How AI and Computer Vision Are Changing Construction Monitoring 

Podcast Transcript

Steven Rossi-Zalmons:
Hello there and welcome to the Construction Revolution podcast. My name is Steven Rossi-Zalmons and here on the show we explore the latest trends, technologies, people, and organizations that are revolutionizing or disrupting the construction industry and are changing what the industry will look like tomorrow. Today we’re excited to welcome N K, the co-founder and CEO of Track3D, a reality intelligence platform that is redefining how construction teams monitor progress and understand what is happening on site.

With a long track record of founding and leading technology companies and deep experience in engineering and computer vision, N K has helped position Track3D as a leader in accurate real-time construction visibility. Join us as we explore how Track3D is giving project teams clear insights, reducing rework, and helping the industry move towards smarter, more predictable project delivery.

Hi, N K, how’s it going today?

N K Chaitanya:
Hey Steven, thanks for having me. Really good looking forward to this conversation.

Steven Rossi-Zalmons:
Yeah, me too. Yeah. Thanks. Thanks for joining us and yeah, let’s jump right into it. So why don’t you tell me about you and how you started Track3D and where you got the idea for the company.

N K Chaitanya:
Perfect. So firstly, let me introduce myself. Steven, I go by N K and I’m the CEO and co-founder of a company, Track3D. We started Track3D about four years ago. And when I say we, it’s me, Vineeth and Kiran. We are the three co-founders of our company who have known each other for more than 20 years now. In fact, me and Vineeth have been working together in different startups of our own over the last 15 years. And all three of us have been working together since the last nine years now.

The idea of Track3D is something that I’ve been always working all throughout my life, although not applied to the construction domain. Like technology such as 3D vision, AI, robotics, IoT and extended reality, especially the convergence of these technologies is what’s been particularly exciting for me. Started out in the healthcare domain, doing neuro rehabilitation, but then furniture augmented reality, smart wearables, home automation, industrial automation.

To our last startup, when all the three of us got together, we were creating digital replicas of entire cities. And that’s how we got introduced into the world of construction, because cities do a lot of construction, and we were helping them track what is happening with respect to the construction sites. Then we realized that construction is a very unique problem because every job site across the globe, even mostly today,

Data from the ground is predominantly being recorded in manually. And the technologies that exist are not simple enough, fast enough, accurate enough, or even cost effective enough to be deployed at massive scale. So we thought that is a unique opportunity for us. And we were uniquely placed given our experience of last 20 years working in these technologies to build something that is super impactful for the job site. And that was the genesis of Track3D. So I think…

It was destiny. was all these technologies that we were working towards and we were at the perfect timing as well to actually create this impact.

Steven Rossi-Zalmons:
Right. Yeah, that’s great that you’re able to leverage your previous experience and keep working with the same team. And we’ll get into your experience a little later because I’m curious about your long standing entrepreneurship. So before we get into that, can you tell me a bit more about what exactly Track3D does and who are your main users and customers?

N K Chaitanya:
So Track3D automates construction monitoring. We are obviously an AI first platform, but before going into all this jargons, right? In a very, very simple way, if I were to explain you, we help you answer four questions. The four questions are how much of work is done? What is that work done? Where is that work done? And is it done right? And we do it automatically without any manual inputs.

So for us, the inputs are a moving camera. The camera could be a drone, a 360 camera with a person walking by it, a robot with cameras mounted on it, laser scanners, any visual source on any vehicle. So the data source doesn’t matter. The vehicle that is carrying it doesn’t matter. All we say is we need a moving camera so that we can quantify and qualify progress for you.

And the main users of Track3D today are, we started with general contractors, but now we have trade partners as well as owners using the same information to get overall operational insight on what is happening on a day-to-day basis and coordinating, communicating and planning their activities around all of this.

Steven Rossi-Zalmons:
Yeah, that’s great. So to give everyone a better understanding, so what would the traditional process look like of monitoring this progress and then how does that change with Track3D?

N K Chaitanya:
Absolutely, Steven. So even when we got started, right, our assumption was things were a lot more automated, things were lot more streamlined. But my original experiences of going from job site to job site were things were mainly happening either on pen and paper or Excel. These were the two predominantly used tools. There people walking on job sites, eyeballing the work that’s done and saying,

Okay, 10 % of ductwork is complete or 50 % of painting is complete. No one could actually because it’s super inefficient and it’s very manually cumbersome to go ahead and measure everything on a day to day basis of everything that’s being done. So we saw every job site even today, majority of the progress that is happening on site gets captured manually. Now manually captured data is either incomplete, it’s inaccurate,

it’s inconsistent or inaccessible. These were the four problem statements that we sought. And we said, you you know, it takes away the joy of building, right? If you have to only report, measure, you know, track everything that’s happening, that’s not real fun to do as well. And that’s why people are inefficient doing it. We said, can we apply our expertise in this technologies to automate those processes so that people can focus on what they love doing the most, actual building?

Because you think about it, right? If I can draw parallel, like building a Lego set, you have drawings. Those are the instructions to build. You have your Legos, you go ahead and build it. But say if I asked a kid to document after every step of what you have done, like yesterday to today, what you have done, document everything, it would be so painful. No one would love to do that. So we say people should not be even asked to do that.

Why can’t we automate the process? So that’s precisely what Track3D does.

Steven Rossi-Zalmons:
Right, yeah, that’s a great explanation. Yeah, definitely. We can see how, yeah, no one likes tracking everything and it makes life a lot easier. So I’m curious, like you mentioned construction, typically not a lot of, especially in this process, not a lot of technology before you guys came along. So when you’re pitching this new tech to contractors, do they have concerns or pushback or resistance sometimes and how?

what are those sort of common things and how do you overcome them?

N K Chaitanya:
This is something I say quite often, Steven, basically. Construction industry has unfortunately the bad reputation of being technology laggards. I say people still use Excel, people still use pen and pencil, not because they’re not technology forward. In fact, average job sites that we have worked on, know, the companies work on 200 to 300 applications on average. That’s so much technology exposure that they have.

They try to manage all the solutions together. But the problem is not them being not tech forward. The problem is the solutions were not being good enough so that they can be actually deployed in massive scales. Like reality capture, for example, has existed for quite some time. It existed from the time you would take pictures from site and actually put them onto floor plans, arrange them on documents or folders. So the reality capture has existed for quite some time.

To my original point, what we found out was why the massive adoption was not happening because these technologies were not simple, fast, accurate, or cost effective enough to be deployed in massive scales. And that was our window of opportunity. So we don’t put it on the industry at all. The industry wants to solve this problem. And that’s why they are very eager for something that actually works. But that actually works is what

people have had bad experiences on. They have burned their hands with people over-promising and under-delivering many, many times. You have to give it to them, right? As an innovation manager at a really good construction company, I am not jealous of their job at all basically, because they get pitched by vendors every day promising the sun, moon, and the stars.

But when they actually look into implementation, it’s complicated implementation, takes in so much of their time and effort, and the value is still not there. So I think the barrier that we have to cross while pitching Track3D is just getting them to trust it, actually, that there is value and this doesn’t need too much of your time. It’s extremely simple to implement, and it seamlessly fits into your workflow. We don’t ask them to do anything new. You’re already capturing the job site.

let us leverage the data and make it more productive for you. So yes, that is why we are seeing some good, really massive traction and scale. People start with us on one project and they very soon scale it to even 100 projects plus. So it’s about just proving that value and creating that trust.

Steven Rossi-Zalmons:
Yeah, yeah, absolutely. I think that’s, it’s an interesting point. And you’re not the first person that I’ve spoken to, to make that like, definitely construction has a reputation for not adopting technology and being slow, but it’s not for the lack of trying and lack of desire. It’s a lot of people are trying to force fit other solutions into the construction industry or to your point, they’re too expensive and over promise and under deliver. So yeah, it’s interesting that

You’re not, a lot of people keep saying that, but it’s good that there are solutions coming out that now are being adopted and pushing the industry forward. So going to sort of touching on some of the points you made earlier, like you’ve started a lot of tech businesses in your career and even with some of the same team that you’re working with now. So I’m wondering how you found the construction industry getting into the industry and

getting on site and also just talking to people.

How has your experience been starting in the construction industry and some of those other industries you worked in in past?

N K Chaitanya:
Right, so as I told you earlier, the convergence of these technologies like 3D vision, AI, robotics, IoT and extended reality, right? This have always been something we are super passionate about. And in fact, our core team at Track3D has come from all our previous organizations that we have worked like starting from 20 years back. And so we are extremely proud of the fact that some of the people actually who are still working with us and we are partnered up to actually solve on this.

So technology side of things, we always had a good grasp. But what made it particularly applicable for the construction industry, especially at this time, was the advancement in technology itself that happened. Like there three major things that actually advanced, right? One is the core sensor technology. Like you look at cameras and sensors right now, they’ve become like much, much better than what they were even one year ago, lived 20 years back. So you have…

super like you know a 360 camera for example with the latest ones that are coming up they are high resolution they capture a lot of data quickly and they are low cost actually like it’s now practical to have a 400-500 dollar camera on every job site irrespective of its shape size or geography so that is the accessibility piece of it but that is the hardware second thing is you know advancements in robotics and technologies as well like drones

or even the four legged robots right now that we have. These also are becoming more and more accessible every day. Like you now can get a robot for under $5,000 that is already practical to deploy on a job site and takes a person’s job of capturing data on a daily basis. Drones have matured so well actually over the last 10 years or so, right? Now for doing your earth work measurements, your survey and all of that.

it becomes super practical. Again, it was not possible even 20 years back. So that is one more level of advancement. But the number one has been AI, right? You know what is possible with AI right now, but the way you can actually segment, track, measure different kinds of objects without requiring a lot of pre-training is what makes this information again super accessible and the cost prohibitiveness is off.

Like typically, if you wanted to build a machine learning model to identify pipes, for example, you had to actually build, collect so much data, manually label it, annotate it, then create a model so that it could work. But now you can be AI first, where you don’t need to do all those steps. And you can automatically not only say it’s a pipe, but you can read the actual text present on that pipe in real time.

and actually give you all the specifications of that. So this was not even possible like, you know, six months or one year back. That rise of advancements that have been happening is what we have been betting on, you know. We say this is the right timing because the technology maturity is there. This makes it accessible for every job site. And you can actually create a massive impact at scale.

So irrespective of whether it’s a 10,000 square feet retail unit or a million square foot data center being built, you can have this technology providing value at unprecedented scale, at unprecedented cost and accuracy. So that balance is possible now. And that’s what actually prompted us to get into now because we thought timing was right and we had the right expertise working in this technology. We had seen this entire evolution. So we said,

the we have we are at the right place right time and if we get the right partners to work on this one I think we can create a massive impact and that’s where we got extremely fortunate as well to be working on this.

Steven Rossi-Zalmons:
Yeah, yeah, it’s great. It’s like you’re analyzing and being able to leverage all the experience. So speak on that note as well, like from your previous experience, what have you learned from those and especially in terms of like making a startup and scaling up so quickly, like with you’ve received a lot of funding and growing super fast. So how did you learn from those previous experience to be able to grow Track3D so quickly?

N K Chaitanya:
Well, mean, that is something I can almost write a book on, things not to do in a startup. So we had such great experiences, such great learnings all throughout, especially having done both hardware and software startups. So we got a unique perspective. how do we approach problems? The number one thing that we learned was starting with problem first. Like if you divide our career,

The first 10 years of her career was only about building cool tech. We didn’t care about what problem it solved, me and Vineeth. We both were like technology nerds. Like we were building gesture controlled smartwatches that could control drones and robots. Now don’t ask us why. We did it because it’s extremely fun to do it and we were enjoying it every day. Now what we realized is that while that is very important, you do something which we are extremely passionate about.

But more important than that is actually solving a problem, actually solving a pain point for the customer. That is how you go to market and the actual product market fit matters. And we had that learning in a slow way, like we would have wanted it to be happening faster, but it gave us a lot of learning, right? And now we also realized that to get to the problem statement, the only way you can do that is actually meeting at

where the problem occurs. Like you can’t be hypothetically like our problem was at the job site. So I can’t be sitting or reading through theory or going through encyclopedias about construction and trying to learn about it. The only way was getting to the ground, shadowing people, asking the right questions, asking questions generally, there are no right or wrong questions actually. So asking questions, being genuinely curious about things and take

technology out of the equation completely. Technology is a means to just solving the problem. Most important thing is how well do you understand the problem.

that became important. The other things that we also learned was from the go-to-market part. It’s very important at the early stages that we have very narrow focus. So we were very deliberate when we started with Track3D as well. We didn’t want 100 customers who like us. We needed five partners who would absolutely love us. They would say that, hey, I cannot actually survive without Track3D.

was our core intent. wanted to make that and we like, you know, this came as a deterrent to when we were pitching to investors, but we had this clarity that instead of making 100 people who like us, let’s get five partners who absolutely love us, who become our references, and then we can grow from that. And to select that five also, what we did specifically was we went after the top contractors who had tried out every competing solution.

because that is the hardest ones to get, right? They’re already enterprise with these solutions and we not only have to make them shift to us, but also scale with us. So I think these were some of the hard things that you had to do and we were very, very deliberate about it. And obviously, luck had to pay a good part of it. But I think being strategic about it, being deliberate that this is what we have to do and we’ll meet them at the source of problems.

Kiran, Vineeth, that’s all we do even today. We go to job sites, we try to learn as much as possible. And that has helped us actually to scale to this factor so far, Steven. I think we’ll continue doing that. This is something we really, really enjoy as well.

Steven Rossi-Zalmons:
Yeah, that’s great. Yeah, it’s great to hear and thanks for the insight into the learnings. yeah, it’s great to see how you’re scaling and clearly at work this time. I’m sure maybe you’ll continue to start even more businesses, but there’ll be more learnings ahead. So speaking of the future now, how do you see

Track3D evolving over let’s say the next five or so years and even shorter term is there anything that you’re working on now that’s really exciting that you could share.

N K Chaitanya:
Absolutely right. And the thing that we are excited the most about Track3D is, know, becoming the operating system of an asset, a physical asset across its entire lifecycle from design to construction to operations and maintenance, eventually to demolition. So I see we becoming the central operating system for every physical asset. We started by being the progress monitoring part of it.

which was the verification part of what is installed, is it installed right, and is it installed to the right timeline as well, so that we can help get projects completed within budget and schedule. But we believe this core piece is the central part that binds all the other elements together. And our focus for the next five years is quite a long horizon. But the near future, what we are…

really excited about is how our engine actually ties up with all the existing systems and even uncovers the next level of insights. And this is through the AI agent framework that we are building. Like, are doing the actual production tracking, but say we have the manpower data combined with it. Now we have productivity. You have the material data with it. You have inventory management. There’s so many things that can play together and this core data can unlock.

that is super, super exciting for us because this is extremely valuable to all our partners, all our customers. And that is what we’ll have a relentless focus on in the short term. How do we unlock more use cases and create more value for them? But eventually the game is how can we actually become the central operating system for any physical asset across its entire life cycle, not only during construction.

Steven Rossi-Zalmons:
yeah, that’s a, that’s a big goal and an exciting vision. Yeah, that’s, it’s interesting. mean, even just unlocking all the, the stuff with partners and other data. Yeah, that’s definitely across construction, construction, especially like being able to integrate all the data is becoming more and more important and they’re looking for that. that’s great. So that leads me kind of right into the next question, which

What do you see the future of a fully digitized and well-monitored construction project looking like?

N K Chaitanya:
and

they are excited to be working towards building that future. So that is what we are doing right now as well, right? Is we believe.

people and robots coexisting and it being so efficient that know people are performing the task they precisely know what has to be done when it has to be done where it has to be done and robots are aiding them as agents doing that so I see AI robots and people coexisting together to create physical assets in unprecedented scale unprecedented time

and unprecedented quality. So these three parameters coming together, I think both from the short term and the near long term perspective, right? We are super excited about what is possible with that. And we believe Track3D can be the central operating system on making that possible. Robots, human beings, and AI engines working together to unlock this level of productivity.

Steven Rossi-Zalmons:
Yeah, absolutely. So then sort of a couple more questions here and then next one sort of as someone who’s new or newer to construction but comes from a you mentioned like big tech robotic background what were when you came into the industry what are some other innovations like apart from Track3D that you’ve seen that excite you the most in construction?

N K Chaitanya:
Right now, again, the fields that I have been working on are super exciting, like robotics, right? I see robotics doing drywall installations actually, end to end. And you know, I see because of the labor problem that exists right now, how these robots can aid people in actually helping things get done faster. I see robotics also being exciting in data capture mechanism, like drones and indoor robots just collecting data so that people don’t have to do it.

These are some of the near term things that are super exciting. Long term things again are like, know, manufacturing and, you know, going on those lines for construction. Can things be massively automated at scale? These are again, problem statements that have not been solved yet, but if you solve it, there are some massive impacts that you can actually create. 3D printing was one more thing which is exciting, but has not scaled to its true potential, right?

So these are things that you’re always constantly looking up to and working with some of them, partnering with some of them. And obviously you cannot stay away from AI and the use cases that AI can unlock for the construction.

Steven Rossi-Zalmons:
Yeah, yeah, absolutely. Yeah, it’s interesting from, especially with your robotics perspective that, yeah, there’s so much new innovation on that side that’s really exciting. So to wrap up here then, if people are interested in learning more about Track3D and getting started, what does that sort of initial process look like?

N K Chaitanya:
Well, you can always look out at LinkedIn. LinkedIn is where we are most active at, at Track3D. And you can always reach out to me as well at nk at Track3D.ai. So happy to connect on that and have further conversations because Track3D we say we are the most simplest solution to implement. it’s the fastest time to value the most accurate solution possible. And it’s extremely cost efficient to scale irrespective of the scale of the project.

Irrespective of that, I am a very curious human being. I always like to have conversations, learn, and see if we can actually work together in enhancing the productivity of the overall construction industry. So any conversation right from a student to a CEO of a company, I’ll be very happy to have. I’m always eager to learn, so happy to connect with anyone. And thank you for this opportunity, Steven.

Steven Rossi-Zalmons:
Yeah, absolutely. Yeah. Thank you for your time and your insight. And it’s great that you’re so open to connecting with everybody. So yeah, thanks again for your time and it’s been great learning about you and Track3D.

N K Chaitanya:
Thank so much, Steven It was great having this conversation.