Andy Verone:

Good morning, good afternoon, good evening, and welcome to the first installment of True Talks, brought to you by Contruent. My name is Andy Verone, and I serve as the Chief Strategy Officer for Contruent. I will be your host for today’s webinar. Today, I’m so excited to be joined by two colleagues, one past and one current. I also think of these fellows as my friends. So, Rich, Burcin, welcome to True Talks. Thanks for helping me kick off this webinar series. Burcin, do you want to lead us off with a brief introduction?

Burcin Kaplanoglu:

Sure. Thanks for the invite. Great to see you, Rich. I’m excited to actually be talking to you guys. I think we have a lot to cover. We have a lot to talk about. And I think this is going to be fun.

Rich Humphrey:

Yeah, again, thanks, Andy. I’m excited to be here at the first True Talks. We’re going to be talking about this many years to come as the kickoff of what should be a pretty exciting series of events.

By way of introduction, I’m the Chief Product Officer here at Contruent. But I’ve been an industry technologist for about 20 years and in the industry for around 30 years. I’ve worked with companies like the Army Corps of Engineers, construction research labs, and Clark Construction, to name a few. So, I’m excited to be here and talk about this really interesting topic.

Andy Verone:

Thanks, guys, Rich and Burcin. I really appreciate you joining us. And I’m excited about this. We had a last-minute lineup change. And these things are hard to predict, even with the best-laid-out plans of the schedule. For those of you joining us, Dave Anderson, a friend of mine for over 30 years and a long-time colleague, had a last-minute customer conflict and wasn’t able to join us today. However, he did help us with building the content, and Dave will join us on future True Talks. So, we appreciate his willingness to be part of this.

Guys, if I think about the design criteria that we laid out for these, one, it’s really to go deep on a topic. And it’s really your point of view, your view of what’s happening in an industry that the three of us are truly passionate about. I mean, I’ve been involved in engineering and construction for just a little over four years. Rich, you talked about your background. Burcin and I have had the opportunity to work together; we’ve been colleagues and friends for a little time. So guys, really be authentic, which is what we’re after here. And Burcin, I was going to poke fun at you being in your autonomous vehicle driving into downtown Chicago. So you’re definitely being authentic, my friend, and I appreciate that.

Lastly, we want these to be brief. We’re just inundated with content, so it has to be short. They’re going to be conversation starters, and then we’re going to continue those conversations on our various social channels. So, a big thank you to everyone who’s joined us live. And then, of course, we’ll have this as a recording.

Guys, let’s get started.

Andy Verone:

Burcin, it would be remiss on my part if I didn’t ask the first question to you about the amazing event that you and your team hosted at the Oracle Industry Labs. It’s the Contractor Cup Drone Challenge. On all accounts, everything I’ve seen on social was just what an amazing event. What are your takeaways from that event? How did we harness AI in that Drone Cup Challenge? Burcin, over to you.

Burcin Kaplanoglu:

So here’s what we did. We hosted some pilots. We set some targets for data collection. And when I say data collection, you know they’re flying drones. We originally based it on NIST standards for data capture for drones, but we really modified the course. Again, we talked about what criteria we are after. Time was one, but quality was as important or probably more important because we also don’t want to put pressure on them to rush their data capture.

Imagine this, there’s a course drones fly. And they take images. So, this is the data that’s collected. But what they’ve taken images of is some of the targets are QR codes, some are photos, and some are material, like equipment and information. As an example, they had to go to an elevation and take a picture from a window. That’s not a reflection there. And it was an HVAC equipment spec that actually was on the side of an HVAC equipment, you would see. Why we did that is because we were trying to see what it takes actually to collect the data, how long it takes, and then really see the skills. We also judged the pilots based on their skills. Now, how does this tie to AI? Eventually, it’s all about data.

Some pilots did an amazing job with the quality of the images and not reflections and some stuff, like they’ve taken the QR codes, but they miss one corner because it wasn’t fully aligned when if you missed the QR code, one corner it’s not going to be right. So that’s one. The same thing with computer vision. Like if you take a picture of a label flying in the air that’s feet above your head, you’re doing this, and you don’t even have line of sight. Because you have some line of sight. But it’s not that clear because you have to look from a small screen. It’s really important how you capture the data.

What we have learned is skills vary significantly. And even though we gave the same course to the pilots, even how which route they took and how they did it. And each time we’re humans, we learn from it. So every time they learned and did one flight, they would do differently the next time. So again, I think it goes back to quality, quality, quality data quality really matters for running Vision Services, running anything that we’re going to use for artificial intelligence. That was one of the key takeaways.

Andy Verone: 

That’s great, Burcin. It looked like an outstanding event. And I know you guys were testing and capturing just amazing amounts of data. So appreciate you sharing that with us.

Rich, if I throw it over to you, I mean, your job every day is building products. You’ve done that now with multiple firms. You’ve got great experience. What do product leaders who serve capital-intensive industries do, and how are they harnessing the capabilities of AI? You can talk specifically about what we’re doing at Contruent or just what you’re seeing out there in the industry. If you can provide a couple of examples, we appreciate that.

Rich Humphrey:

Yeah, I mean, I can tee off what Burcin was describing too. I mean, as a technologist providing solutions into the industry, there are a lot of potential applications for AI in the construction space. And it’s all around data, as he said, whether it’s data capture, data management, or data insights, and I think we see that AI is really currently an enabling technology. Whether that is using some of the underlying tech stack related to natural language processing, or generative AI, or that becomes at the forefront right now, in terms of, uh, some of the things that are going on, but also just basic logic systems. And as Burcin said, computer vision, just to name a few of those underlying technologies and how we leverage those in our solutions to make data capture or management and analysis more effective, is really how I see it and what our strategy really in Contruent is so. 

So, some examples might be in a project controls environment, which is what Contruent is. We are looking to figure out things like how do we better predict project outcomes. So, for example, if I’m doing estimating or scheduling or specifically cost forecasting. How can AI actually be applied leveraging historical information and trend information to do things like better predict cost forecasts? Today, Contruent provides out-of-the-box, tons of calculation methods and ways to time phase data. But we’re looking at ways to predict your costs based on the data that you’ve had for previous projects. 

We also look at it as a productivity assist. Now there are lots of workflow wasn’t tasks that construction folks are engaged in constantly. If we can provide a better interface, and we’re leveraging AI, for example, like a chatbot, as the interface for them to assist their workflows, and to pull up information they need to rapidly get through the process in an effective way. That’s great. And we’re looking into that. 

And then just things like data insights. We leverage in the industry a lot of well Excel to do data analytics, but also things like dashboards, reports, business intelligence, but what if I could just use an AI to ask it a question about my projects, like, hey, what if I, if I want to show me all the projects or tasks or activities that have the risk of being behind schedule or behind cost, and that those insights are just surfaced to me through natural language or a simple kind of Chatbot? Again, those types of enablement to our existing workflows to make everyone more productive and ultimately make the data more reliable and trustworthy and higher quality are the types of things that we’re looking at at Contruent and kind of resonate with our customers. 

Not to mention, kind of last one, which would be things we’re looking at around performance optimization, as Burcin described, how can we use AI to better track progress and feed that into our system through things like computer vision and BIM modeling?

Andy Verone: 

Yeah, Rich, really appreciate that. Burcin, you guys have two very different viewpoints. So Rich is off building AI capabilities for the Contruent customer base and future customers. But Burcin, you are always out there testing. You’ve got these wonderful facilities in the US, UK and Australia. What are those AI capabilities that you’re seeing that the industry should really be excited about? What are those ones that you guys have in the labs now that you’re just seeing that could be a game changer? 

Burcin Kaplanoglu: 

Well, I’m gonna tie to what Rich said. One is, first of all, let’s classify what artificial intelligence in massive fields. First of all, machine learning has been out there for a long time. It’s a great field. You can take deep learning, you can do a lot of data, use a lot of data, and come up with some predictions, as Rich said, so that’s not a new field. It’s been out there for a long time. And, I’ve seen a lot of adoption of it in, I’ll say, last ten years. 

The second one, I would say computer vision. So that is about recognizing things, counting things. We do like measurements and things like that progress reporting, so it is getting there. The challenge with that is accuracy often. It comes to like 90%. The last 10% is pretty challenging to actually achieve. And everyone in the industry who actually deals with the tech knows this. So, I’m not saying something new. It’s just the challenge of the last ten percent is pretty complicated. And we, as engineers, expect 99.9999. So that’s our expectations from machines, we have that expectation. So it is getting there. But I’ll say it’s been a journey. 

And the next one as Rich mentioned large language models, like ability to actually talk to your data, call it like, talk to your cost, talk to your schedule, talk to your plan, talk to your budget, like all those things, I think that’s a totally new interface in some ways that’s how I see it. And it’s really opening up new ways to actually run our projects. 

So if I were to classify those three, I still see there are massive opportunities in machine learning today. What can we do with the existing data and how we can actually harness it? There is a lot happening in computer vision in manufacturing. So, at the lab, we don’t just work with construction, we work with nine industries, and I’ve seen so many things that Oracle technology help our customers in manufacturing, whether it is high tech, whether it’s automotive, whether they’re producing food, like any mass production. So there’s a lot of technology in the vision that’s already being deployed in these factories that is definitely applicable to us. 

And then I think large language models are, that technology also is not new. It just didn’t happen in a year or two. It’s been evolving. It’s been becoming more mature. I think the biggest change in this time was public actually had access to it before. Before, I’ll say a specific industry use cases were actually applied. So that’s what created the excitement, and that continues. But I really think that all those three fields have tremendous opportunities for years to come.

Rich Humphrey:

If I can just jump in on that, too, around things like language models and natural language, it’s all come to the forefront. But a lot of these technologies are being used in other industries. You mentioned manufacturing. And what’s happening from our perspective as we build construction technologies, tuning those into the construction space, because for what it’s worth, the language that we use in construction, and the objects that we identify, are all getting trained now more than they have in the past. So that’s where the software vendors, technology vendors, have a good place to play is take all the work that’s been done over the last 10 years in these technologies and tuning it to the problems that we have in the construction space. 

Burcin Kaplanoglu:

And now that you say it let me give you a specific example. So I have seen this example where you can like materials. You call it drywall, glue board, there are so many names to it, right? So we use different things, and they show up with different names and schedules. You can easily get a large set of schedules and figure out who built them and which country they’re coming from. Again, we have you can easily also see that some people provide a lot of detail. And some do not like it is again. The human factor is so important. And data is unstructured. So I think the keyword is unstructured. Because a lot of people, when they think about this, they think that oh data has to be structured. It needs to be in a certain order. Well, that’s not really true anymore, we can actually really use unstructured data and make sense out of things, which is pretty fascinating. If you think about it, you actually have to do that. 

Rich Humphrey:

If I could piggyback on that with an example. I know you’re also really familiar with Burcin, is in the computer vision space. This is something that I also built in one of my last software’s. It’s. Basically, we take tens of thousands of just photos, let alone create reality models out of those or doing drone passes to build models. And the biggest problem is that those photos are largely unstructured. And if you want to go if, heaven forbid, want to go need some backup for some report or a change order or heaven forbid you end up in litigation, being able to access a particular photo or event is important. 

One of the simple things that is widely usable now is just indexing photos based on the information in the photo itself because the AI can recognize pipes, rebar, and construction equipment, and we’ve been able to OCR things off of photos for a long time, all get all enabled via AI to index that photo in a way that’s quickly searchable and can be organized and more structured without human interaction. 

So it’s a very simple use case that is actually something that most users trust that type of AI, whereas in other cases where there’s an engineering output, they’re a little more skeptical. 

Burcin Kaplanoglu:

And I’m gonna jump in one more time. I know this is a conversation, so I’ll share a personal experience. Now that I’m in the car, this is even more relevant. I had a small accident a month ago, so luckily, I was driving 10 miles an hour nothing happened. But the issue is a clean process. First of all, going to the website guides me through the process. I have seen so many places in the insurance process that actually use computer vision and other algorithms to actually do it. It didn’t actually force me to download an app because, most of the time, it does. It didn’t. Everything was native on the browser. At the end, it actually had me do a 360 walk around the car because I know where that’s going. We all know where that goes. You do 360. It guided me step by step on how to do the 360 accurately and capture the right angles so that they can easily build 3D Point Clouds, like from that data.

Now, talking about an industry where our audience is asset-intensive construction engineering, I’m talking about insurance. I’m talking about personally just going through a website, and then look, it did these things like a picture of the insurance of the other driver, well, it automatically recognized everything in that picture. And then, the other thing is, I took pictures of the car. Well, it mapped the car, it looked at the underlying data in the photo, and it actually put in the right location on the map. So it used the geo data that’s in the picture. Like all these things, typically, an example, I didn’t remember the street name. I don’t know what street it was. But because the photo is taken, it knows exactly the location of the photo. From the photo, the underlying data metadata that’s in there. It just actually said to me, is this the right location? I’m like, yeah, it is the right location. You’re absolutely right. It took me a second. I had to zoom out and say, “Wait, okay.”

So, many industries are using these now. And I think the example I’ve given in insurance is a personal experience. But we are all seeing this. And we might not recognize this happening, but it’s already happening around us. Now, how do we take all this to bring engineering construction? There are so many opportunities, I don’t even know where to start. So I think that it is becoming very relevant for us.

Rich Humphrey:

And if I can add to that, too, I mean, the analogy for construction of your use cases, someone, just an operator, just took these photos of construction equipment to do general equipment inspection at the job site. And then the other thing that we talked about is data capture, but in that example, you’re talking about, it’s using the phone sensors, is using the photography, as using the GPS location, there’s a lot of kind of IoT, and other types of things going on, and other technologies that are also just all connected together with AI and other things, which is one of the other interesting, exciting things going on, is we’re talking about AI, but it’s just part of a lot of other enabling technologies that are coming together all at the same time.

Andy Verone: 

Yep. For sure, guys. Listen, great, great conversation, you make my job really, really easy. Listen, the CPM schedule is really important to both of our organizations. So if you think about what Contruent does with cost, schedule integration, the P6 schedule is just a vital part of that version. You and your team have the world’s greatest CPM scheduling capabilities that you offer to customers all over the globe. Rich, what are you guys seeing in scheduling? How are you using and applying these types of technologies with the schedule?

Burcin Kaplanoglu:

I’m going to show you a specific example. So we shared this in September of ‘23 at Oracle Cloud World. When you receive a request for a proposal, you have to click the turnaround and actually create a schedule. Typically, the way it works is you have an estimating department, you have a project management department, you have schedulers. This is typically a three to five-day exercise. But the goal is you need to actually really get quickly a schedule and the cost and all those things tied together to put in your proposal. 

Well, what we have done is, again, similar to what Rich said, using multiple technologies, not just large language models. So that you’ll be able to go to a chatbot and actually upload to request for proposals, so it synthesizes and understands what it is. The chatbot asks you specific questions. And the structure type, certain things. As an example, it recognizes in the RFP that says there’s a podium, so it reads through it, it actually recognizes certain things, asks you questions, and based on the input you provide, it actually can generate a schedule. 

Now, this schedule is not something you’re going to take and just use it right. This is your starting point for a scheduler. Typically, when we asked the schedulers what they used to do typically today is they would literally go to a piece of schedule, similar project, and start changing things right. That will take a couple of days to basically use an existing schedule to update and create a new one. In this method, something can take you three days, potentially, to take you three hours that’s the change. It’s giving you a good starting point. Again, humans are still involved. There is human factor. Schedulers are very, very important. Our knowledge about projects is really important. But it’s a good starting point.

So this is a great example, I would say, that boosts productivity and also helps to move things along faster. You’re responding to an RFP. Every minute counts, it’s really important to do your due diligence. So you can spend less time thinking this way, like the example Rich gave. You can spend less time trying to format it and actually align it with the requirements, and more time looking at how I can deliver this project faster. What are the techniques and methods I can use to actually do that? So that’s the example, I think, is really, hopefully, will resonate with the audience. 

Rich Humphrey:

And I think if I can add on to that, too, it’s similar if you’re asked to start corollary around bidding and estimating because you can apply some of the same logic and create quick estimates or bids. And it’s not just to make it happen faster and have that be the result of a good start. And we see, I see kind of the value being, like, if you’re going to do an estimate in the past, you do your quantity takeoff, you build estimates. And you’d usually have two people doing the same estimate at the same time. So you can do quality verification and make sure that you don’t have human errors and omissions in the process. 

By doing it twice, well, that process that Burcin described either for scheduling or for estimating could just be a verification of what a human might do in terms of am I in the right range? Instead of not just relying on it to get the answer. And I see that resonating with a lot of our customers because it means they can be more productive. They only need one person to do it. And multiple people do an estimate. But instead of another person or team building the schedule or estimate just to do validation. 

Andy Verone: 

Fantastic guys. As we’re coming down to the end here, I do have really two more questions I want to squeeze in before the time is up. Guys if we exclude drones, because obviously drones, in my humble opinion, are robotics. And we know the use cases and most large capital projects have drones in use and capabilities. Where do you see robotics going? Burcin, I’ll start with you. You’ve tested various robotics for the industry. What would your prediction be on robotics in the next couple of years? Where do you see it going? 

Burcin Kaplanoglu:

I see massive investment in humanoid robots. Which is when you say humanoid, they form factor. They look like us. And massive investments, and a lot of research done. And again, this is another field that’s been researched for decades. But the changes in the last two years, especially 12 months, have been quite fascinating. And think about it, our job sites are designed for humans. It’s designed for us to walk around and move things around in all this. I think that is my guess that’s coming in the next couple of years, that is going to have an impact. How big in what it is, time will tell. But I see really moving fast. And this is probably one of those that we need to keep an eye on.

Andy Verone: 

Yeah, Burcin, I appreciate it. Rich, any feedback on that?

Rich Humphrey:

Yeah, I mean, I agree. I think the human-like robotics is pretty interesting. I think that the drones and even data capture is a great use case for robotics. I’m gonna spin it a little bit differently, and I’m going to align to my or more my background too around civil projects, where we’ve been doing things like machine controlled machine guided earthworks, but that is an area that’s calling for full autonomy in terms of the driver is really adds a lot of cost to that type of equipment if they can eliminate the driver and have it be completely autonomous, and you basically have a full robotics driven, GPS, Model Driven earthworks solution. It’s very similar to a huge large-scale CNC machine, where you can have the AI actually program the most productive path of earthworks and then let a robotics tool, like a heavy piece of machinery driven by the model and GPS, actually carve out the earth or a CNC program. And that’s what I find really interesting. We’ve been using machine control automation for a long time. And that’s the next evolution in terms of robotics and AI for large earthwork moving projects. 

Andy Verone: 

Yeah, that’s great Rich. Thank you. Guys listen, we are just about at time, I do want to end it with one final question I want both of you to answer. If we were to have this call in 12 months, what would the AI headline be? So, I know you both have articulated that things are moving super fast, but what would it be? What would the headline be in 12 months, Burcin, we’ll start with you. 

Burcin Kaplanoglu:

So I’m not gonna do a technology prediction. I’m gonna do something different. Typically, what we do, I think adoption, my prediction is, there’s going to be a lot more adoption. There’s going to be a lot more embedment of these technologies in the applications. We’re doing it, industry is also doing the same. So I think adoption, we’re gonna start seeing a lot of adoption of the technology itself and not even recognize that actually, that’s the technology. It’s just going to become more and more embedded into our applications. 

Rich Humphrey:

I was gonna say the exact same thing, the AI technology is going to continue to advance very rapidly. And it’s just going to find its way into almost every construct and software in general that we have. But it’ll be very subtle, as users won’t always necessarily know that some workflows are being driven by AI. It will just become a common case. And just very pervasive in how they use software solutions, among other technologies.

Andy Verone: 

Yeah, guys, I appreciate it. Listen, I think for me, this is the one technology that no matter if you’re at the dinner table, or if you’re at the board meeting or a business meeting, everyone is talking about it. And Burcin, you hit on an earlier where the consumers, if we would call them that, the people not even in various industries, they’re hit with AI every day. Burcin, you gave a great use case with your insurance example. I mean, it’s around everyone in business or out of business. So I think that’s part of the challenge. There’s a lot of conversation. 

And guys, listen, we just scratched the surface on this today. And I appreciate it. You guys shared some amazing insights. I want to thank you both for joining. I know you’re busy. Burcin, I really appreciate you. I know you went out and took a little extra leap for us today, you had a conflict in your day, and you ended up joining us from the car, and man I appreciate it. Most would have just said hey, I can’t do it. We need to reschedule it. But I knew you wouldn’t do that to us. I knew you would find a way and I really appreciate it. 

For the audience, I really want you to pick this conversation up on our social network. So, under social media channels, let’s continue the conversation. Rich and Burcin both have agreed to do that. So if you’ve heard something you want to double-click on, please do that on social. We’ll be back. Again, at the end of April. We’re lining up a very special guest for the April True Talks webinar episode. We’re really excited about that. And we’re going to do these monthly. And again, for that is just to really go deep, be authentic. And let’s talk about topics that matter to the industries we serve, guys. Thank you both, man. Really appreciate your time and efforts today be safe, and we’ll see you soon. 

Rich Humphrey:

Thank you.

Burcin Kaplanoglu:

Thank you.

Andy Verone: 

Thanks guys.