Design-Build Delivers

AI, Data and the Design-Build Owner: A Conversation with ARKANCE

• DBIA

Share episode

Have thoughts? Text us!

Artificial intelligence is no longer an abstract idea for design-build teams. It’s already changing how projects are delivered, managed and maintained. In this episode of the Design-Build Delivers Podcast, DBIA’s Director of Virtual Design and Construction Brian Skripac is joined by three leaders from ARKANCE — Jason Kunkel, Senior Practice Manager; Nick Miller, Director of BIM Services; and Justin Friedman, Practice Manager | Facilities Management — to explore how data and AI are reshaping the Owner experience.

Together, they unpack how design-build teams are using data to:

  • Turn information into intelligence that drives better decisions.
  • Build a culture of trust and governance around data ownership.
  • Identify “quick wins” to make AI adoption practical and scalable.
  • Shift focus from building smarter to owning smarter.

It’s a conversation about the real-world impact of data — how it’s captured, shared and leveraged across every phase of a project’s lifecycle — and what it means for Owners navigating the next era of delivery.

Resources:


Access all our free design-build resources and learn more about Design-Build Done Right® at dbia.org.

DBIA members are shaping the future, one successful collaboration at a time.

 

Octfinal

Thu, Oct 30, 2025 2:52PM • 44:21

SUMMARY KEYWORDS

AI in design build, data organization, owner's perspective, AI adoption, data transformation, building systems, proactive maintenance, data governance, digital twin, AI challenges, data quality, collaboration, practical applications, future planning, AI readiness.

SPEAKERS

Nick Miller, Jason Kunkel, Brian Skripac, Speaker 1, Erin Looney, Justin Friedman

 

Erin Looney  00:03

DBIA, back in August, we looked at what AI already means for design build practitioners, how it's being used on job sites to coordinate teams, predict risks and make day to day work more efficient. That conversation with Brian skripak and Paul hedge path showed how AI is already changing the game for design builders across roles. I

 

Speaker 1  00:21

would compare this to them 20 years ago. Everyone is a little hesitant, right? But once we got a couple jobs under our belt, and then the people were starting to see that benefit, then you started to get the need to have it on projects.

 

Brian Skripac  00:33

It's a level of excitement and apprehension, the fear of the unknown. As soon as you show value in that, everybody's kind of head turns pretty quickly and they say, Well, wait, how did you get that information so quickly? Or

 

Speaker 1  00:44

when you're doing instructions of agent, you're almost giving it a job description like you would a person in HR department. So we're telling the scheduling agent, hey, you're a scheduling expert. You're going to look for these types of things. That really is acting more like an employee than it is a piece of software.

 

Erin Looney  00:59

This time, we're flipping the lens and talking about data and the owner's side of the story, especially, how to turn years of scattered information into Insight strategy and smarter building performance. I am Aaron Looney, and this is the design build delivers podcast brought to you by Archons. Three members of the archons team actually joined me for this conversation, Jason Kunkel, Nick Miller and Justin Friedman joined Brian, who himself is a recurring character on the show at this point, and they got granular about where owners can start with AI, what challenges they're facing and how design builds collaborative structure can turn data into action. In our August episode of this show, Brian and I talked to Paul hedge path about AI in design build projects primarily there in terms of its use by practitioners. This time, we're taking a bit of a shift toward the AI design build owner data relationship. So from what we learned in that episode and then our chats to get ready for this episode, it was clear there's one issue for many owners, lots of issues for owners, but this one issue of dealing with years of fragmented facility data. So why don't we kick the conversation off with that? Jason, talk about those first practical steps for owners to get that data organized and usable.

 

Jason Kunkel  02:13

We've got information everywhere. It's parsed out, it's broken out. Brian brought this up in a previous conversation, though, it's that Venn diagram I think we've all seen about people, processes and technology. And the technology is there. The processes are easy to develop. The challenge is going to be the culture. It's going to be the understanding of how to use it, what to use it. And there are so many factors involved here that we've talked about over the years, terms of training, in terms of understanding and taking risks. But then there's facets that we may not have considered in culture and with our teams before, things like security, things like, how do we want our, you know, our sustainability, footprint, things like that that are involved in terms of AI that, like I said, we haven't really thought about, we haven't talked about, so it really is going to take kind of a deep look at our internal teams, at what's important to our groups, and, you know, how our groups work, and what they can accommodate and take off. To me, I'm a nerd. The software is easy, the people are hard. So getting the culture to change, getting the ideas to change, getting willing, being willing to take reasonable risks, is always the hard part.

 

Erin Looney  03:31

Brian as our resident DBIA connective tissue, and from what I hear, still a nerd, but also good with people. What do you see?

 

Brian Skripac  03:38

Well, I think, and I appreciate Jason, dropping the people, process and technology, you know, looking at that across design, build projects in general, the question I was actually going to ask is kind of relaying that back and Jason, I don't know if you saw our recent primer, the VDC project leaders roles and responsibilities on design build projects, but you know, that's really looking at leadership From the project team to get the owner that information, and we'll talk more about how that information gets delivered later. But what are you seeing on the owner side, with the people right, who's leading the charge? You know, we've worked together on projects in the past, having that leader on the owner side. How are you seeing a lot of your owners work to really be the leader on getting that information at the end of the day and then maintaining it.

 

Jason Kunkel  04:22

I listened to the last podcast you all did kind of comparing this to BIM in some ways, and it's a little bit the same way. We certainly found some owners who understood the benefits of BIM right off the bat. They got it. And then there are some out there said, Give me BIM. I want all BIM, which, of course, is not a thing, and we're seeing that here too. They say, well, I need data for my AI, give me all the data. And that's just draconian, impossible. So the owners that we're seeing we're asking this and understanding this already, kind of have a good understanding of what information either they have or what information. Action slash data they need. It's not just some sort of random concept out there. They get it. It's like, okay, I need to know X, Y and Z. I don't need to know a, b and c, and so it's a very easy transition for them to go into. All right, how can I leverage this? What can I do with AI after here, those are the ones we're having more conversations

 

Erin Looney  05:19

with that relationship. Brian, you already said, thank you for bringing up the people process technology. I actually have dreams where those words are just in neon lights.

 

Brian Skripac  05:28

Dream. Nightmare. Are

 

Erin Looney  05:30

you sure? Call it what you will. And usually I'm trying to run away from those neon lights. But those relationships are important. But But obviously, in addition to the relationships, there's also the theory and the buzz and putting that into practice. So owners hear AI. They hear data transformation everywhere. We see article after article about it. We we see it just sort of show up in all of our conversations, almost. But what does starting that actually look like? Justin tell us how an owner can assess their data landscape before they even think about bringing in AI tools

 

Justin Friedman  06:05

today, when we talk to a lot of owners, right, when we ask them, kind of where they're at in the whole data situation, and what do they kind of want to do with the data, they're not sure. So I think one of the recommendations that we kind of always put out there is think about the end goal, like, what do we want to do with the data? What's going to be meaningful for us. And if we can kind of define what that looks like, then we can work backwards. We can then start to look at, okay, what data is actually available to us at this very moment, what systems are we using? What data is it tracking? Is the data reliable? And then I think once we kind of understand that, when we look at the gaps, what data is missing with what we currently have to get to the end goal. And if we know what our gaps are, then we figure out how we actually fill those gaps. It's really about not looking at where we are today, but where do we want to go. And if we can really define that, the other pieces kind of fall into place. And then if you really start to break that down, and you meet with all the individual stakeholders that have skin in this game, about taking the data, transforming it into something meaningful. We understand how they want to use it in their day to day. How is this data going to help them, not only from the very high levels in the C suite, but also the boots on the ground? What's going to help them do their job more efficiently, as well as the building run more efficiently and be safer for the occupant? What do we want it to do for us? And if we have that definition and it's clean, reliable data, then we can work on the AI side of that and be okay. We want AI to help us take our data and do this automated in a way, and send out meaningful metrics that we can use moving forward and reduce that overall decision making process again, as well as reducing the overall operational downtime.

 

Jason Kunkel  07:39

I want to double down on something. Justin said there about the stakeholders. It's got to be all the stakeholders as well. They can't do this in pockets. Usually we find that this charge is led by one team. This is a great opportunity to talk to all the teams, which requires leadership, and it requires some tough conversations sometimes, but this is a great chance to bring all the teams together and kind of have these possibly new

 

Justin Friedman  08:04

conversations, and Jason just to kind of continue down that road. And I think this kind of goes back to the first question, right? Is when we meet with those different stakeholders, whether they be design, build or operate, stakeholders, everybody has different data points that they're looking for. And a lot of times when we talk to owners, it's like, hey, what data would you like? And you're like, we want it all. That's one of the biggest things we try to stay away from, because the more data you have, the more data you actually have to maintain. And who's maintaining that data, who's keeping it clean? And as we all know, you can have the greatest software in the world, but if the data is not accurate or reliable, then the metrics aren't going to be reliable. And it basically becomes a meaningless exercise. To Jason's point, it's not only understanding the data that all the different stakeholders want, but the data that's actually needed. So we're

 

Erin Looney  08:48

actually going to dig into that a little bit more about the who, what and where. Justin So, nice little foreshadowing, Brian, before we go on to the next question, your perspective is a little bit different because you talk to different owners from all over the place in your travels. So what are you seeing owners taking advantage of in terms of data and AI?

 

Brian Skripac  09:07

I think it's just what Justin said, Right? Owners, they want everything. Hey, it's all there. Just give it to us. We'll figure out how we're going to use it. I think the big challenge there, and I know in my past experiences working with owners, is if they want everything, but they really only need 10% of that. That other 90% is waste. It's waste in the process of structuring it, turning it over, thinking about how they're going to using it, storing it, doing all these other things, just from a data perspective, let alone how is that informing some AI application to benefit them down the road? So if that's immediately becoming outdated, it's wasteful. It's just not a strong piece of the puzzle. It's not going to drive value in that. Going back to Jason comment or previous conversations like them, I want everything. Well, what do you need to do to maintain and I think our cons obviously understands that, and you know, when they're working with clients, being able to drive. Describe that, but there's some consistent data messaging that you're seeing right? Every owner is asking for XYZ kind of perspective, and where are they using it, maybe a little bit more specifically in AI applications today. Where are the big benefits that they're seeing

 

Justin Friedman  10:14

right now? Where AI is, in terms of the operational aspect of a building is really around building systems, right, turning off, you know, utilities where floors might be less occupied, I think those are the types of areas where we see AI today in terms of actual building maintenance. You know, we still know we need the boots on the ground for that. Somebody's got to repair equipment, right? But maybe it's being more proactive, and where we're seeing certain activities in certain areas of the building, that we can avoid those kind of major pitfalls of major capital expenditures on assets that break. So we kind of get that heads up or insight, if you will, by AI turning off certain things within the building based off the activity at the building. So I think that's where we're seeing a lot of the savings at this point. But that's, again, I think there's a lot further we can go with that

 

Nick Miller  11:01

to expand upon that, I think we're seeing a lot of drive in the kind of quote, unquote, invisible AI right? Things that live in the ecosystem they're already working in. They work in the background to parse data, make decisions, help solve problems. We're not seeing a huge push towards driving massive AI tools into place to solve huge business challenges. It's really more specifically like, can you work in the background and help resolve the little things that come up on a day to day basis that are minor attributes to the the overall contribution of what's happening inside of the building life cycle, but completely solving, you know, everyday issues that people run into and can take advantage of these tools and technology.

 

Justin Friedman  11:40

One thing that would be great to see, right? Is if you had some kind of AI engine running that was like, hey, I want to reduce this building's utilities cost by 10% over the next five years. If you could somehow get AI to kind of tell you the direction you could head in to make that happen, right? That would be great. But I just don't think we're there yet.

 

Erin Looney  11:58

So, Nick, we're going to stay with you for a little bit because you were quietly waiting for your moment while we went through the setup there. From the number of times we've said, AI and data in this episode alone, it is obvious that's a non negotiable part of the conversation, this relationship. But owners, you know, as you've already kind of hinted to they worry about who controls the data, how proprietary information, sensitive material or managed. How do you ensure sensitive project or facilities data remains protected? And talk a little bit about ownership,

 

Nick Miller  12:30

as we've talked about here, owners are now receiving more data than they ever have throughout the project life cycle, and they have this legacy history of information that they're working through as well. Because of this, we're really encouraging that owners become better data landlords. They need to set strict rules for their project participants, their internal team members, the software vendors that they work with, and they really need to shift from, you know, kind of passive data handovers to more active data governance, where you have policies that are driven through technology and contracts to actually help implement, manage and take care of all of this data that you're receiving, more holistically for your organization, when it comes to the protection and kind of keeping secure data separate or segregated, right? It's something that we help consult with clients all the time. We recommend kind of an anonymize or tokenize approach, where we're taking sensitive data that's got PII or equipment information or serial numbers, we're making that more generic in the purpose of our broader search capabilities or our larger data sets, we can segregate that data into a separate layer and then tokenize it on its way out, so that we've got a better understanding of the insights that might live inside that data without the secure information that we're trying to keep from the broader public understanding. The other option is to bring the AI into the controlled environment and have it gain its insights and then leave without ever bringing the secure data outside of the cloud, hosted environment where that secure data lives. There's kind of two different approaches in terms of keeping data secure when it comes to owning the IP and the output that happens throughout the project life cycle, which is where it gets a little bit murkier, right? What we encourage our clients to do is own the output. The output should be your IP. If I'm using an AI model to train for predictive maintenance over the lifetime of a facility, the output of that training exercise should be my IP as the owner, and I should take that with me from the project exercise that should be something that I mandate and standardize throughout that project life cycle, and ensure that the learnings from the AI tools are something that I take with me as an owner.

 

Erin Looney  14:43

Okay, so let's move down the road a little bit and look at projects after they're delivered. Once a facility is operational, how are owners defining what data they actually need post occupancy? Let's kick it off with you. Brian, I

 

Brian Skripac  14:58

think that's an important part of. Of really developing a project delivery standard where the owner, you know this is where you break down, is as Jason and Justin talked about, right what's actually needed, defining very clearly what what data is needed, how it's structured, how it's formatted, and how it can be consumed. What owners want varies quite differently. One type of owner versus another is going to have different needs. Sure, there's going to be a lot of overlap, but how are they using industry standards? How are they using common data structures? How are they receiving it? How are they consuming it into their applications, which vary quite widely, is key. So I think, much like our cons does with clients you know, being able to sit down and have that discussion and say, What are you looking for? Let's document it. Let's put it in the front end of your deliverables, in your contract, in a delivery standard that this information is required. It's probably less important about who's delivering it. Let the project team figure that out, right? Because it could come from multiple people, somebody like a VDC project leader is probably on the Design Build Team, is probably the key aggregator of that and really the gatekeeper, quality control person that's going to receive that information and turn it over at the end of the day. But it's a process that starts before you know an RFP. RFQ actually goes out. So once you have that information, kind of sifting through what you have and understanding what's missing, what's needed, that allows you to plan moving forward,

 

Jason Kunkel  16:28

you know, focusing specifically on the owner side of things. And this is more of a question, and I don't know if there's a specific answer from our end the work we do, we feel like we've got it nailed down in terms of design and construction, collecting that data, how to organize that data, how to share that data. From the owner's point of view, I've always been curious if there have been changes to specs and requirements in terms of kind of that ongoing data collection they're going to be in these these buildings for 30 years. Are they defining specific rules and expectations around their you know, mechanical equipment, how they expect that to get collected and organized? We haven't said digital twin yet, so we'll see digital twin now. I feel like that buzzwords got overlooked by AI at this point. But you know, how are they kind of organizing and defining and requiring that data structure and from us, the work we do, the people we work with? It's kind of that handover. I don't know, Brian, if you've seen anything in terms of requirements and expectations around the ongoing data collection, I

 

Brian Skripac  17:28

think from our perspective, the clients that I've worked with and still have relationship with, is is maintaining and evolving it, right? I think the worst thing you can do is say we have a standard here it is, and walk away from it, having that partnership to make sure it evolves and continues to grow is critical. I think the next step of it is maybe less about what it is, but how you receive the information which kind of rolls into the whole IoT conversation and that back and forth that you start to see of digital twin. Not to start throwing more terminology out there, but how does the facility respond to what you're doing? And this becoming more of a conversation, kind of similar to AI right, where we're capturing this data about how we're using a space. Is it occupied? What is it? What's the thermal comfort of it? All of these different things. Now you're feeding that back into the system, and now how do you respond to it to make sure that you're maintaining things at a consistent level to achieve the performance requirements that have been set. So it's less about what additional information do you want, but how do you consume the information that's available that you're building or your asset is giving back to you, whether that's at a at a macro level, at a building or a campus, or a very micro level, a specific piece of equipment.

 

Erin Looney  18:45

It's interesting to compare us talking about occupying a building for 30 years, 30 plus years, in terms of this data, to compare it to actually last month's episode of the podcast, where we talked about the physical occupancy of healthcare and education facilities and that planning for the future, and I'm assuming, without having the requisite knowledge, these things do go hand in hand. How you manage your data and how you deal with what you all are talking about certainly impacts how people move and exist within the spaces you're building. Because we're talking here about technology and these almost nebulous sort of things that people can't necessarily see when they walk into a completed building to use the space. I'm assuming there is a relationship between what you're talking about in the end user. Am I creating connections that don't exist, or is that fair?

 

Jason Kunkel  19:37

That thing is absolutely fair. Nebulous is a great term for what we're talking about. Now, I certainly think to the end user, it will be invisible. I feel like we need to hit a critical mass of data also, before we can really start leveraging this on a large scale. Some of this information been collected. It might be just kind of floating out there in a data lake somewhere that can get cleaned up. I don't know that a lot of. Owners have been collecting what we are going to find really useful five years from now. I think the chat we're having now is like, Okay, get ready, clean up, but I'm really excited for the chat we're going to be having in 2030 which is horrifying to say, but

 

Erin Looney  20:15

no, it's not, because it gives me an episode for 2030

 

Justin Friedman  20:19

You're welcome. I would also say like, I think the bar is constantly moving. What we know today and how we use data today, like you said, in five years, that could be completely different. So I think where AI could also play a part in that is making the pivot easier. Potentially, we already kind of have those models established. Maybe AI is going to give us the ability to make those pivots again a lot quicker than we normally would, whether it be in one system, manually creating fields or, you know, and then gather all the data input and format that data into the correct field, so that then fills into other supporting applications, like a common data environment and aggregating all the data together. So I think there's also that aspect of to where, yeah, you're 100% correct, like right now, the actual users of the space or the occupants of the space. It's behind the scenes for them, they don't recognize it. As long as they're comfortable, they're going to go on about their day. But it's it's for us, as owners as well as operators of the building, that's where we really see the benefit is in our day to day of how we're maintaining the facility, the things we actually have to fix. You know, if we can again, change that and make it more again, proactive versus reactive. We see it on both sides, where operators get to spend their time doing something else within the building that needs to be adjusted, or as owners, we're also seeing the impact financially, where we're not spending as much time and money in certain areas of the building where we would if we did have this kind of technology in place.

 

Erin Looney  21:38

The Design Build delivers podcast is brought to you by Archons, your technology optimization partner, helping design build teams streamline workflows with Autodesk solutions, expert support and real world training ready to work smarter and faster, get Tools insights and schedule your AI readiness check at arkons.us/dbia cons.us/dbia

 

Erin Looney  22:06

we were just talking about going forward five years, but I'm going to go back to 2023, at least. Since I started here, we've talked a lot, a lot about AI from that August episode that we were just mentioning earlier to last May's conversation about augmented reality. Brian was there too several recent posts on the design build delivers blog, all of which I link in the show notes for this episode. This is clearly a topic that's not only not going away, it's ramping up. So just in that means we need to think about it much more strategically in our work, which you have said, Tell us, how does the use of AI, particularly in light of what we're saying here about data and the owner's role. How does that change the ownership experience once the ribbons cut?

 

Justin Friedman  22:48

In the near term? I don't really see it changing that much in terms of the actual maintenance aspect of repairing equipment, things along the way that, again, any building engineer, space plan or facility manager would do. I think where we see it right now is the actual building systems, as we talked about earlier, right? The ability for the building systems to recognize a certain floor is not occupied. So let's turn off the air. Let's turn off the lights, you know. Let's save money where we can save money just by noticing that a certain part of the building is not again occupied. But it also comes down to, I think, under performance of certain areas of the building, and the BMS or the BAS can actually notice that whether it triggers an automatic alert or creates a work order, whatever it may be, or it orders a part for you as a part goes down, maybe you have an AIP set allows you to order inventory based off things that are broken, but at the end of the day, somebody again still has to be in the Building to replace that equipment or turn off that alarm. So I think near term, yes, I think there's autonomy around being able to control the building and its systems. There's not a way for us to streamline, I guess, would be the best way to say it fixing the building. But if we can find a way for AI, in the long term, to extend that life cycle of being in a building from 30 years to 40 years, right? I mean, that's where I think we're trying to get to. Let's get the most out of the building, and let's not get out of the building simply because it's broken down. We can't fix it at this point. It's going to be too cost efficient to actually do that. We need to find a new space to go. So let's try to extend that building life cycle with the use of AI being proactive, and then finding a way to actually, then reduce our overall operational costs.

 

Erin Looney  24:24

You know, my Alexa already does order, try to order light bulbs for me without really asking. So I'm assuming it can be done, if they can do it in my tiny apartment. Eventually we're going to get

 

Brian Skripac  24:35

there. To add to what Justin just said, too, it's, you know, you have that proactive maintenance, right? You're trying to extend the performance of the facility in a positive way. But I think kind of, there's almost another step for that, for owners, okay, this is how we utilize this facility when we go to build the next one. How do we take those lessons learned for how we use space, how we manage this, how we ran through. Things, certainly for higher ed, healthcare, these other owners with these large real estate portfolios, who are, you know, somewhat of a serial builder, right? They're expanding their portfolio. They're adding they're renovating. That also informs future projects, not just the maintenance of the one. So while they start there, it's also about, how do they reuse and repurpose that information on the next design build project that they're doing. Yeah,

 

Justin Friedman  25:23

Brian, that's actually a great point. I mean, when you call out, like higher ed and healthcare facilities, and we talked about the bar constantly moving for those kinds of industries, I mean, they always have to change the metrics of what they're reporting on for their individual campuses to, you know, state and federal agencies to get funding, and all those fun things that kind of happen outside of what we think about in our day to day. It's critical for those kind of lessons learned. I'm glad you brought that up. We

 

Erin Looney  25:47

need to, of course, think about the industry and the people on the ground alongside these inevitable technological evolutions. And Justin mentioned the people in the building who have to actually be there to install and repair the physical building, you know, and stopping Alexa from ordering 4011 batteries just because she sees one is ready to be changed. So Nick, let's talk to you. Hear from you, first on the archons perspective on this, and then Brian, I want to hear from you about where DBIA enters the chat. How should design build teams and owners work together to set those expectations around the data?

 

Nick Miller  26:18

This is more than ever a conversation that has to be started very early in the process, and it has to be very deliberate. We're finding that the data exchange is as important as the building handoff. The reliance on good data, as we've talked about over and over again, is higher than it's ever been. So the owner, in our opinion, needs to really generate a data requirement specification, something that defines the handoff that they're expecting. It should go into the data standards, what asset data is required, ensuring that we can drive these facilities management processes downstream and all these other space planning tools and other pieces to our puzzle that go into the life cycle. Management of the building has to be outlined right up front. You also need to get into the level of information needed, right we Jason talked about BIM earlier, and level of detail comes up all the time for broader data needs. Level of information is also very important. What assets are connected to this? What information do we need to connect at the asset level? What data needs to be collected when I install this machine to ensure that my facilities management team has what they need. All of those things need to be defined upfront, so that the team understands what the expectation is, and they can ensure the subcontractors in the field are gathering that information at the right time, so that they're not retroactively trying to go back and figure out what happened in that space. Additionally, we're seeing more owners start to tie data quality to payment milestones. So not only is it enough to have the building built, but actually understanding that the data you receive drives the processes that you need post handoff are super important. So aligning those payment milestones to data quality is something that we're seeing more owners start to do to really ensure that that building is operational, but then also functional within the broader ecosystem that they're trying to stand up as well. We've

 

Erin Looney  28:06

talked a lot so far about what can work, what does work, what about what doesn't what are some of the common barriers you see to meaningful AI adoption among owners. One

 

Justin Friedman  28:17

of the biggest things right out of the gate is the knowledge gap. AI, isn't, you know, it's not brand new by any means, but I think we don't really fully understand AI, how it can be manipulated to do certain things within our industry, and until we kind of bridge that gap, I think all of us are kind of looking at it from a what ifs perspective, and not so much of, oh, let me go do this, because everybody's really unsure specifically About what AI can do for them individually. I also think, like we've talked about siloed and unreliable data, right? Everybody has their own data that they're managing, and whether or not it actually is correct or if it actually is going to be useful for another stakeholder, we certainly see that barrier all the time. And I think, to Jason's point, the handover process, what we want gathered during design and build always isn't what we get at handover massive portfolios. You know, owners can have all different types of buildings within their portfolio. So how is AI going to actually manage an organization's different building types? Right? I mean, we work with tons of organizations that have a lab, they have office space. I mean, there's all different types of building types, and so having AI being able to essentially operate and autonomize your different building types is also, I think, a barrier. That's one that's not necessarily self inflicted by any means. Data Standards, I think that's one of the biggest things. We also don't see a lot of organizations that have standardized data, and they don't have the standards in place to wear at handover they get the data they need so that AI can then potentially transform it into metrics that they're looking for out of the gate. Those are probably some of the big ones

 

Jason Kunkel  29:44

want to expand on what just talk about the knowledge gap. There is active white noise around this topic right now there, there is people are getting inundated with everything, and everything's AI, and this is that, and it's just confusing. I think it's just the media and the. World coming in and forcing it in on people now is also a challenge.

 

Brian Skripac  30:04

You know, it's also what you see and understand, is AI and media, right? People with images and fake images and videos that drives fear and anxiety. Certainly, there is some of that. But I also think the AI conversation just gets lumped in and generalized as that. You know, the opportunity for us to have this conversation and talk to industry leaders from our cons is the opportunity to say, Hey, these are the practical applications in the building industry, construction industry, this is how we're using it. This is where the benefits are. And it's not just chat GPT and helping people get beyond what Nick was talking about, the security side and the risk management of, oh, if somebody gets my data outside, how are we going to manage that? I need to wall this off. And there's certainly an opportunity for an owner to have a confined space for their data to live that's secure and for them to interrogate it and investigate it. And that's an opportunity. It's not a free for all out there, right? Your information doesn't have to go everywhere. It can be secure. And there's partners that we have out there that can help, like Archons, that can help bridge the gap on that, which I think is really important. I think you're right, Jason and Justin just listed all of those great opportunities, and I think that's the part that I feel we lump it into what we see on the news or on social media. And there's so many practical applications beyond that that owners can take advantage of and see real value in I think

 

Justin Friedman  31:21

we're also missing use cases until we really know what, like building engineers, facility managers, those kinds of folks that you know would probably benefit the most out of this kind of AI transformation around data for owners, it's kind of hard for us to really build something right and understand how they want to use it. I think right now, there's a lot of theory behind it. It's just we don't, again, don't have the concrete use cases of like, hey, it'd be great if AI can do this for us. And I think we need to start asking those kinds of

 

Erin Looney  31:47

questions. And I know it's different across different industries, but Brian and I have talked about this at length. There is the ongoing fear of, go back to old South Park. They're going to take our jobs. And the reality is, it was Brian who said it. I you'll have to tell me where you found it first. But we've really relied on it. It's not the actual technology that's going to take the jobs and take these roles away. It's actually the person who knows how to use it effectively. And it's not, it's not going to be a substitute for everything we've been doing. It's going to be a tool. It is a tool I'm seeing that I'm learning it's it looks a little different. It's wrapped up in different paper, but that seems to be the standard across regardless of what you do for a living.

 

Brian Skripac  32:28

And I didn't say that, I'll be the first to jump in and say I heard it somewhere else. I don't know who said it. They probably heard it from somebody else, because now it seems like a really profound statement,

 

Erin Looney  32:36

but they've heard it from chat GPT. That's who said, Yeah, somewhere

 

Brian Skripac  32:39

yeah. But I think it's important, because what we miss with all of the, as Jason said, white noise and anxiety about it, is the opportunity for efficiency. Just over the last two months, going to Autodesk, university, pro, core, groundbreak, everything was about AI, right? But showing practical things of, how do I find information about a project? How do I get this? How do I connect this? What can it automate about my processes and do some of the mundane tasks in the background? That's the big value that we have to our advantage right now. You know, we don't want to be going through, searching through documents, looking for text or, you know, spec books are two inches thick. How are you going to sift through that and find what you're looking well, if it's a PDF, you type in an AI finds it. Here's every single reference all the way down, and here's your risk opportunity with this, if you're looking through a contract for certain things, there's all of these applications to just manage the massive amounts of information and get you the answers you need that you can take advantage of, even before we get to these amazing preventative maintenance opportunities that we're talking about for owners, which is there, sifting through the information and giving you feedback and ideas and next steps without having to do it on your own, saving you time. That reminds

 

Erin Looney  33:50

me of an example that actually just happened the other day. I was in a meeting that I recorded, and it was with our CEO, Lisa and our Vice President of Communications, Danielle and I looked through the transcript, looking for what we had decided about a very specific thing we talked about. And every time I read it, I went, I have absolutely no idea where we landed. Am I just, is it me? So I threw that into chat GPT, and I said, Here's my question. Help. And Charlie, which is what we call her, has returned, yeah, I don't know, it was like, Oh, good. It just reaffirmed that I wasn't missing something. But the time it took for both AI and I to find Charlie and I to find that we had no answer. I thought, Man, I'll never get that time back. So we did figure it out. Eventually. We talked here quite a bit about noise, and we know now what it looks like. So Nick, what does it look like to cut through this cacophony, especially for an owner who just wants practical results,

 

Nick Miller  34:47

I strongly encourage people to ignore the messaging that AI is going to solve all their problems today, we are very much finding that narrow focused AI is the easiest and fastest thing to. Implement it solves everyday challenges. I very much like to use the analogy that we are in the AOL days of AI right now. It's fun, it's cool. I'm sure it's probably more than chat rooms and downloading illegal music, right there's probably to Jason's point earlier. Five years from now, we're going to be talking about way cooler stuff in the industry that we do with this technology. We're very much in that kind of early phase discovery of

 

Erin Looney  35:22

it. Do you think that kind of hype actually slows teams down, like they're waiting for some perfect all in one solution, instead of just getting started,

 

Nick Miller  35:32

what we try to encourage clients to do is take a problem first not tool first approach. We find lots of clients who go, oh, this tool is really cool, and then try to back into something that is a problem that they can find that that solves, instead of the opposite, right? Find things that you don't like. What are the three most expensive problems you have? Let's go look for tools that specifically address those issues. It makes more sense to work that way. If you need to build something custom, it's more viable because you've defined what you're trying to accomplish. It goes back to the all the information and the data we're gathering. What of this do we actually need? What don't we need? Let's define that right up front, like I mentioned earlier in the the episode here, prioritize the invisible AI. Find tools that you're using today that are adding those features and figure out how they can impact the workflows you have that's going to be significantly faster and more impactful than trying to collect all of your data into a data lake and then figure out what you're trying to do with it and what AI agents you should be applying or subscribing to to make that useful for you. And then, to Brian's point earlier, we find like starting with a quick win document centric AI is the simplest and easiest things our clients can do. You've got this huge repository of data available to you. It's your historical tribal knowledge of everything you've ever done. Apply an AI to that and allow it to make it easier for you to find those surface those insights. Have we ever done this before? When did we tackle this? Where did we use that roofing membrane? All of those things are simple. They're easy to implement, and they can gain immediate, tremendous value today from the data that you already have available

 

Erin Looney  37:09

to you. You just touched on something there. That's another one of those things. You hear a million different places, and it's just because you can doesn't mean you should. How many times I've worked with somebody who's like, there's this great new program, and I go and how there is no and how it doesn't really add anything. So going forward, then thinking about planning, how can owners set realistic expectations, you know, and avoid these common pitfalls of things like, look at this shiny toy, let's figure out how to use it. And venturing into the Sci Fi realm, how can owners set realistic expectations when you integrate AI into planning?

 

Jason Kunkel  37:44

Be clear, I love sci fi ambitions. I love the shiny things.

 

Erin Looney  37:48

So do I.

 

Jason Kunkel  37:51

I think I would encourage folks to do about the least sci fi thing of all, go talk to your peers. You know, there's conferences out there, I know, and FMT is one. I'm sure there's plenty of others out there. Build that kind of network and understand what your peers are up to and what your peers are doing and what's worked for them and what hasn't worked for them. Often, there's a tribal knowledge in groups. I've found that on the owner side, on the facility side, you know, there doesn't seem to be that fear of holding things close to their vest, they're certainly far more willing to share as long as they have the opportunity to share. And I think just kind of getting in front of people and chatting about it is a great way to kind of start learning what's working, what's not and what to avoid.

 

Brian Skripac  38:36

I agree, even your design build teams, your builders, your architects, your engineers, you probably have those trusted advisors there as well that you go back and work with. They're going to have experiences on what they're seeing, what they're delivering to other owners, and what opportunities are. I think that's a good thing to have that back and forth. Maybe it was Justin that said, before you know, how do you balance that? We want all this information right? Having that partnership to be able to challenge one another and have a dialog. Sure, we can give you all the stuff that we have, but what do you really need? What are you experiencing? Okay, I can answer that. I can give you this information. We have this readily available, or we have this and it needs structured this way for you to use it. Those are the teams that are going to be delivering that information. So in addition to talking to your peers, which is, which is totally critical, have those same conversations with the people who are bringing that, that current, unstructured, maybe analog or barely digital information to you, and talk about, how do you evolve that as a process that's beneficial moving forward? And have a why for an owner like Nick was saying, right? What problem are you trying to solve? Hey, I have this. Maybe it's not. How are you using AI, I'm having this trouble of getting this information that I need. How are you delivering this to other clients? Is there a better way that we could be doing this than we're doing now? Have those partnerships, have those conversations.

 

Erin Looney  39:54

If the robots ever do take over, it's going to be because someone did something with some new, shiny technology. Technology without thinking it through, and the robot saw an opportunity. So let's plan, let's strategize, and let's wrap this up with some opportunities. Let's talk about now the biggest opportunities for owners to take advantage of data and AI in the future. And then Brian, I'd like you to tie it back to how design builds collaborative process, sets them up to get there faster.

 

Nick Miller  40:22

The immediate impact is taking advantage of it by focusing on micro improvements and define some very clear metric. Again, this notion that we can use AI to solve all of our problems is exciting, but it's very sci fi in nature. Focus on very specific things. Can we identify the top cause of delays in our schedule based on open RFIs, very clear metrics with a defined purpose, focus there and then work backwards to understand what kind of a tool you need to implement or build to put that into place. There's a million areas of potential cost savings within a project lifecycle, I think, starting small and kind of understanding the problem you're trying to solve and which tool applies best is absolutely the easiest way to get going.

 

Jason Kunkel  41:06

100% agree with Nick on the Start small. There are absolute wins and things people can do now, if I'm putting on my sci fi hat a little bit, I do like looking down the road once we get the cycle, once we get the data cleaned up, once we have the information there, looping and plugging it back in to the decisions at the beginning of design. Again, I love coming full circle with this. I love the idea of this and just asking questions that we weren't even expecting. Why did I only get three years out of this? We're supposed to have five. Oh, it turns out, because it was installed on a Tuesday and it was raining on that Tuesday, there's going to be these relationships that we weren't even expecting or thinking about that we're going to be able to improve how we do things down the road, and that's what I'm excited

 

Brian Skripac  41:45

about. Tying this back to design, build that contractual relationship, having that streamlined focus to deliver information. It's very similar in the way traditional project information goes through the project team to the owner. It's no different there. But also these ideas of building collaboration and trust, I think that starts to mitigate some of the challenges, or maybe fears about, how do we get the data? What data is coming in? Is it reliable? Has it been quality controlled? And having the owner working with somebody on that design build entity like the VDC project leader, can start to be the one to streamline the organization, the being the recipient of that information, being the one to quality control it, and delivering it to the owner and really serving it that as an interface in this evolving conversation of this digital information that gets translated from the project team to the owner, is where we're going to continue to continue to see success, and having that one to one relationship from those two sides of the contract, I think, is a big value for design build teams

 

Erin Looney  42:48

across the many conversations DBIA has been having about AI lately, from this episode to the August episode in design build delivers blog posts from DBIA and from A number of guest contributors. It should be pretty clear by now, AI is not going to wait for us to get comfortable. It's already shaping how design build teams, plan, deliver and maintain the built environment. What we're seeing is the same through line that defines design, build itself, collaboration, trust and shared problem solving. Now AI doesn't replace that foundation. It builds on it the success stories we're hearing show technology works best when people in process come first. If you'd like to explore more, you'll find links in the show notes to dbias new free VDC project leaders, role and responsibilities on design build. Project primer that Brian mentioned, design build delivers blog posts on AI, including companion pieces written by guest contributors from across the technology and AEC landscape and a couple other design build delivers podcast episodes on the topic, each offering slightly different but still practical insights on how AI is impacting data technology and leadership in design build. Many thanks to Brian skripak, again, Jason Kunkel, Nick Miller and Justin Friedman for joining me and to Archons for powering this and every episode of the design build delivers podcast. Learn more at arkons.us/dbia

Head Shot

Erin Looney

Host
Head Shot

Fred Yi

Editor