Real Estate Addicts

Brian Vizarreta has been acquired twice in four years, from StructionSite to DroneDeploy to Procore, and he came back to explain what all that money is actually betting on. He and Marc, Ray, and Dan get into facade mapping that finally settled a four-fix water leak, why the most valuable thing to ask an AI is what isn't there rather than what is, and whether any of this makes sense for someone building six-unit multifamily instead of $300 million towers. 

What is Real Estate Addicts?

Marc Savatsky, Ray Hurteau, and Dan Rubin open up and discuss their own deals, budgets, and mistakes every Tuesday.

Not guru content. The stuff that decides whether a project makes money: what a triple decker really costs to build now, the retroactive Newton tax bill that landed on unsuspecting condo buyers, the helical pile surprise that cut a deal's IRR from 70% to 40%, the eleven-week-old demo company doing $6K a week against a $20K burn. They read their own numbers on air, including the ugly ones.

A typical week is some mix of: a breakdown of one of their active projects (an 1874 adaptive reuse in Amesbury, a Dorchester three-family held fourteen years, a waste and demo business), a guest who does one thing extremely well which include zoning attorneys, architects, labor lawyers, energy consultants and brokers. Ray's Multifamily Minute on the news that actually moves the needle, and a round of Overrated / Underrated / Appropriately Rated.

Heavy on Greater Boston and New England, because real estate is hyperlocal and pretending otherwise is how people lose money. Useful anywhere the same problems show up: permitting, capital stacks, subs, tenants, and the gap between what the spreadsheet says and what the building does.

For investors, developers, GCs, agents, and anyone still holding a W-2 while they figure out deal one.

Co-hosted by Ray Hurteau, Dan Rubin (Instagram: @rhinvestgroup), and Marc Savatsky (Instagram: @choose_boston)

Follow the Real Estate Addicts Podcast on YouTube: @RealEstateAddicts

How are we doing this? Welcome to the Real Estate Addicts Podcast with your host, Marc Savatsky. Choose Boston. Ray Hurteau, RH Investment Group. Dan Rubin, RH Investment Group. And joining us today is our guest. I am Brian Vizarreta. Two-time guests. Two-time guests. Last time was episode 83 in 2022, about exactly like four years ago now. Yeah. Well, you're 31 now and you're back with us. And the last time I was here was through two acquisitions ago. So you've been busy. Yes, tell us what you've been up to. So you were when we first. had you on you were working for StructionSite which had i think just been Or was in the process of being acquired by DroneDeploy. Correct. Yep. So, but prior to that,

Start us the beginning because I like your path. It's an unconventional. Right into like the assistant project manager. Then I became a PM. And now like you kind of like zigged and zagged and landed in a cool spot. Or you went like straight to tech? Yeah, well, all right. So, quick note before we get into this, everything I say here is my own view, not my employers. I'm not speaking on behalf of DroneDeploy or Procore. It was a little bit of a combination. So, I come from two generations of builders, and I've always kind of been on this path of you have to go into construction. But in college I would always do a lot of hackathons and I always kinda loved playing with the tech. And at the time, the concept of like a technology division did not exist. Uh so I you know coincidentally was able to go Essentially started my own little company out of a hackathon where I then met So people from Consigli.

And what ended up being a consultation conversation, I ended up You know, napkin sketching my job offer. Uh did you study you didn't s did you study like computer science or c engineering or did you actually study No, straight straight civil engineering. I was going to go the traditional PM superintendent route. Uh, but uh, you know, took definitely took like a bit of a pivot right after school where I got the opportunity to go to Consigli and build out. A reality capture division. Which was exciting. It was everything from drones, laser scanning, 360 cameras. Uh and then I kind of hit a wall where I felt like we we did a really good job implementing it. But I wanted to go back into the field. So I went back in there as a superintendent. Oh, interesting. Andrew Jackson, cool. Yeah. So Brian was working on a major commercial project in Cambridge. And at the time, the facade was something which was super complicated. And it was for a big institutional client. And so Brian, part of Brian's job was to figure out how to sort of document the exterior wall condition because there's massive liability in dollars. And like, I like to point out, very rarely is a superintendent getting on a condor lift or pump jacks.

Scaffolding and like really putting their nose up into the wall, but you kind of figured out some cool ways to do that, yeah. Yeah, so it was definitely some of our you know biggest wins for facade mapping. But I did personally love getting in a boom lift, I love lights so. But on this one building, there were these four very interesting architectural roof, like little slivers. Right. Didn't have any function other than it's look cool, just all study. And he's interesting. So you're using drones to map the facade at a regular interval. So you can see the progress and steps aren't buried. Like you can fly a drone once, and after the metal panels are on, and it's not really going to do much except give you a pretty view. But if there's three different layers of waterproofing and a peel and stick membrane, followed by this counter flashing, followed by a sealant, like. You kind of have to like keep flying the drone to like prove that you did all of those steps. And it's interesting because you never really.

When you're documenting You kind of see a combination of two different types of photos that people take. It's either a cell phone photo from the ground. Or it's a ton of really close-up photos that you have no idea where on the elevation they are. And I think the best thing that we were able to do is we were able to do these facade maps to where you can click into each, you know, a specific waterproofing detail and go click in and see the entire history. It's like a Google map, but of a facade of. Exactly. And I personally was on the exteriors. So every time it rained, I would have a mini anxiety attack. And I think people don't talk about this side of being in the field and like, you know, obviously from building. Every time it rains, you're at home with your family. And you're probably still thinking, is there water in my building right now? Oh, yeah. Some of us think of that more than others. Some of us think of that more than others. Just had a conversation about water on the way here or the way down. Yeah, we did.

Mm. I just solved a water, knock on wood, solved a water problem. We had a leak. I'm doing a building in East Boston, and this one wall kept showing water. And we got you mentioned that it was like you had to hold off on insulation. Yeah, we had the floor. JF Top Floor has held all the flooring off this room. It's become its own little like mini construction schedule because we kept trying things. And, you know, it's kind of like being a detective or a triage nurse. You know, house MD, if you will. But like, we try to fix this and then look at that, then water test it. And we had a good water test this weekend on account that it rained like hell. What was the issue? Ultimately, the issue was where the deck. The building bumps in on the top floor and the deck on the outside where it met the exterior wall right there, there was a scupper and there was a bad waterproofing joint. Um but we thought

We thought it was the roof above. We thought it was the pump jacks that had left. Pinholes where the pump jacks were screwed into the parapet. Like, I did find one of those. So I thought that that was a good culprit. It wasn't that, although we did fix that. And then I thought the siding or maybe the vents, because we had like nine dryer vents stacked side by side up there. And I was like, could that be it? And finally, we thought the parapet wall top. Wasn't properly capped. And we kind of redid that, just built. And then finally we. Found it now. You borrowed my thermal camera at one point for a different issue. Yeah. Did you get your own now to try and find it that way? I do have my own. I mapped it with a moisture gun. Like, she is a, it's like a stud finder, but it's showing up moisture levels. It's like with a magic marker and this thing on my mapping, I'm like, this percent, that percent. And it kind of tells a story. Even with that, there's so many different trades that touch an exterior, right? Back and forth, back and forth. It's never black and white. It's always like.

Ah, well, you know, that insulation was left on the roof for two months, or and they all bled each other. Yeah, and it's everyone else's fault. And like, I think the, the, My the unofficial title of what I say the benefit of reality capture is, is it's just an unbiased view of reality. And that's really all it is. Existing conditions documentation. Visual existing conditions, I don't know. Something like that. I dunno. Uh the the the combativeness is It's a risk mitigation game. Like at its core, construction is about risk transference. And like if you really play the game, it's you know, the owner is just who cuts jacks. And then the G C is has the responsibility. But if they do their job right, it's gonna flow downhill to the appropriate subcontractor and so on and so forth. But um I think it's it's kind of set up that way. Yeah. Now when we had right after we had Jan, we were working on our project in Lynn.

Dan had a cool like Had his construction helmet. 3D printed something that like screwed into my hard hat. Yeah, yeah, yeah. So he attached the camera to the 360 camera to the helmet or to the hard hat, excuse me. And then Dan would walk the sights. And it would take pictures at like a staggered interval. Now before we started hopping on here, you mentioned you showed us an example of some really cool interesting Aspect of it where now you can just kind of click on a surface and you can just kind of port over to it. So It's doing way more obviously doing way more than what it is now. Tell us what's James, like being able to talk to. AI and like asked me like, hey, bring me to where this spray paint was on the floor. Yeah, I guess there's almost two parts of it. Like, there's the actual technology and what you're able to get out of a$600 360-degree camera, which is we're able to leverage a lot of the

Photogrammetry 3D engine capabilities, which creating maps of the floors, the ceilings. It's a 3D mesh that's measurable on the underside. But then there's also the AI side, which is You know, it's no different than you putting any photo in ChatGPT and asking, hey, what am I looking at? It's just that on a much larger scale where every photo that's

photos per second is getting context. So if I want to say find me a blue trash bin. Oh, but what I think is the most Interesting part about it is the negative query. So, you know, show me every door that doesn't have hardware on. Things like that. And that is where it helps you create punch lists and all that stuff. Exactly. And it's not just being able to ask, is this there? It's what isn't there. Yeah. And understanding what should be there based off other contexts that you have in the project. Yeah. That's it. Yeah. Instead of having to do it manually. Yeah. You know, and walking a hundred-unit building and seeing what. What's missing? Now you just ask or you walk around and you make your list and you're like, you know what? Like, I made my list or I might have missed something. I'm looking at 70, you're looking for 75 different things to work with the AI because it saves you all that time. Right. Now, that's what I'm saying. Is you just walk it and you're like, give me a punch list for this trait. I think people are.

It's changing, you know, AI in general is changing the way we work. Yep. In general, it you always used to make the original file and then you spell check to go edit it. And now we are starting with the AI, creating the overall document, and then you're editing it based on your background knowledge, things like that. Yeah. Be interesting to see if municipalities adopt that, whereby like at final inspection, you need to provide a construction photo documentation behind the walls. That seems excessive. If I were the mayor of a town and uh

in construction inspections from like a municipal Inspectional services perspective is very flawed. It's impossible to be done really well. You're just, you're not there enough. You can't possibly know everything, see everything. But like, They preserve your as-built drawings at the end of the job so that some future buyer or a litigator or what have you can come into the and request those records. I think it'd be good. If I were a, again, municipal official, a mayor, I would say, like, yeah, you should also have photos of the whole thing. But like From like what, how many stages and how many photos? And how do you get it? I mean, primarily. Do you do like a rough and then a final? Yes. Prior to drywall and then after. I think a lot of that even photo documentation can be means and methods in a way of like, you know, you can you're just gonna if you put more restrictions, people are just gonna check boxes in very various ways.

Right. But the biggest thing that we're seeing is a really interesting shift is we had to build our own division one spec that owners would request. That took into account all of this new information because the last Division I spec that they're using is older than me, is very vague in the sense of we need periodic photos taken from the same. angle, you know, named in a specific way. But there's very vagueness of like How much, how it's organized, how it's handed over. It's just. Take photos is really what most Division I specs kind of sound like. And if you're building a project that's over$300 million, you kind of Should expect because you are paying people, it's part of their job to photo documentation as part of field operations. and just having clearer stances of what you're expecting. That makes sense. When I was a youngin', we used to create, we like the assistant super's job was to create these weekly updates for the executives at the construction company. And we'd email them weekly to the executive assistant who'd put them in a binder and put them on whatever GM CEO's desk.

Not even a PowerPoint. It was just like a binder. There's a binder. Physical binder. So, do you? I guess my question. So, I have two, I guess, two-part question. In your mind, like, You're going from like almost like startup. To pretty corporate? Like, or is that kind of like? I don't know, do you have Any reservations about that? I think let's just start with The last time I sat here. I was inStructionSite.

And I think now going through this second acquisition. You're realizing that It's not necessarily, it's not just Right? Acquisition, that's not the story. The story is that. bet on reality capture and photo documentation has gone larger and larger. Right. It went from DroneDeploy. Understanding that there's an interior part of the context that you need to collaborate and that should all live in one place. I guess the second part of that question. And I'm going to play devil's advocate. Like what what can the smaller guys use? to help them. in their projects. Like, you know, if we used StructionSite because it was a small company and it was great and you gave us a free trial. But like, you know what I mean? Like, what, what can we use and what can like most of our listeners use? To help them and their projects. When I met all of you, And I started talking about 360 cameras, your first reaction was like, I don't need that. You know, I'm building, you know, five, six unit multifamilies. And then you got that larger project and you're like, okay, I can see it. But there was still a skepticism of like,

Do I really need this? But having used it, you're. My shift has changed of like, oh, this is something we should really carry some money for. Um Yeah. The evolution from our end was We were doing projects, and I think on the first few projects, we just never took pictures of anything unless it was like an issue. Then we went to, okay, we're at this phase. Let's walk through with our phones and take pictures. That was psychotic. Uh then we went to Matterport.

Two stages of the project. Yeah. And then we met you and tried out the StructionSite. version of it. And it was great. And then it switched to DroneDeploy and now it so it's So it sounds like it's just Part of the evolution. Well, I think it's also a part of how big you are. And like, I don't agree with that. I think any builder should. I'm in team documentation for obvious reason

At a size where we are. I don't think unless you have a full-time super or Like if you're on, unless you're on site every day or multiple days a week. Well, this leads me to my next question. We've got AI, we've got autonomy. Do we ha now when The other problem is you're walking around while people are working half the day. Is there an opportunity to have Something done. Like it covers the grounds, interior and exterior, autonomously. In the exterior, people are using dock drones every day. Essentially, you know, imagine like a square dock. Doors open up, butterfly doors, drone comes out. It's like a Roomba, but like for a drone. Exactly. And it's plugged into power in Ethernet, and you're essentially able to, you know, it does its mission using our software. Lands, uploads the data. Real refresh every day. Refresh whatever you want. It's doing on large, large sites every couple days. To a point where they need it done so frequently, it's more cost-effective to buy instead of paying someone full-time just to do this.

We have like Reolink cameras on the exterior of our jobs now. And it's great, especially early in sight where I can see the whole site. But as soon as the exterior walls go up, my view is significantly cut off. And um Then you think about like, well, maybe I put one on the inside, but like, to what end? Then I can see the living room in one unit. Right. Like, where, where is there a way that I can basically want it at your access points, like, who's coming in and out? Well, yeah, but like, I would like to see, like, even within the units, like progress stuff. Yeah. Well, I think that's where the autonomy comes from. And you're not. I'm not saying you don't you can't go to the job site every day, but you take a a very mundane task, right? Like no one really wants to be that bad. The 360 cameras and the drones and all that. Like the hardware is the cheapest part of it. So let me ask you, it's interesting because we're talking about this.

You know, the interior side, right? You can get a rover. But then there's also like the human aspect of someone leaves drywall in the doorway. Rover can't get in there. Right. There's like the actual slight conditions. And then secondly, Uh hiring. Hiring gets harder and harder. Yeah. You can If you can find exactly who you want right now, you're probably going to go hire them. But if you can't, I just remember. every summer internship. the amount of like opportunity we would give an intern. To be able to do one 360 walk and produce an output that is beneficial to the entire project team, from the PX to the Super to the PM. And it gave them almost like this scalable way to help what other scenario. Someone provides that much value in 10 minutes as a brand new entry-level person? I think it's going to change the way. We even define roles. Like you're going to be able to hire. freelance photographers in a way, as long as they get their OSHA certification and things like that.

Yeah, maybe you don't have to buy it. Maybe you can just lease it for a specific job instead or something like that. Or you just, like you said, just pay the guy from Circa to come. Three times a week to walk the job. Yeah, yeah. You have a labor foreman on site full time? Yeah. Hey, just walk around. That's the most important thing we're focused on: making it easy that anyone can go pick it up. Yeah. I mean, When I gave you the cameras and you were able to put photos on the floor plan, it took maybe five, ten minutes before you're like, okay, I get it.

Right. I get it. So it's just the smaller jobs are are harder to absorb. costs like that. So you would need like it just you know. Um But what I will say there too is, you know In every scale of a project, that scenario flips to, like, you also can't afford for something to go really wrong on a small project.

Well Not at all. I guess it it depends on what that cost is and what the what the Cause like, I don't know how would Would you have been able to find that water intrusion issue on Mark's project if you were documenting it? No. Yeah, it I mean, yeah. There's certain things that you're not going to be able to do. But there's certain stuff that you always have to manually inspect. And they're like, it's like a one-off thing. So I don't think you want to throw the baby out with the bathwater, you know? There were. I even thought at one point it could have been our framer. He did the decking, and I thought maybe he. I did some. We were using pedestal pavers, and I thought maybe he did some kind of a ledger board support around the primer to hold up one side of the paver, and that he just screamed through the memory, but he didn't, and it was all flashed properly. But, like, they could be. But something as venine as somebody's walking around after everything else, they drop a hammer or a drill or something and they poke a hole in the roof. But sometimes it's a photo of everything not being messed up that you need to not back charge the wrong person. Yeah.

That's a great point. Sometimes people are like, oh yeah, like nothing's happening on the floor. I'm like, that's a great time to document it. Because if that combination is like, oh, that palette material sat there for three years. That might be a very different conversation a year and a half from now. Yeah. I mean, the other interesting thing that we're starting to see with uh your technology's the ability to forecast. progress specifically. So if you're taking enough regular Interval progress shots. It can figure out, like, okay, you're drywalling this much per week. You'll need another three weeks to finish drywall in the building. Like, integrate that into like three-week look aheads or whatever and their final schedules and all that stuff. It's another job that I had to do when I was, you know, in diapers. Which was steel progress. Like every day, the steel guys would finish. They would leave the site. I would go out with colored pencils and crayons and be like, looks like they flew this steel beam, that column, this. And then I date it, you know, September 8th.

And then same with concrete foundations. Where did we pour what day? There is something we all learned here by highlighting. Yeah. Printed out drawing. Yeah. Cause you understand how work worked. But there's also opportunities. I remember there was a team in DC that didn't do like this type of. I think it was like I don't know if it was like a new Sarnafil type roof or like some new exterior system. And they needed to understand how we did it on the last project that I was building. I was like, here's the photos. Go check it out. and they just went through all the details. one by one. I think there's an educational opportunity there too. Yeah. And then, also, just from quality standards, you kind of want everything done the specific way, and not everybody has the time to be everywhere at once. And this kind of helps at least scale you, right? Like your billable rate. It's probably gone up.

Right. And like you would love to be in every single place that you can be, but An intern walking around with the camera might be able to scale you when you start going on five, six different projects. Yes. Scaling, yes, for sure. Yeah. For sure. I think the other cool thing I was thing about too is like Comparing Like, um proposed to what was actually built. You know, the thing that I sent you guys the other day with like. Uploading 2D or just like prints, and then AI being able to basically. like render a 3D Building on demand right now. On demand with all the finishes, and like you can move stuff in real time. But then taking that. and comparing it against what was actually documented.

To think about, you could be able to like say, all right, well, this was what the architect proposed. Is this what was actually built? Now, what's interesting is. I think right now a lot of people are in this. this like chatbot phase where you ask it a question, you get a response. But the direction that I think is going to be the most exciting where a lot of people in your shoes are going to say, ha. Maybe I didn't 100%. like understand what the capabilities were. is when it starts creating actions for you and having inference or at least pointing to things. You know, when things start happening in the background for you is where you're going to feel like when it starts automating and actually.

Proactively telling you what you should be doing. Right. Like you walk through, forget how it was built versus how it was designed. If you're walking doing walkthroughs during it and it's like, hey, there you know, there's supposed to be some Vents here, some lights here, or the wall was not supposed to be this long. And then it gives you that as an action item proactively. That's incredibly helpful. If you know where you want to go, right? You're A, I want to get to Z, it's very easy because you can say, great, infer everything in between. I think very often people are like, I'm at A. I don't know where I want to go. Do something for me. And they get angry when it makes them the wrong thing. Right. Um, And unless you can be very, very specific. or explicit. in what you want. Then I think people have a hard time with the output. Right.

Yeah, you just gotta be you just you just need to make sure that you're very succinct in what you are asking it. Well, the other thing that I've been doing recently, and we talked about this, is like using one AI Like to write a prompt for another AI. I've tried that. Yeah, you know it didn't go well. It didn't go well. I had uh I had uh Gemini try to write something for me to set up like my first Claude. I wanted to make Claude my executive assistant. And it was like, this is just made up things. I can't do these things. I'm like, okay. Well. Give it a nice try. I'll let you know about this. You just have to free agents within Claude to do stuff for you. Like I've created a Claude agent. So like I get so many emails from my kids' school, from like the PTO and like just general, I get emails. You can't be bothered with this kind of email. I get legally emails from like the teachers, the principal, the superintendent, plus the PTO stuff.

So I created a Claude agent to basically take all of the emails that I get on a weekly basis from my kids' school and then generate summaries for me. So I don't have to like go in and read them all because I get like 12 emails a week or texts and all that stuff. So simple. I think. Saves me time. Does your agent. Sorry. No, no, no, go ahead. You'll have to share I know this is a bit of an aside, but like Is just running all the time, or just schedules to run all the time? It runs. I'm so confused with agents, and I sell it. And you can have it run as much as you want. You can right, you can run it. But it's not just sitting there actively. But it could run in real time in the background and like scanning your emails. Well, when you create skills and stuff, you don't use as many tokens. I'm teaching a class right now. My first class was this past Friday. It brings up interesting points about what do you teach? How much AI do you let the students use or not use? And what we're leaning on a lot is presentation skills.

And so there's an RFP, where it's a project management class, and there's a whole set of construction documents and an RFP that was provided to accompany these drawings from the owner. And like the students are broken into teams, and their job is really over the semester to like build a response to the RFP. But like, it's not enough just to like hand that in anymore, because like, sure, Claude did that. Like, I want you to stand and deliver. Like, I'm going to ask you questions. I'm going to put you on the spot. You're going to have to give me context and color. Like, it's not enough to say anymore. Like, it's a three-story building. It's like, well, what happens over, you know, a column line? Why six? You know, like, how come the roof is doing that? And what are we going to do to waterproof that weird. I think that's what's one thing that's going to set a lot of people apart when like with the AI stuff is just

communication skills. and presentation skills and just being Socially not awkward. Well, yeah. Cause there's so many people just that hide behind a screen now. And they just can't be in social situations. And if anybody, I was gonna say, if anybody is listening to this, and if you want, if you're still debating where to go. As far as a career, whether it's development, et cetera. Get a job.

For a construction company, construction manager, if you can. Superintendent Project Manager. Uh You learn to talk to every walk of life in the same day. Yeah. You learn interpersonal skills. It's not, I think a lot of the old school way of just I get to yell at everyone and they have to listen to what I do is going away in a, in, in a good way, I think. And the amount of, you know, the, my, a lot of my.

Would run off a lot of good handshake deals and respect, and you know, making sure people have runway so they could bring the manpower and things like that. Uh A lot of that you can't you can't vibe code your way into. And I'm glad'cause I I don't think you could figure out a way to do that. So yeah. Makes sense. Amen. Well I think A pretty good episode, we've covered a lot of ground here. Yeah, uh, thank you everybody for rating, listening, subscribing, uh, and following our YouTube channel. We appreciate your support. And I'll catch you on the next one. Congrats. Yeah. Congratulations. Later, four more years. Bye.