya AI

00:34and has spent the better part of a decade building AI for businesses, long before everyone had an opinion about

00:40it.

00:40We get into the magical aha AI moment, that superpower feeling that you get when you realize what AI is

00:48truly capable of, why it can be dangerous, what happens when builders go DIY on AI, and why clean data

00:56matters more than cool tools.

01:10Rich, Paolo, thank you for joining me.

01:13Good to be here.

01:14Great to be here.

01:16All right, so before we get into AI, can you both give me like a 60-second version of how

01:20you ended up at Constellation?

01:22Rich, let's start with you, and then we can go on to Paolo.

01:25Yeah, I mean, it's a story on how a lot of entrepreneurs and software technology people end up at Constellation.

01:33I sold my last business to them about six years ago.

01:37My primary focus was predictive analytics in the housing market and sold software to builders and other real estate finance

01:46sectors.

01:47And then over the last three years, I've been primarily focused on building out AI agents that help businesses automate

01:56their workflows as well as their day-to-day operations, sales support, et cetera.

02:04Okay.

02:04And Paolo, you spent years as a production home builder before crossing over kind of into the software side.

02:10So what made you take that leap?

02:11So actually, I started in software, so I was trained as a computer scientist and started on the software side

02:16and then, you know, went into the business solution side as you do, as you grow in your career, right?

02:21And I ended up running IT organizations for production home builders for many years.

02:26And in the course of that, you know, one of the companies that I was working for used Constellation, used

02:33one of their products as their ERP.

02:35So I got to know Chris and his team really well.

02:38And then as we went through acquisitions and all different things, you know, I reached out to Chris at one

02:42point after one of the final acquisitions I was in and just said, hey, by the way, I've gone to

02:48all these builders and each builder is solving the same problem over and over again.

02:54Why don't we do something that is more industry-wide and do it from a Constellation perspective and provide it

02:59back to the builders?

03:00And that was the genesis, essentially, of the data services team that we stood up with in Constellation.

03:06So, and I've been there almost six years now.

03:09So, Rich, you have spent, you know, such a long time working in AI.

03:15I feel like it's, you were working in AI before it really became a thing, like the buzzword that it

03:20is, is now.

03:22And I'm curious to know what that world looked like when you started and what's the biggest shift you've watched

03:28happen?

03:29Yeah, I mean, I mean, my background is mathematics.

03:31That's what I went to school for.

03:32I mean, I've been a serial entrepreneur ever since, mostly on the technology side.

03:38And I would say that even though I studied quite a bit of kind of chaos theory, statistics, and some

03:45sort of machine learning before it was called kind of machine learning back in the day,

03:49I, over the last 15 years, when I, with my last company I mentioned earlier, mostly focusing on predictive analytics,

03:59which is really just saying I'm going to use past behavior to predict future behavior.

04:03And it's kind of how we predict weather patterns and things like this.

04:06But it's a very complicated data science, requires a lot of big machines, requires a lot of, you know, mathematics,

04:12et cetera.

04:13And then, of course, along came the chat GPT moment where language is now kind of the interface, if you

04:21will, to this kind of same kind of technology I was working with.

04:26And it became very, I guess, kind of groundbreaking because for the first time people could talk to their computer

04:32and they could leverage the same mathematics, actually, that's been around for quite some time.

04:36It's just the same algorithm.

04:38But now, because it has language and it has the ability to talk, you know, people think it's, you know,

04:42human.

04:42So, very exciting because what used to be just a couple of people, nerds, I guess, in the back of

04:49a library talking about vectors and vector spaces and nearest neighbor algorithms is now the hottest thing on Instagram.

04:57And I can't, you know, I'm scrolling through Instagram and I'm seeing people talk about, you know, large language models

05:03and, you know, vector spaces.

05:07And it blows my mind that it's become such a popular component of our society.

05:12But it's deservedly so because I do think there's magic in math and I think it affects all industries.

05:18And that's what is what's probably the most exciting part about it is now everybody in the world can actually

05:23interface what has been around for quite some time.

05:26But the power of it is just immense.

05:28That leads into the exact question I wanted to ask you next, because, you know, when you were talking about

05:34the magic, right?

05:36And I feel like everybody has this magical moment, this aha moment when they're interfacing with AI and they realize

05:43how much power is at their fingertips.

05:47And then that gets your brain going, your imagination running wild on what's possible.

05:52But that's not, you know, that doesn't always make it a good thing either, right?

05:57That can be dangerous for someone.

05:59That can be dangerous for someone specifically that's running a business where you kind of really get out over your

06:04skis.

06:06Have you seen much of that?

06:10Yeah, I mean, I think there's three elements to AI that are, you know, can be quite daunting for a

06:16business, right?

06:17First of all, you know, it's complex.

06:19I think people tend to underestimate how complex using AI, deploy AI, especially when you're talking about building products or

06:26enhancing your business in a more deterministic way, because people don't realize that AI is very much based on probability,

06:34which means it's never going to be 100%.

06:36And that's okay when you're talking to somebody like this communication we're having here.

06:39Maybe I don't get every word right, but people understand the context of what I'm trying to say.

06:43But if you're dealing with financials or home building or plans or diagrams or support, you know, it becomes a

06:50very difficult thing to manage and control.

06:52So this is why software companies are so very much more critical than ever when dealing with this new age

06:57of walking into AI.

06:59And then the second thing is cost.

07:01You know, it's nice to have your ChatGVT or Claude, you know, on your desktop and you're paying 20, 200

07:08bucks, whatever it might be.

07:09But deploying this at scale in a business where there's multiple things that you have to be considered when you're

07:15deploying AI, especially for mission-critical components, your cost control is going to be a challenge.

07:21And you want to make sure that you have a partner that's also monitoring that too, right?

07:25You can be very easy to get into sprawl where everybody's using AI at a degree that becomes almost like

07:31cost prohibitive.

07:32And you don't want to stop using AI, but at the same time, you don't want to be spending thousands

07:35of dollars on stuff that you have no value in.

07:38And so I think working with companies that monitor that, manage that for you and make it part of your

07:43legitimate business is more critical than ever.

07:46I think the role of software businesses is changing, but it's only becoming more critical.

07:52I don't know.

07:52Paulo, you probably have some ideas around what you're seeing as well.

07:55Well, one of the other things I was going to say is, you know, you also have to be aware

07:58of your intellectual property stuff as you're dealing with AI, right?

08:01So putting your stuff out there, just throwing it out into the, you know, the wild west of AI is

08:08not really a good model because you're actually giving a lot of your stuff away, right?

08:13You have internal processes.

08:14You have things within home building, a secret source that you guys use that's a competitive advantage for you.

08:20And if you're not really looking at AI holistically and understanding how all of that data is used on the

08:28back end, you're essentially giving that information to competitors potentially, right?

08:32So you've got to be careful of some of those things as well in terms of just leveraging AI and

08:39getting over your skis.

08:40You know, I think that was one of the things you were mentioning there, potentially putting stuff out there that

08:44you shouldn't.

08:44And it gets very, you know, you've got to be very careful with public companies as well, right, in terms

08:48of what information you're putting out there that could be non-public information that folks could potentially jump on and

08:55see it filter through on the other side.

08:58So a couple of things there that just came to mind.

09:00Yeah, I think the key word to take away from that, you know, that brief conversation is the word control,

09:05right?

09:06You want to control cost, you want to control compliance, you want to control usage, and you want to control

09:11outcomes.

09:11And, you know, in order to do that, you know, you have to have a sophisticated platform partner that will

09:16guide you through those roads.

09:18Paolo, from one of our previous conversations that you were relaying to me that you've, you know, you were around

09:24production home building whenever the big tech push was just getting sales teams off of paper, you know.

09:30And, you know, 15 years later, here we are, we're talking about AI.

09:34Do you think that the industry is, I mean, the jump is here regardless, but do you think the industry

09:39is ready for it?

09:40Or do you see steps being skipped as far as the AI adoption and integration?

09:47Yeah, I definitely think there are some steps that have to be taken, but the industry is ready for it,

09:50right?

09:51The industry has been ready for a good set of tools, in my mind, because I'm a data person, you

09:57know, around data.

09:58And when I think of AI, AI is built on top of data, right?

10:01So, the industry has been crying for good data quality tools for a long time.

10:09So, yes, I think the industry is completely ready for it.

10:11You know, if you think about what I've been doing from a data standpoint, it's really been about providing reporting

10:17and analytics on top of data, right?

10:19Which is almost a baby step to the AI side of things.

10:22People wanted to answer questions about their business.

10:25You know, what options are selling well?

10:28What's my margin across these things?

10:30They had all these questions, but they didn't have a mechanism to literally voice it.

10:34So, they had to go and run reports, and they had to go look at dashboards, which is kind of

10:39where I've been focused for a long time.

10:41And then you think about that.

10:43Now you put in front of them the ability to just ask the question.

10:47You know, just when the question comes to mind, I don't have to go run a report, pull that data,

10:51stick it into something, do my sorting to get to the answer I want.

10:56You can just ask the question now, which I think is the beauty of what I'm seeing on the AI

11:00side of things now, right?

11:01You ask the question, you let the AI model do the work behind the scenes to go say, oh, you

11:05need these data sets, and then you need to filter it by this, and then you need to pull this

11:09piece.

11:10And you put it all together, and here's your answer.

11:13So, the industry is definitely ready for it.

11:16The one caution I will have is that what I've seen consistently across home building is data quality is not

11:23great.

11:23And what I mean by that is there hasn't been a focused effort within home builders to force people to

11:33use systems in a way that derive quality data because there really hasn't been a benefit to them in the

11:39long run.

11:39So, they've used systems essentially in a way that benefits the single end user potentially, right?

11:47So, they'll take fields in a database and stick stuff in it that was never meant to be because that's

11:52the way they could get value at that localized level.

11:58When you think of that across the business, now, I need the business to have clear visibility into information.

12:03You really need discipline around your data, and that comes down to discipline about how you use systems as well.

12:08And so, that's the one caveat I'd put in there.

12:11I do think there's still work to be done to help builders or businesses in general manage the quality of

12:18their data.

12:19Rich, you had used a house or housing metaphor, house building metaphor in a previous conversation that we'd had that

12:26I want to unpack.

12:27You know, the stuff that matters most in a build is generally not, you know, what most people see, the

12:32foundation, the plumbing, the electrical work.

12:35What's the equivalent in an AI build?

12:38And, you know, I mean, it seems like a pretty obvious answer, but can you give me some details of

12:42what happens when companies skip the foundation?

12:46Yeah, sure. So, I mean, just to be on the extreme side, so a lot of people, let's say they're

12:51using a tool like Claude Code or some other tool that might be out there.

12:55You know, maybe they have a, you know, one of their, you know, cousins or something like that comes in

13:01and says, hey, I can build this software for you or whatever it might be.

13:04And they typically, they'll use prompting, right?

13:07It's kind of called vibe coding.

13:08And they'll just talk to the AI, and the AI kind of spins up this really beautiful front end.

13:12And then people go, oh, that's exactly what I want, right?

13:15So it's kind of showing you the window dressing, you know, the outside of the house, the curb appeal, if

13:20you will.

13:21But, of course, as we all know, software is a lot more complex, a lot more deep than that, right?

13:25You have many, many layers of software that go from the data layer to the logical layer to the security,

13:31the routing, I don't know, all things beneath it.

13:33Now, AI can certainly help build all those things, but it's a lot bigger than just building the interface, right?

13:39The, you know, I can, you know, see those movies where the, you see the front of the house, and

13:43then somebody goes up and pushes it, and it's just like one, you know, it's a two-dimensional, it falls

13:46to the ground.

13:47So foundation is everything.

13:49And when you build software, you start with a strong foundation, which is typically not the UI, right?

13:53It's not the user interface you start with, although that's a great way to prototype and to kind of come

13:58up with some really cool ideas on how you want it to look.

14:01But it doesn't get you to the way you are actually building something that scales, that's hardened, that actually will

14:06run a business.

14:06So this is the, I always use the analogy of Greek mythology, and, you know, I guess it's pretty, I

14:12guess it's Odysseus or whatever, the new movie's out, right, with Matt Damon, is the sirens on the rock.

14:17AI, to me, and I've been doing AI for 30 years, right, in the sense of, like, how it's evolved

14:22from machine learning to today.

14:25And it's like the siren on the rock with the new generative AI.

14:29You can see what it can do so easily, right?

14:32You can prompt it, and it can build this stuff so easily.

14:35And you kind of want to, you know, you're attracted to it because it's so amazing.

14:39But the reality is it's very dangerous if you do it the wrong way, and you waste a lot of

14:43time, you spend a lot of money, especially if you put this stuff in production.

14:46So I think there's, like, I mentioned this earlier, is there's this beautiful partnership now with users and creators and

14:53builders.

14:54And builders meaning, like, not just builders as far as home builders, but software builders.

14:58And the users have kind of gotten to more to the middle.

15:01And now they have the ability to tell us, right, software companies, what they want, what they want to see.

15:06And it's our job as builders, as software companies, to create what they want, right, to listen to the customer

15:12and build it, but build it in the right way, right?

15:15So it's no different than, like, if you're a home builder and you're going to build a custom home for

15:19your customer and the customer comes in and they start showing you pictures of the house they want, right?

15:23They're not showing you pictures of the plumbing and the PVC pipe and the type of cement, you know, that

15:28you're going to use.

15:29They're showing you pictures of, like, the aesthetics, the taste.

15:32And this is exactly what AI is perfectly aligned to do is it's meant to create and be a channel

15:40for us as humans to show taste and show judgment and to help us accelerate.

15:45But it certainly is not going to do the heavy lifting, right, as far as what we need to be

15:50done, right?

15:50So this is exactly the balance.

15:52So I always like to think, I mean, homes themselves is a great analogy for software in general.

15:58And now with AI, certainly people have the ability, especially our customers or, you know, builders themselves, they have the

16:04ability to kind of show us better what they want.

16:06And it's our job to really listen to them and build it at scale.

16:09And I think from a collaboration standpoint, that's been fantastic, right?

16:13Because if you think about it, builders are coming to us and saying they want a certain piece of functionality

16:19in a system, but they're not used to the software world.

16:22They're used to home building world, right?

16:23So they're telling it to us in their words, right?

16:26And typically you've got to have an engagement manager or someone who can almost translate that into what we're going

16:31to do from a requirement standpoint.

16:33Now folks are bringing us like little prototypes and saying, this is exactly what I want.

16:36And we're like, okay, I got it.

16:38Let us take it now and put it into the platform, right?

16:41And I think from that standpoint, it's been a fantastic evolution and essentially communication almost between folks who are non

16:49-technical and folks who are technical.

16:50It's been good from that standpoint.

16:52Yeah.

16:53And that's the beauty of AI.

16:54We can finally talk or say, explain, right, to what we want.

16:58And it will, within a very short period of time, build out this kind of proof of concept.

17:05It's no different than if you were to build a legal document or build software or anything, right?

17:10There's always human in the loop.

17:11There's always human judgment that's part of it.

17:13It might give you a great first draft, but there's always work to be done.

17:17But it is a great accelerator.

17:19Yep.

17:19Yeah.

17:20I have a friend of mine that he bought a house and him and his wife.

17:24They got into the HGTV rabbit hole and he ended up tearing his whole house down at the studs before

17:33he realized that he didn't know how to do electrical work, didn't know how to do plumbing, didn't know how

17:36to, he didn't really.

17:37They knew what kind of paint that they wanted and what kind of flooring that they wanted, but had no

17:42idea how to put it together.

17:43And it's one of those things where it's like, you know, just because you can hammer a nail doesn't mean

17:47that you know how to build a house.

17:49But, you know, someone can hear this conversation and take away, you know, like, you know, you say that, but

17:57you're the software vendor, right?

18:00So how do you, how would both of you kind of respond to that point?

18:04Because the part, you know, the AI brings about this democratization of skills and possibilities.

18:11And so how do you respond to, to that, to that very, you know, obvious kind of question?

18:17Yeah.

18:18So, you know, I had a, I worked for a CFO once and he, the one time he sat me

18:22down and he said, we're home builders.

18:24What we do is we buy land, we develop land, we build houses, we sell houses.

18:28It's that simple, right?

18:31Home builders are not application architects.

18:35They're not folks who support applications.

18:38They're not folks who understand how to do versions of applications.

18:41They're not folks who understand security of applications and things like that.

18:44So from my standpoint, when I look at it, it's like you have core competence.

18:48You really should focus on the core competency of home building because that is who you are.

18:52Your identity is, you know, you're the best home builder out there.

18:56Let us focus on those other pieces.

18:58Let us focus on what it means to be scalable in an application.

19:01You know, you can put together a great proof of concept, but what happens when you've got 35 million records

19:06in there?

19:06You know, what does that application do?

19:08And are you going to put a whole business process on top of that?

19:11Go to the public companies and now you've got audit involved as well and they've got to be, you know,

19:15responsive to that.

19:17It becomes a pretty interesting question.

19:19So typically what I say to them is, you know, look, there's a couple of things that are going to

19:24benefit you hugely from an AI perspective.

19:26One is internally I can do a lot more with a lot less now, honestly, you know, even if I

19:32look at my own team and how we were projecting to scale the team up over the next little while

19:37because of the business that we have, we've been able to not scale the business from hiring people anymore because

19:44we actually have significant efficiencies gained from using AI tools, honestly.

19:50And I'm talking like five, six people, right, in the course of a year that I was planning to add

19:55that I'm not adding anymore.

19:56So one thing is you should, your software vendor should be a lot more responsive, I think.

20:03You shouldn't see huge increases in cost from your software vendors because they should be able to do more with

20:10less.

20:12And ultimately, we should be able to provide you better tools that are more reactive to your business, right?

20:19So instead of having to go through this process of asking a ton of questions of either us or our

20:23applications, you should be able to now use tools that we provide you that just can do all that for

20:28you, synthesize it down and give you your answer.

20:30So that's kind of the way I tackle it.

20:34Yeah.

20:34And I don't think it's an either or question here.

20:36I think a lot of times people think, oh, it's this, it's, I go this way, do it myself, or

20:40I go and buy something from somebody else.

20:42And there's like a pros and cons argument.

20:44I think AI has introduced this new dynamic where it is very collaborative and you can do both.

20:52And the fundamentals of software have not changed.

20:56Everything's been accelerated.

20:57Don't get me wrong.

20:58And that definitely compresses time.

21:00That means better software, faster, potentially, you know, cheaper down the road, you know, as far as how we develop

21:06it.

21:06And I think all those things will have its dynamics and play out in the market, just like any sort

21:11of innovation, right?

21:11Any technology innovation, you know, even with home building, I'm sure that there's been over the decades of home building,

21:17there's been places where I've seen now people can print a home, right?

21:21And a 3D printer the size of a, you know, of a, of a pickup truck or, or pour cement

21:27faster.

21:27But it doesn't change the fundamentals of, you still have enough architect, you still have to build out, you know,

21:33permitting, the complexity of software, just like complexity of home building will never change because there's a certain bar that

21:41it remains.

21:42Just like as humans, we live in a very complicated world and we've introduced this new technology, but it doesn't

21:49change the fact that we're human, right?

21:51And that humans want to see and operate still in this very, very real three-dimensional world.

21:57And as much as the hype is around AI today, you know, fast forward two, three, four years, there's going

22:03to be a great reckoning around the realization that we still very much live in a human-based world.

22:09And as much as AI is going to help all of us be more efficient, more effective, and certainly will

22:13introduce really dynamic things to all of our industries, as Paolo said, you've got to focus on what you're good

22:20at.

22:20This is what we do.

22:21This is what you do.

22:22And the good news is that it's going to benefit both of us.

22:25But I don't think anybody wants to, you know, wake up at 3 a.m. and worry about their database

22:31scaling or have to pull a report for an auditor or when the systems go down and you lose millions

22:36of dollars of business because somebody forgot to back it up.

22:39These are the realities of running a software business.

22:42And I don't care how good AI is.

22:44If you forget those things, you know, you put yourself in risk.

22:47Yeah, and I know that, Paolo, this is for you, that Constellation spent years building unified, anonymized data set across

22:56hundreds of builders and multiple ERPs, which powers what, builder metrics, still AI.

23:04Can you walk us through what all had to happen behind the scenes to make the data clean, trustworthy enough

23:11to run AI on top of all of that?

23:14So this was, you know, when I talked earlier about being in different home builders and solving the same problem

23:19over and over again, it was essentially that.

23:22It was how do we take this data that we've got inside our systems currently and make it usable from

23:28a reporting and analytics standpoint, you know.

23:31And then for me, you know, I wanted to solve that across the industry because it seemed silly to be

23:35doing it one builder at a time.

23:37It seemed to make more sense, you know, if you had a partner in the space that could manage it

23:43across multiple builders.

23:44So Constellation was a natural fit there because we do ERPs across multiple builders.

23:48We have hundreds of customers.

23:50And so we were able to take that data from all these different builders across different ERPs and look at

23:56the pieces of data that really had value to the organization.

23:59There's no value really to a home building organization in a foreign key to another table, right?

24:04So if they were just pulling data out of databases and looking at some of this raw data, there's a

24:09lot of that stuff that is really just for the application side of things.

24:14So we stripped all of that away and we just said essentially, okay, these are the things that are important.

24:18So, you know, if you think about jobs and projects and communities and shell buildings and all the things that

24:25make up home building, we synthesize that down to the pieces that were really relevant to running the business and

24:32being able to take that data and do something with it.

24:35And so we put a lot of effort into actually going across all these different ERPs and saying, okay, so

24:40when I'm looking for job number in this system, it's this field, in this system, it's this field, and then

24:45putting it into a single unified data model that mapped across these systems.

24:51And that gave us a very clean set of data now that we could then do reporting and analytics.

24:55I can build one set of reports and dashboards on top of a set of data.

24:59It doesn't matter what ERP you're on, right?

25:01So that was the thinking behind it, yeah.

25:05So what's the – so for a builder listening, like who's pulling reports into Excel, you know, once a month,

25:11what's the – like the tangible difference between their current, you know, reporting world, their interface, and what they can

25:17do whenever they're plugged into this data set?

25:21Yeah, so, you know, you can – at an individual builder, you can go – you can have a data

25:26team that pulls your data, cleans it up, and provides your information back.

25:30There's nothing that stops folks from doing that.

25:33And a lot of the large builders do that, right?

25:35They have the capacity.

25:36So first of all, there's a cost component to that, right?

25:39You need a data team who understands data, which is not cheap resources.

25:42So you've got to stand up a data team to do that, to pull your data for the most part.

25:46That's one thing.

25:48The other thing is, because we are doing this at scale across multiple builders, we also have the ability to

25:52give comparative data back.

25:54So we can say, you know, you know how you're performing on a certain metric within your organization, but how

26:01does that stack up against other builders in the whole ecosystem, right?

26:05Whether it be in your market area or state, you know, more broadly across the country, if you want to

26:10look at it that way.

26:11But we have the ability then to say, okay, this is how you're performing.

26:15This is how your competitors in your market are performing.

26:20You're ahead of them or behind them in that aspect.

26:22Here are some things that we can see in the data from them that you could do to be more

26:27operationally efficient or more cost effective in your purchasing, for example.

26:32So that's one of the things that, you know, we can bring to the table.

26:35The other thing we can bring is, because we're agnostic on these ERPs from a data set standpoint, as happens

26:42in home building, you get acquired, you change ERPs.

26:46We manage everything from a reporting and analytics standpoint where you don't really have to do anything.

26:50Your reports and everything just keeps working, even when you switch out the back end.

26:55So those are some of the benefits of utilizing a standardized reporting platform like ours.

27:02This actually leads into a question, you know, you were talking about, it's not an either or, not necessarily a

27:08black or white, it's a collaboration, right?

27:10And between the vendor and the client.

27:14The foundation of all of this is good, clean data.

27:19And I'm curious to see or to know how you both view the responsibility for the providing of that good,

27:31clean data from a relationship standpoint, right?

27:34Like where, where does that, you know, your role as a vendor bleed into, you know, helping a builder or

27:42a client provide good, clean, better data to operate on top of?

27:47And where's, what is their responsibility when selecting or harvesting that data for you?

27:52Does that, did I, does that question make sense?

27:55I don't know if I phrased that correctly.

27:56Yeah, no, it totally doesn't.

27:57You know, from my perspective, it's a two-way street.

27:59So generally, so one of the things that we did is we found that in-home builders, again, because it's

28:05not a core, data is not necessarily a core competency of builders.

28:10We wanted to do all that messy work.

28:12So we took on the challenge of doing, you know, doing our best shot at it.

28:15We said, these are what we feel is important from a data standpoint.

28:19This is how we're going to map your data set to this.

28:21And, you know, even, even builders using the same ERP backend implemented completely differently and used things completely differently in

28:29the same system with the same screens.

28:31So we had to go through that whole process of mapping all that stuff and getting down to a common

28:35data model that we could use.

28:37So that's, you know, that's part of the process.

28:39The other part of the process is then to take that learning back to the builders and really help them

28:44understand how, at the end of the day, if you put some discipline around how you use the system.

28:50So, you know, one of the CIOs that I worked with in the past said, you know, everything from the

28:55keyboard forward is mine.

28:57And, you know, what you do from a process standpoint in your business in the field is yours.

29:01But if you're going to be entering data into the system, this is explicitly how you do it.

29:06This is how you use these fields.

29:07This is how you use these screens.

29:09And the dividends that get paid back on that are huge down the road, right?

29:13It's hard for the person who's actually doing that data entry to understand it initially until they start seeing, you

29:19know, the fruits of that coming back.

29:20So it's a bit of a learning process that happens both ways.

29:23One, you help them clean it up and then you take it back and say, okay, let us help you

29:27now actually be better at how you deal with your data in the first place so you don't have to

29:31spend a bunch of money with somebody else to go clean it up for you.

29:34Yeah, just to add on top of that, I mean, I think we're also in a very exciting time where,

29:39you know, if you do have your systems of record and you do have your data straightened away, right?

29:44And you've partnered with a strong software provider that's, you know, providing that, right?

29:48Cleaning it, providing the data warehouse, kind of what Paul was mentioning.

29:51I mean, there's this great unlock with AI that's happening because before, quite honestly, you know, and I'm not speaking

30:00specifically in home building, just across the board with software, no matter what industry you're in, we do sense that

30:06we put a lot of data into these systems, right?

30:08I mean, we all have CRM systems, we all have support systems, we're pumping data into these systems.

30:13But, you know, to get it out was somewhat limiting because, you know, if I wanted it, like God and

30:18Paulo, you know, kind of mentioned earlier in the call is, you know, if I want to say, hey, I

30:23need a report that tells me my sales for the last two months from this region, right?

30:26You'd either have to have a data expert or you'd have to build that report.

30:29And, you know, that's an inhibitor, right?

30:31That's a hard thing for a CEO or CIO or anybody really to build dynamically, right?

30:36So we would just kind of like, ah, well, I don't really need that report.

30:38Or maybe we have to hire somebody to pull it out and sell and we have to spend two, three,

30:41four hours to pull it together.

30:43But now with the introduction of, you know, large language models and the ability to talk to software, talk to

30:48data, it is now 10x the value because now you've unlocked that capability for anybody to talk to their data.

30:56So I think now the value of all of that data that you might be collecting collectively or in a

31:02single company over the last decade or more, you now have the opportunity to really, really dive into that for

31:08the first time at scale.

31:10And it's not an inhibitor that software companies had.

31:13It was just we had to wait for a technology like a large language model, AI, to come to play

31:18where it could translate language into queries or into, you know, formulating this data in a visual manner or, you

31:26know, a spreadsheet or whatever it might be.

31:27So it's an extraordinarily exciting time.

31:30I've been in data business forever.

31:31And, you know, like I said, it took smart people, very expensive people, a lot of time to generate what

31:38AI can do now, you know, for you based on just a simple prompt, right?

31:42And this is the fantastic element of, I think, what the Home Builder Group is building is, you know, putting

31:50AI on top of this strong data set is really a massive unlock for companies.

31:55I mean, you're able to now think about things like what could affect my business and decision making.

32:01You know, a great Home Builder makes four or five really strong decisions in a month that can impact their

32:06sales, can impact their costs.

32:08And a lot of times that data is there, but it's unreachable or maybe it's just not in the language

32:14that they would understand.

32:15Or maybe even to some degree, sometimes Home Builders don't even know that they could query it that way, right?

32:19They just live in the reporting system they have, but now they can actually start investigating and drill into and

32:26have a conversation, you know, with their business.

32:29That's, I think, the most exciting part of what Paul and his group is building.

32:33Yeah, let me give you two examples.

32:35And this is from, you know, being in industry with different Home Builders, right?

32:37So I've known Home Builders where to get the margin on a home that has been sold is a two

32:45-week process, honestly.

32:47Now, maybe I'm going back a little bit, but it was to get that generated in terms of finding out,

32:52you know, what options were put on the house, all that kind of stuff,

32:55was this lengthy process of running multiple reports with the data team, and eventually you'd get your information.

33:01I worked with another builder where as soon as a home sold, within five minutes, the CFO had an email

33:08in his inbox that had a detailed breakdown of the gross margin on that.

33:12Now, tell me who can pivot more quickly, right, when something happens in the market, somebody who's waiting two weeks

33:17for the data or someone who, you know, is getting it at their fingertips.

33:21And I think AI probably accelerates this, right?

33:23But, and again, it has to sit on good data because if you take AI and you put it on

33:29top of bad data, you can also get some anomalous stuff that doesn't do you well in the long run

33:34and can certainly inhibit you.

33:35Because AI will take what it's given, you know, across a large set of data and then will figure out

33:42the patterns and generalize and give you, you know, it's based on everything that I've got.

33:46But if you're giving it a lot of bad stuff, it's also going to give you bad stuff back, right?

33:50So it's not a system that will save you from yourself, essentially.

33:57So at the end of the day, you know, this notion that we've spent all this time building this clean

34:02data set, we honestly weren't thinking of AI at the time.

34:05We were really thinking just of reporting and analytics.

34:07But as I said earlier, you know, the natural progression of that is the language ability to ask those questions

34:13as opposed to running reports and getting that data back.

34:16So I think we're in a great place.

34:19I'm super excited, like Rich was saying, you know, in terms of what's ahead of us because we have great

34:24data, we have great tools.

34:27And, you know, as we show this to our customers, you just see their eyes light up.

34:32It's almost like you've unlocked a superpower for them because for so long they've been trying to ask these questions.

34:38And in many ways they haven't had the verbiage to do it because they didn't understand the technical components and

34:45now they can.

34:45And it's just it's great to see that on their faces when they actually when they actually visualize it.

34:51Yeah, the greatest investment, the greatest investment any business can make right now, you know, when it comes to AI

34:56is not AI itself is.

34:59It's really spending on their data, getting access and capability on software that gives them exposure to their data and

35:09then layering AI on top of that is how the natural human user interface.

35:12Right. The language is now the new UI.

35:17I mean, I love the fact that AI can get me a spaghetti recipe and summarize a document, maybe write

35:26me an email, but that is not ROI for a business.

35:30There's no outcome that's going to provide you some sort of glorious result at the end of the year in

35:35your P&L.

35:35The P&L is going to come from decision making, access to information and using AI to not just produce

35:43a report, but talk to the AI and work through the challenge.

35:47Why is that? You know, why is this number off? Why is this home less profitable than this home?

35:51Dig into it for me. Tell me exactly. And just literally talking to it as if it's the smartest person

35:56in the room and then discovering areas where you can save thousands of dollars or make more money or opportunities

36:02that you would have never seen.

36:04Right. It's the known unknowns and the unknown unknowns. So all this starts with data.

36:10And obviously, this is why Paul and I spend our entire lives, because, I mean, every decision we make as

36:14humans is based on data, data, what we understand or know to be true.

36:18But if you could unlock that for your business, I mean, it's the best investment you can make.

36:26Everything else around AI, in fact, AI itself is all based on data.

36:29And so if your company needs to grow, needs to be profitable or needs to do any decision moving forward,

36:37there's no excuse why you shouldn't be investing heavily into your data, because that information is going to be what

36:42you're going to use.

36:43And AI is that new tool to allow you to access that data.

36:47But without the data, the AI is kind of, you know.

36:50Right. Well, and a huge factor in that equation, as well as the right partner, right?

36:55The right find the right vendor, the right software provider.

36:59And, you know, what do you encourage builders to ask of their vendors, providers to assess on whether, you know,

37:09whether they're not, whether or not they're actually bringing AI into the organization?

37:13Like what, you know, what is a partner as a good technology partner?

37:17What does that obligate you to do?

37:19You know, their expectations are continually raised around vendors.

37:25In both of your opinions, what should those higher expectations look like specifically?

37:31Well, my opinion, first and foremost, I would say, what's your expertise in this vertical, right?

37:36What's your expertise in my domain, right?

37:39I mean, I think there's probably going to be, you know, I mean, I can't even probably count how many

37:43companies are going to come around, you know, and try to sell you something that's AI that they built, you

37:49know, vibe coded or something like that.

37:50They have no industry experience, have no depth and no, you know, no customer base from learning.

37:56So do not discount just because we have this great new technology and we're all excited about it.

38:02Do not discount industry expertise and know-how.

38:06And I think that's number one.

38:08That'll never change, right?

38:09Vertical expertise and domain expertise like Trump AI every day of the week.

38:15The second thing I'd ask is, how are you going to empower me and my business to leverage AI?

38:21And basically what they're asking is, how are you going to produce data and organize my data, whether I have

38:28one or many ERP systems, whether I have disparate systems data, whatever it is, how are you going to help

38:33me create that layer of confidence, as Paula was mentioning earlier, right?

38:38How do you give me good data that I can strap AI on top of it, whether you're providing or

38:43they're using any other tool, but I need access to that data.

38:46I need to be able to make decisions faster as well as better decisions, you know, for my business.

38:53You know, those are the two primary things I would focus on.

38:56You know, what do you know about my business and how do you make me smarter?

39:01Forget about technology, forget about everything else.

39:03You know, just figure out, like, these are the outcomes I want.

39:06I want somebody I can trust and data I can trust.

39:11If you give me those two things, you're going to help their business, right?

39:14For me, the first one is key, you know, the vertical expertise.

39:19You know, there are some very large ERP vendors out there that happen to have a home builder module.

39:24How much have they invested in that in the last X amount of time?

39:28How many resources do they have who truly understand the vertical?

39:31You know, look at Constellation Home Builders.

39:33That's all we do.

39:34That's all we do.

39:35We do home building.

39:37We do platforms for home building.

39:39We do the data of home building.

39:40We do, you know, the ERP.

39:41We'll do the scheduling.

39:42We'll do everything related to production home building.

39:46And we have 300 people who have expertise in that, right?

39:50So to me, I think that's the key thing.

39:52Do you have a partner who understands you?

39:54I mean, any time you go into a business, you really need to have a partner that understands your business.

39:59Not just that wants to sell you a piece of software.

40:02They need to understand the fundamentals of your business.

40:05And if you look at, you know, even our organization, we have a lot of people from industry who come

40:10into Constellation Home Builders, which is key, I think.

40:13You know, we don't just have software folks and we don't just have support team members.

40:17We have a lot, a lot of people who've worked in industry for many years.

40:20So I think, to me, that's a key one for sure.

40:25Okay.

40:26I got, we're running up on our time here.

40:28I've got two takeaway questions for both of you.

40:31And I'll, Rich, if I can start with you and then we'll close out with Paolo.

40:35So five years out, what does the builder who got this right in their AI adoption compared to the one

40:43who either ignored it or they went a little too far down the DIY road?

40:48What does the future look like for them in five years?

40:52And what's one recommended move to make within the next six months?

40:59Yeah, I mean, obviously, AI is moving very fast.

41:01I mean, coming from the software world, usually we would see size, you know, sizable adjustments, right?

41:07Whether it be chips or whether it be software, maybe every two, three years.

41:10I was even telling this to my, one of my developers is that we used to have to rebuild our

41:13software every three to five years.

41:15But it seems like now we're on a three month cycle.

41:17So things are accelerating.

41:18So five years is like forever.

41:20I think if you were just to kind of skate where the puck is going, I think you have to

41:24assume that AI is going to be extraordinarily intertwined into your business.

41:31Every single person within your team will be using AI, whether it be at a tool or maybe even managing

41:36an AI agent that's doing automated work.

41:39So I think builders who get it right, they're going to be training and focusing on their team and their

41:44staff on getting them to become AI ready.

41:47And then working alongside their vendors who have a very clear roadmap on developing AI into their systems and kind

41:54of really working on that human plus AI recipe, which is what's going to give them the best ROI.

42:01I'm a math guy.

42:02I'm the first guy who loves technology.

42:04If I could outsource my, anything I can in my life to technology, I would.

42:09But I've been working with generative AI and, of course, machine learning prior to that for years.

42:14And every day that goes by, I become more and more confident that it is a human plus AI world

42:19in the next five years, as much as anybody would tell you, because there's judgment, there's taste, there's liability.

42:29As long as we're still working on the US dollar or Canadian dollar or whatever, the monetary system exists, and

42:35there's people that are promising to deliver product and people receiving that product, we are very much still in a

42:41human-based world.

42:42And that's not going to change in five years.

42:43I don't care how good AI gets, right?

42:45We still are still at the top of the food chain, if you will, in that world.

42:49So I do think, though, that being AI literate, making sure the vendors you're partnering with are AI literate and

42:55have a very clear roadmap on there.

42:56That's how it's going to benefit you.

42:58Every decision that's going to be made in probably less than five years, but certainly in five years, is going

43:05to be AI-supported.

43:06Maybe not AI-driven, but it's going to be certainly AI is going to be in the room making those

43:09decisions with you, including, which is another reason why you have to have really solid data, and making sure that

43:16you get more data.

43:17I think AI in the room speaks both ways, right?

43:20AI will not only help you capture more data about industry, about conversations, about what's in the heads of your

43:27customers or in the heads of your teams, but capturing that and making better decisions is going to be a

43:33key component.

43:33But the people who get it right are just going to dominate their market.

43:38I mean, nobody can argue against the fact that with the right AI, the right data, the acceleration that you

43:45sense and feel in your business at the cost comparative to somebody who is non-AI and is not leveraging

43:55data to the degree you are,

43:57I just don't see how there's a competitive advantage or there's no competition.

44:04There's just no way that you can compete with a business that has gone AI-native and really thought about

44:10that in a careful manner.

44:13I think slow is fast with AI.

44:14I think you want to be methodical and you want to make smart decisions, not get too bought up into

44:19the hype, but understand the fundamentals of what AI can do for your business and be practical and honest.

44:24AI is not the solution to everything, right?

44:26Not everything is, you know, have a hammer, not everything is a nail suddenly.

44:30But I do think it's in five years, I do agree with kind of the sentiment of what I would

44:35call the general buzz of AI.

44:38In five years, if you are not embracing it and your team is not embracing it and using it on

44:45a functional day-to-day level, you know, you will suffer, your business will suffer, right?

44:50You won't have the cost control, you won't have the additional capability of organic growth profits.

44:56That would come your way by using it.

44:59Yeah, and I mean, I'd echo many of those sentiments and I'd probably be a little more aggressive about it

45:05and say,

45:06if you're not already starting to down the process of understanding how AI fits within your business, you're already behind

45:13the ball a little bit, honestly.

45:15And, you know, I think you asked, you know, one of the things I would certainly ask any company who

45:22is thinking about AI in their business to put some effort on the data side.

45:27I know we've beaten this horse to death, but if you're going to invest somewhere right now, I'd say the

45:32data side, get that sorted out.

45:33And it just opens up the whole path for you.

45:36The other thing I would say in terms of, you know, five years down the road, I think, honestly, the

45:41market will sort that out.

45:43The market's going to get rid of the folks who don't embrace it because you're going to have people who

45:47will be able to – home building is cyclical.

45:51As we all know, it goes up and down, right?

45:52The ones who can pivot, the ones who can really adjust in real time based on good data are going

45:59to be the ones that really profit and survive at the end of the day.

46:02I think – and especially with what AI brings to the table, I think it accelerates that.

46:08And so if you're not doing AI and your competitors are, you're at a significant competitive disadvantage.

46:14So the market will sort out what happens in five years.

46:17The ones who get it right will be the ones that are still around, honestly, I think.

46:20Yeah.

46:21And just to kind of double down on the answer, the second question you had about what they should do

46:24in the next few months, six months,

46:26is realize that the benefit of AI is only going to achieve that 10x, that multiplier, that seismic jump in

46:36productivity if you invest in making sure that you get your house in order.

46:41Get all of your data and have full visibility and confidence in that data, and then your AI will take

46:47you the rest of the way.

46:48So don't jump.

46:49Don't put the cart before the horse.

46:51Invest heavily in data – you know, in the data side of the business now.

46:56And AI is going to only love what – you know, that investment, take that investment, and give you a

47:01solid ROI on it.

47:03Yeah, and just going back to what we said before, you know, find a solid partner that you are comfortable

47:07with,

47:08that understands you and your business, and do it as a partnership because your core competency is home building.

47:14And, you know, the partner's core competency should be all this other stuff that makes this magic happen to help

47:18you drive your business forward.

47:21Paolo, Rich, thank you both, gentlemen.

47:23It was a pleasure.

47:24Like I said, I appreciate you taking the time to come and speak with us.

47:27No problem.

47:28Great.

47:29Yes, sir.

Category

Transcript

Comments

Recommended