Welcome to Nexus, a newsletter and podcast for smart people applying smart building technology—hosted by James Dice. If you’re new to Nexus, you might want to start here.
The Nexus podcast (Apple | Spotify | YouTube | Other apps) is our chance to explore and learn with the brightest in our industry—together. The project is directly funded by listeners like you who have joined the Nexus Pro membership community.
You can join Nexus Pro to get a weekly-ish deep dive, access to the Nexus Vendor Landscape, and invites to exclusive events with a community of smart buildings nerds.
Sign-up (or refer a friend!) to the Nexus Newsletter.
Learn more about the Smart Building Strategist Course and the Nexus Courses Platform.
Check out the Nexus Labs Marketplace.
Learn more about Nexus Partnership Opportunities.
Episode 183 is a conversation with James Dice and Brad Bonavida from Nexus Labs, as well as Devan Tracy from Lockheed Martin.
Episode 183 is a conversation with James Dice and Brad Bonavida from Nexus Labs, as well as Devan Tracy from Lockheed Martin. In this episode of the Nexus Podcast, the Nexus Labs team breaks down the top stories relevant to energy managers, facility managers, IT/OT managers, and workplace managers.
Introduction (0:50)
At the Nexus (3:23)
Digitizing Operations & Maintenance (07:30)
Integrating, Connecting, and Securing Devices (16:40)
Smart Building Champions (25:28)
Sign off (32:55)
Music credits: There Is A Reality by Common Tiger—licensed under an Music Vine Limited Pro Standard License ID: S706971-16073.
Note: transcript was created using an imperfect machine learning tool and lightly edited by a human (so you can get the gist). Please forgive errors!
James Dice: [00:00:00] Hey friends, if you like the Nexus Podcast, the best way to continue the learning is to join our community. There are three ways to do that. First, you can join the Nexus Pro membership. It's our global community of smart billing professionals. We have monthly events, paywall, deep dive content, and a private chat room, and it's just $35 a month.
Second, you can upgrade from the pro membership to our courses offering. It's headlined by our flagship course, the Smart Building Strategist, and we're building a catalog of courses taught by world leading experts on each topic under the smart buildings umbrella. Third, and finally, our marketplace is how we connect leading vendors with buyers looking for their solutions.
The links are below in the show notes. And now let's go onto the podcast.
Welcome back to the Nexus podcast. It's James here. I have Brad. We're missing Rosie this week. Hello. Hello. Rosie's on vacation. Uh, and then we have a special guest who I'll introduce [00:01:00] in just a second. Um. This podcast is designed to give you a preview of everything going on, um, in our content at our events, uh, and obviously what we're doing right now at at Nexus.
So, um, if you want the full experience though, this is just a preview, go over to our website and sign up for the newsletter and that will give you everything we publish each week along with everything going on with our events. Um, our special guest today is, uh, Devin Tracy, smart Buildings lead at Lockheed Martin.
Devin, do you wanna introduce yourself?
Devan Tracy: Sure. Hi everybody. Thanks for having me, James. And Brad, great to be here. I have been with Lockheed for. 12 years. My background is mechanical engineering, got a PE in hvac and I started the Smart Buildings program I think back in 2018. We modeled hours after Google. I had the opportunity to meet Darryl Smith back in the day and [00:02:00] we've been going ever since, starting with FDD and expanding out to other industrial internet of things technologies.
So I love Nexus and. Huge supporter of this community,
James Dice: and you guys have thanks to yourself and all your hard work. Um, and your team, obviously, um, you guys have one of the most mature smart buildings programs out there, so we've been thinking a lot about where all of the building owners in our community, where they sit in terms of maturity of their programs.
And what we learned next last year at, at Nexus Con was that a lot of building owners showed up. We had over 62 organizations. A lot of building owners showed up and were kind of like deer in headlights a little bit, like hadn't really got their programs off the ground and I think yours, they can look at and say like, well, where can we be in eight years?
And it's probably fun for you to look back on the journey.
Devan Tracy: It is. Yeah. It started out with just me and we proved [00:03:00]the value pretty quickly and. Then the need for a team to actually sustain the program was, it was realized pretty quickly that this, you can't just make wave a magic wand and, and hope that the technology will fix everything.
And so that's been crucial to our success.
James Dice: All right. Well, we're excited to get into some AI stuff with you in just a second. Let's talk about what's been going on at Nexus. Um, we've been on vacation a little bit, so all of us took a couple weeks off. Uh, we're kind of resting up for a big push ahead of Nexus Con this year.
Um, Brad, you want to, you want to, uh, talk about what's been going on?
Brad Bonavida: Yeah, I mean, apologies for the delay in the podcast publishing here. I know like when I listen to podcasts a lot and then someone doesn't produce on their regular schedule, you're like, what's going on? So yeah, we were enjoying our July, but we're back in the saddle now getting ready for Nexus Con.
Um, aside from that, we do have some. Pretty exciting new partnerships. We've been working with some [00:04:00] companies that, uh, are really leaders in the space that we're excited about. Just to rattle off a couple, um, we're working with McDonald Miller, which is a big mechanical, electrical, plumbing and HVAC controls provider in the, uh, Pacific Northwest.
So they're kind of our first, like we're calling it technology enabled services partner. They're really like leading. They're a leading provider of adding these technologies like FDD into the services that they're offering to, uh, their clients, which is really cool. Um, we're partnering with ar, which is a massive, uh, consultant, but they've got a really cool smart buildings division that's doing amazing stuff, helping big.
Uh, commercial real estate owners get to their outcomes by, you know, creating a data layer and getting all their data in one place and putting smart applications above that. And then lastly, uh, J two Innovations as well, which is a Siemens company and they are doing HVAC controls in a way that kind of no one else is in.
I mean, maybe James, you can add something [00:05:00] there. I know you just talked to Scott, but the way that they are, um, enabling the entire technology stack at the HVAC control layer is pretty unique.
James Dice: Yep. Yep. Um, we will talk about them at some point in more detail. I think it's really cool how they templatize.
Control system. So
Brad Bonavida: yeah,
James Dice: the old world of controls, and it was funny talking to Scott a couple different times, but, um, he's been doing this for 30 years, but it, it is funny to talk about how this is like cutting edge, but, um, a typical, the old controls world is that you might program a VAV to talk to an air handler, and a, a programmer might do that.
Like they're doing it from scratch, right? And it might be this like custom expensive solution and they're, they're developing the ability to just say, well, you know what? That's a VAV with hot water reheat. It's talking to this air handler and the system just basically, uh. Does programming, graphics, alarms, f, d, D points, all that based [00:06:00] off of predetermined templates because as we know, the systems are the same across buildings.
There's just a little bit of variation. And so using the ability for templates and wizards to, and we did a podcast with them like going through that a couple years ago, that people can go check out, um. Yeah. Um, I think it's really cool to have McDonald Miller as a partner. Um, it's our, our first sort of regional contractor to come in and say, like, I'm, I want to tell smart building stories, which is really, really cool to see.
Um, the, the investment in smart buildings from you. We, we expected to see it more and more, but, um, it's not a startup with a venture backed marketing department that is doing this. It's, it's them basically saying like, here's, here's our role in, in the ecosystem, and I think that's really cool.
Brad Bonavida: We, we just watched them demonstrate how they kind of pitch their FDD solution to a new, uh, building [00:07:00] owner.
And it's spectacular. Like just the way that they've integrated FDD into all their services. So there's a fault and like they're taking care of it before anybody even knows that it's an issue is it's pretty incredible. They're cutting edge on that for sure.
James Dice: Awesome. Alright, let's jump. What we've published in the last few weeks.
While, while we took some time off, we didn't stop publishing. So we have a, a good couple of essays if you were also on summer vacation. Definitely come check these essays out. Brad, you wanna start with the
Brad Bonavida: AI one? Yeah, sure. So this was for our, our digitizing operations and maintenance audience. So like facility managers, those type of people.
Um, it's funny, I was thinking back to. Nexus Con 2024 and we, we sort of avoided the term AI 'cause there was so much hype around it, like everyone was just saying it and no one really knew what they were talking about. And the amount, it's changed now that we have like, so many real world examples. Um, so that's kind of what inspired this piece is like.[00:08:00]
Demystify how people are using AI for facility management to somebody who might not be like deep in the weeds with it every day.
James Dice: When we, when you say ai, you're talking about what has been built on top of the generative AI wave. I think some of the people that commented on this piece we're, we're telling us to basically remember that we've been using.
Regression modeling and expert systems like FDD for decades. It's not like we're saying like, this is now ai. Right. We're just saying like, here's the ways in which generative AI is being, you know, totally integrated in with products.
Brad Bonavida: Yeah, I know exactly what you're talking about. There was some good points brought up about Yeah.
How regression models have been used in the past and I, you know, I made a big infographic about like the different versions of this, and that comment made me think you could actually do like a fork of this infographic is where like. AI in terms of like LLMs and Chacha PT came into the industry, but there's this other fork where like [00:09:00] regression models have been doing similar things for decades.
Yeah. So totally recognize that valid point. Um, when, when I did a deep dive, this is not like a industry standard, but my brain thought there's, you know, three versions that I saw in terms of like how FMS are applying. Generative AI and agentic AI today. Again, aside from the regression models, so just really quick, you know, version one is like your standard just using Chate or Gemini as your support throughout your job, which is effective, but it doesn't know anything about your building.
It's just answering questions. It can't make decisions. Um, and then we are seeing a lot of. Technology vendors who are adding retrieval, augmented generation on top of that. So think about like an LLM, like Chacha, PT or Gemini or whatever. But it also knows all this data that's in your building at your dating data layer.
Like I don't know what, what the set points were yesterday and what the temperatures were yesterday and what the fault codes were yesterday. So the information it's giving you back is much more [00:10:00] rich with. What is relevant for your building? And then on last, I put my version three as when AG Agentic AI comes in and I see AG agentic ai as simply put, it's starting to make decisions for you.
So if it knows something's wrong with the building, it can actually react to that and either make a command or, you know, um, integrate or force a workflow to, to occur so that someone has to go out and do some maintenance. Um, so check out the article. Uh, lemme know what you think, but. That's one of the big reasons we wanted to get Devin in here is 'cause um, Devin, I know you were gonna, you know, at Nexus Con you're gonna talk a little bit how Lockheed's starting to, you know, get into this and using AI for operations and maintenance.
So what did you think did, does this article apply to the way you guys are doing it? How are you guys doing it? Um, what can you tell us?
Devan Tracy: Absolutely I, I liked how you broke up the different versions and I would say at Lockheed we're between version one and two. Right now, we're on the simpler end, but I think there's a lot of value [00:11:00] to be harvested there, so I don't think there's any rush for us to try to jump to version four when we haven't even scratched the surface.
On version two, I would say. I agree with your vision, Brad, kind of where the industry's going with self-healing buildings, you know, way finding for assets. I remember seeing, I think Brainbox did a, a video where they had a technician driving to an air handling unit out on the roof and while they were driving they pinged Aria and or asking, you know, what, what year, what model, what's wrong with it?
What alarms. What filter do I need to bring up? And so see that in the future. But I also see so many gaps in the underlying data set, and I think that's one of the other articles we're gonna get to. So again, that's why we're really focusing on the basics. I'll share a couple examples. So. We're going through a huge risk and [00:12:00] resiliency exercise right now for maintenance, and we're really honing in on what are our critical assets and back to basics.
What is our spare parts plan, what are the criticalities? And so we are using our internal LL and to help fill in some of those gaps to help free a draft troubleshooting guide to help create a draft. PM plan. What spare parts do we need? How often do we need to maintain that asset? So that's one example.
Second example I would say is querying building code. So, you know, for example, we'll put in a prompt, you know, I'm a facilities engineer, I need to utilize UFC when designing specs per facility. What type of piping should I use according to UFC? And also include the reference, right? A lot of trust, but verify types of activities.
I would also say we are using [00:13:00] our internal LLM to summarize survey results that we get. You know, those. Basically iPads you would see out in the airports. We have similar setup and there's just too many responses to go through manually, so that's a really low hanging fruit example, writing job recs, I mean, that applies to any industry writing scripts for videos that we're creating.
I should say drafting, not writing right. And then a lot of bluff, like the bottom, bottom line up front for an executive who has no idea what Laura Wan is, let's, let's utilize that to help quickly summarize that in a couple bullets. So again, a lot of value at the simple level. Mm-hmm.
James Dice: Yeah. Can you go back to the surveys?
Who's taking the surveys?
Devan Tracy: Employees. Building. Uh, yeah.
James Dice: Occupants.
Devan Tracy: We, we put them in occupants. Thank you. We're putting them in cafeterias, break rooms, bathrooms.
James Dice: And then what do you do with that data? Um. How does that sort of feed into the program? [00:14:00]
Devan Tracy: So one of our interns actually just created a chat bot to automatically process the data weekly, summarize some of the key complaints, and that goes back to the site lead at each facility.
For them to react to. And we specifically said we're, we can't just put these surveys out there if we're not planning to do anything with them, because people will stop responding obviously. So that's the current process.
James Dice: Cool. Interesting. Yeah. Yeah. Really cool. So what does it look like as you guys try to like go to, to what Brad is calling stage two or, or getting into that rag rag world?
Devan Tracy: Definitely filling in the data gaps. I mean, our buildings are, I think 50% of them are 50 years or older, and we have not done a great job documenting standards, documenting [00:15:00] design specs, and so in order to get to what Brad was talking about, I think we're gonna need to stand up a team two. One, maybe two teams, one to work on what's the process from now moving forward for any new builds, and then how do we start to clean up our backlog in a strategic way?
'cause we're not gonna be able to. Everything. But you know, probably based on criticality,
Brad Bonavida: the, the scary part about that too is, you know, for anyone who's using LLMs and a lot knows that they're all probability based and they don't tell you that they don't know something, they just fill it in with their best guess.
So if you're concerned about the data from your rag model and now you're feeding that with an LLM, that's just gonna give you its best answer. It could. Be hard to uncover what's real and what's not real if you're already worried about the underlying data. 'cause it's going to try to answer the question.
It's not gonna say, I don't know. We don't have that information on that unit, which is kind of scary. [00:16:00] Yeah,
Devan Tracy: and we actually had a deep dive in person last week. All the controls engineers for our business area got together and we were like, okay, this is the standard moving forward and we're gonna agree and, and hold the vendors to it anytime we do a BMS upgrade, et cetera.
And so that was pretty cool. That was our first kind of data cleansing exercise as a beyond just my team. That extended out to the controls engineers.
James Dice: So that's a great segue. I feel like we should go into the next article here. Um, this is an article that I wrote. Um. Published in June, I think it was, um, before there, there's, in my brain, it's before vacation and after vacation.
This was definitely before vacation. Um, and it was, it's called AI Won't Fix Your Buildings Data, how to actually make it useful and. We started out on this when we started, put this in our editorial calendar and started planning out this piece. It was originally supposed to be called Dirty Jobs. [00:17:00] Um, the, you know, the dirty job of cleaning up, uh, dirty data in your building.
It's kind of what you're describing. Um, and when I did the interviews, I sort of, um. You know, we interviewed three or four for four experts as part of this, and what, what each of them kept saying is like, AI is not actually gonna solve this problem here. And it makes a lot of sense. It's like, it's, it, it doesn't know what, um, that sensor that's mislabeled is.
AI is never gonna be able to figure that out. It might be able to take a well named point and add tags to it, but it's not gonna be able to. Do something with that piece of data, with no label or with no modeling or with no relationships added to the data. Um, it definitely can't, and like you're saying, Devin, in the 50-year-old buildings with a lack of digital systems, it definitely can't do anything there.
Right? And so the articles about sort of building this. [00:18:00] Data model of your building up over time. And it's actually kind of, it's unfortunate from where we're at in industry, but it's kind of the building owner's responsibility. Because right now the vendors aren't necessarily taking that responsibility across projects.
And sometimes, as you're saying, your meeting that happened last, last week or whenever, the control system's not being installed with. The data model coming with it, right? It's like the contractors are not taking that as one of their pieces of responsibility, and so this article is about, you know, what vendors are doing, what building owners need to be doing to like basically stand up for themselves and make it so that they have this clean set of data and, and more specifically the data model that then enables AI applications to sit on top of it.
Right. It sounds like a lot of what you've been thinking about right now, Devin.
Devan Tracy: Yes, spot on. I would. What is gonna incentivize vendors to do that, [00:19:00] in your opinion?
James Dice: I think nothing, and I think what's gonna solve the problem is that it's just simply built into the workflows of like, I can't. Set this control system up without one of the outputs being a model that, you know, fully explains the data here.
And so in the article we, we quoted Scott from J two Innovations. That's how the system works. When you set up the control system, it automatically. Sets up haystack tags for your data, right? 75 F does the same thing. Um, I think if we dug harder, we could probably find other control systems that, that do that as well.
Um, we also quoted and cited, um, normal framework. So Steven Dawson, Haggerty talked about how you can't, you can't set up the data layer or you can't set up the integration, um, gateway without. Also modeling and ta and tagging at the same time. [00:20:00] So, so just like baking those into the deployment workflows. Um, FDD is the same way, right?
You can't set up FDD without modeling the data, but what we're trying to do is basically incentivize the players to do it further and further down the stack, right? Um, so that there's not so much effort, uh, higher up the stack.
Brad Bonavida: It's, it's such a good segue from the last article, like if you just started at the like, oh, let's just push AI into this, and you like, the hot take is if you don't have your data layer set up, or you know, you don't have an ontology that makes sense.
Like AI is just gonna expose all of that because you're just gonna be asking questions about data that doesn't make any sense. So, yeah. The other one that came to mind was onboard data. When you were listing companies who were doing this, they got a pretty cool tool of like put your, whatever data you have into their system and it'll, you know, work through that to create haystack and Google DB with probabilities of like how, what each point should be named in those ontologies, which is pretty cool.[00:21:00]
James Dice: We've reached a point in our company's growth where Brad has seen vendors demos and I haven't, so I actually don't. I, I don't know about that yet. So, uh, I'm excited to learn about it. Yeah.
Brad Bonavida: Yeah. We'll have to talk about
James Dice: it. We'll
Brad Bonavida: check it out.
James Dice: Um, so, so Devin, and when you think about like, making this happen in all of the buildings that you guys are renovating slash managing and not planning on renovating, how, how are you thinking about this like.
Of building out that model over time.
Devan Tracy: Yeah, I mean, just to hear that some vendors are, are starting to do this is is huge because it is, it is an unsustainable amount of work to play catch up on, like you were saying. And so to be able to just say, you know, I want Project Haystack or Brick, or whatever, and build in that whole model from the beginning.
Just to have vendors understand what that means and, and do it, I think is a, a big first step. [00:22:00] Yeah. We've had haystack in our requirements for several years, and oftentimes we've had to explain to them what that is. Yeah. And that's kind of the base, the most basic level of what you're talking about.
James Dice: Yeah.
So then it's like, do they know what it is, step one. And is their system capable of exposing those tags? Step two. Um. Yeah, fascinating. One of the things that came up in the interviews that I learned just by writing this article is that there's this gray area that Andrew Rogers called, um, digital commissioning.
And so you have, a lot of times you'll have your commissioning agent if you've chosen to have a commissioning agent on your project. Um. You have an MSI usually, uh, if you've chosen to have an MSI on your project, but even if you pay for both of those scopes of work. You still might have this middle ground where the commissioning agent is thinking from a sequence of operations [00:23:00] and mechanical, is this mechanical system operating properly?
And the MSI is thinking about it, are these systems that are supposed to be talking to each other, talking to each other properly? Okay, great. But we still miss this middle ground where, which is. How is the data modeled and is that data model set up properly? And so Andrew brought up like, we actually need more scope.
Like it's not necessarily just about convincing building owners to do the commissioning, and it's not just about convincing them to have an MSI. It's also about like maybe there's this like extra piece of scope that needs to happen so that when you hand over the building, it has a data model that comes with it as well.
Yeah,
Brad Bonavida: that's, that's so true. Like when I think about traditional commissioning, I think about, uh, the commissioning agent going underneath the elevator and seeing that when the leak detector gets a little bit of water on it, the screen highlights that the leak detector's on, and that tells you nothing about any of the point [00:24:00] names and how they'd be identified in other systems or anything like that.
So it's a real gap still, for sure.
James Dice: Yep. Yep. We can do it. It's not hopeless.
Brad Bonavida: So I, I have one more question for Devin on that before we move on. Yeah. Okay. I'm just curious, like if you, if uh, uh, one of your technology vendors or a new technology vendor came to you today and pitched that they have this. LLM with rag built into their system.
Is that appealing to you today or are you like, man, we have too much to do at our data layer that we're like, I don't believe that it's going to work for us yet. We need to do more at the data layer.
Devan Tracy: We honestly don't think we could trust it, right? Mm-hmm. Because, like you were saying, it's gonna start making decisions based on so many assumptions and if, if it makes one bad decision.
That thing is gonna be shut off. Right. Especially if it's in a politically critical area or, [00:25:00] uh, one of our production areas with critical specs. So I think to get to that point, we have a lot of of work to do on our end. Mm-hmm.
James Dice: Mm-hmm. Interesting. Fascinating. Cool. All right, let's go. I wish we
Devan Tracy: could make wave the magic wand.
Yeah.
James Dice: Yes. Yes. Um, let's go to our last article briefly, Brad. So you published, um. We did a Building Owner meetup in June on space utilization technology. Um, and Brad, you then wrote an article about the top five takeaways from that discussion. So take it away.
Brad Bonavida: Yeah, sure. So building owner meetups. If you've listened to the podcast, you've heard me talk about 'em before, but we get a big Zoom meeting room full of just building owners to talk about a very particular topic, and then they do breakout rooms.
So you've got two building owners who may have never met before who find out that they're working on the. Exactly the same problem in our exchanging notes. It's pretty cool. Um, we're doing one on [00:26:00] August 20th, I believe, on AI for facilities and operations. So basically what we've been talking about here.
Um, so watch out for the invite for that. But the one in June was about space utilization. And Devin, we were excited to have you as part of this podcast because you were a big part of that conversation. And when I went back through my notes about what everyone was talking about, my first biggest takeaway was that.
People or building owners who are doing space utilization? Well, every single one started by leveraging kind of what they already had. It's not like they had nothing and they're like, fill every room with sensors. Then we'll have space utilization. They kind of like, uh, almost MacGyver the beginning of it with all these different ways that they could get data about where people are using spaces.
So I thought maybe you could talk for a bit about. I mean, please let us know what you thought about the Building Owner meetup, but specifically the examples that you guys brought up with what you're doing with your smart restrooms, I thought was, um, you know, very pertinent to leveraging what's there.
James Dice: Yeah. Toilet flushes as a proxy for occupancy is something I hadn't heard [00:27:00] before.
Devan Tracy: Yeah. So I thought that that meetup was, there was a lot of healthy discussion, right? Everyone wasn't like, oh, let's go centralize everything before we even know what to do with the data. There were, there were a lot of opinions on both sides and.
I did bring up the, the flushes and the, the hand washing data points that we already have, though might as well use that. We've been using badge data for a while. I mean, before that we literally spent sent space planners around to count. People in seats. So we've, at least we've matured beyond that badging data, at least for space planning is typically all you need.
But I do think we have more opportunity to do set point adjustments based on the occupancy. But also, why don't we, if we're already monitoring carbon dioxide, does that get us close enough to the needle? So I think we really need to assess how big are the gaps before we just go about centralizing [00:28:00]everything.
'cause. Maybe we'd rather spend that money on vibration sensors on, you know, all the critical pumps versus do we really care about knowing to that granular detail?
James Dice: Yeah. Um, I just want to ask a nerdy question, which is, how do you digitize toilet flushes?
Devan Tracy: I don't, and I'm not gonna say the vendor, but because I don't think I could say it, but it's, it's all just built into the flush valve.
James Dice: Oh, okay. Got it. We had to retrofit
Devan Tracy: the flush valve and we're not going to do that retroactively. But we did say any new bathroom, we're gonna install these. But also for leak de, it has leak, leak detection capabilities as well. Got it.
Brad Bonavida: Okay. Is that connected to we don't we care about
Devan Tracy: when people are flushing the toilets?
Brad Bonavida: Is that actively connected to your maintenance plan? So like you have janitors who are getting alarms that this bathroom's been used X amount of times.
Devan Tracy: Yes, and the idea is [00:29:00] to use that to help prioritize their time and maybe, you know, skip a bathroom if there was only one person in it all day.
Brad Bonavida: Cool that that's O one use case of space utilization is smarter maintenance for your bathrooms.
And I went back through that building owner meetup and I came up with 19 that I put on the infographic in this article, which is crazy like. You talk about space utilization and some people kind of don't know what that means. It's 'cause it can mean so many different things with what you're gonna do for that data.
One other really creative one that I heard during that building owner meetup that I hadn't heard before is, uh, providing the fire department with real time occupancy during an emergency. So you can imagine like something happens, the fire marshal shows up and you can hand that person a tablet that shows like, this is where people are in the building.
I mean, you could imagine how. How much that's going to help them in their job, for sure. Oh, yeah,
James Dice: yeah, yeah. You created a really cool people that are, that are [00:30:00] thinking about Spatialization should definitely check out Brad's infographic from, from that article. It, it displays all 19 of those in a, in a cool way.
Brad Bonavida: And I want someone to show, tell me one that's not on here. That's my challenge. I want to hear, I wanna hear what the 20th one is,
Devan Tracy: but again, for that example, it's like, can we just use. Our phones in the emergency situation. Yeah. Not that the company's gonna track every minute, but in an emergency, have the permission.
To track that way.
James Dice: I think this is a big, like tension point right now in this, you know, we've been writing about workplace tech for six months now, and when I talk to the building owners, it's all about what can I use without sensors? And you talk to the vendors, it's all about why you need new sensors.
And it's just like we're, we're seeing both sides and it's like, well wait a second, you guys aren't exactly talking about the same thing here. You know? That's a great point. So that's another, another way in which we are sitting in the middle.
Brad Bonavida: May, maybe, uh, maybe that's a, a [00:31:00] idea for sensor providers to also be providing consultative services for how they can use what's already there for certain parts of, of what they're doing.
'cause I really like that point. That's, that's definitely what we're seeing.
James Dice: It speaks to the importance of the data layer. Right? Because like, I think I wrote an article about this in 2021. All the different ways you could get occupancy data. In this case, toilet data that has a proxy for occupancy data.
But basically you're, you're, you need some way to basically aggregate all of that and figure out what the best signal for occupancy is for a given space because Devin, you guys might have different buildings that don't have those flush valves installed, right. Um, right. Yes. Uh, and, and you might have some other, you know, badge data over there, or you might go to a different building and it's got a different badge system or whatever, you know.
Um. You need to build ability to aggregate all of that together. Kind of speaks back [00:32:00] to the data model, I guess, in a way.
Brad Bonavida: Yeah.
James Dice: All right. Any, anything else on that one?
Brad Bonavida: There was, it's a whole nother topic, but there was also a really good discussion around the tension between landlords and tenants when it comes to space utilization.
'cause they both want the data, but it's also kind of a, a negotiation. Uh, it helps each one of them with negotiation if they have it and the other doesn't. Right. So you're about. To renew your lease if the landlord knows you're not using the space. They come into that negotiation differently, and if the tenant knows they need more space, they come into that differently.
So it's like they can both benefit from it, but they don't really wanna share it yet. It's a pretty interesting dynamic there for a lot of the commercial real estate land.
James Dice: Yeah, and we put them, I think we put them in the same breakout room that the occupiers and, and landlords.
Brad Bonavida: Yeah. Yeah. That's good.
James Dice: All right, cool.
Let's move to carve outs and close out this conversation. Um, so carve outs are something that's just more fun from your personal life. Could be a book, podcast, [00:33:00] new album, have all been shared. Um, I'll go first. So I was in Hawaii and I read this book, um, called The Overstory by Richard Powers and it's fiction.
And I don't read a lot of fiction. Um. Generally when I'm reading, I'm like trying to teach myself how to run a company for the first time. And um, this time I was like on the, in the plane or on the beach just reading this, this novel and I could not recommend this novel enough. It's about trees and, but it's about humans' relationship with trees.
And it goes all the way from like, and it, it's, it's not historical fiction, but it's based on. Things that have actually happened, like it, it covers the Pacific Northwest Timber Wars, Brad, which you may have read about, that happened in the late eighties and nineties, like before you were born. Um, people were fighting over these old growth trees in the Pacific Northwest.
And, um, it tells this story [00:34:00] over decades. You know, maybe it's 500 pages or whatever, but like. It's kind of like telling the story of humans from the perspective of trees while telling the story of trees from the perspective of humans all at once, and it, it mixes in all these different people's stories all together.
It's just fascinating and he is written other books as well and I'm gonna move on to those. There's one about the ocean and yeah, so The Overstory by Richard Powers, definitely check it out if you want. That sounds great. A change of pace. Yes,
Devan Tracy: I'll add it to my list. Yeah, I like it. And I'll also
James Dice: plug reading a book and then talking to chat GPT about that book.
So I finished the book. A flight home from Hawaii is six hours. I finished the book like with like five hours to go, and I'm not kidding you guys. I bought internet on the plane so that I could talk to Chachi bt about trees and I literally talked to Chachi Bt for like four of the remaining five hours about trees.
So I know. Wow. I [00:35:00] know all about the timber wars. I know, I know so much. So let's ask me later. Devin, you wanna go next?
Devan Tracy: Sure. That's great. I also asked Chet GBT to remind me what some of my book was about, because I've been reading this one page at a time, basically. Because I don't want it to end. Oh. But it's, it's called Joyful by Ingrid Lee, and the subtitle is The Surprising Power of Ordinary Things to Create Extraordinary Happiness.
Rosy Khalife: Hmm.
Devan Tracy: I also love reading nonfiction as James, you were saying, and I love studying happiness and joy, and she just reminds, she's actually an interior designer, so a lot of the book is how do you design spaces to inspire joy, whether that's. The colors that you choose or the, the shapes of objects, but then also things like looking at the clouds for a few minutes.
Or my favorite is watching the sunset every night, free art, and I'm gonna [00:36:00] do that tonight for my birthday. So I love, I love just these gentle reminders of finding joy every day.
James Dice: It's your birthday today.
Devan Tracy: It's my birthday today. Oh my gosh. Happy
James Dice: birthday. You're spending your birthday with us. That's awesome.
Brad Bonavida: Wow. That's a, I like how you slipped that in there, right at the end of the podcast. Yeah. And it is in fact your birthday fact. Your birthday. Yeah. That's funny.
James Dice: Well, happy birthday Devin. Yeah, happy. Thank you. We'll spare the listeners, Brad and I singing voice and just tell you happy birthday. No one wants that.
Brad Bonavida: Yeah.
James Dice: Uh, my birthday was on Saturday, so I'm a fellow.
Brad Bonavida: Oh, dang. I didn't know that. He belated. Happy birthday, James. Thank you.
James Dice: Yes, thank you. Um, all right, Brad, do you wanna go? Yeah,
Brad Bonavida: I got a fun fact. So I was also in Hawaii. I was not with James, different island. I didn't know to Hawaii before I went to Kauai.
Um, which it was spectacular. We had a great time, but if you'd never been to Kauai, the fun fact about Kauai is that there's chickens everywhere, like literally everywhere. And so I did a [00:37:00] small amount of research to figure out what the story was there. And apparently like in the late eighties, there was a pretty bad hurricane.
Released all the chickens from their coops and uh, they started to cross breeded and breed with like feral birds. But really the big story is there's no predators there. A lot of the other islands, they introduced mongoose to take care of rat populations, rat infestations, but they don't have the mongoose on Kauai.
So you, I mean, we were like. Five miles into a jungle hike and there's like a rooster that like walks across the trail. It's wild. They're, they're everywhere. Um, so you can always tell somebody who just got to the island too, 'cause like, they'll be there and they'll be like, whoa, look at all those chickens.
And like, you definitely only say that for the first like six hours you're there until you realize so many literally everywhere. But it's, it's kind of cool. It's like on all of their souvenir stuff and stuff, they've embraced the chicken as kind of their, uh, their mascot.
James Dice: Yep. Yep. I was on Oahu and there's also lots of chickens there, but they do have, [00:38:00] they do have the Mongos.
Yes. Yeah.
Brad Bonavida: Alright,
James Dice: all this has been a fun convo. Thanks for joining us on your birthday, Devin. Yeah,
Brad Bonavida: thanks Devin. Appreciate all the insights.
Rosy Khalife: Okay, friends. Thank you for listening to this episode. As we continue to grow our global community of change makers, we need your help. For the next couple of months, we're challenging our listeners to share a link to their favorite Nexus episode on LinkedIn with a short post about why you listen. It would really, really help us out.
Make sure to tag us in the post so we can see it. Have a good one.
Welcome to Nexus, a newsletter and podcast for smart people applying smart building technology—hosted by James Dice. If you’re new to Nexus, you might want to start here.
The Nexus podcast (Apple | Spotify | YouTube | Other apps) is our chance to explore and learn with the brightest in our industry—together. The project is directly funded by listeners like you who have joined the Nexus Pro membership community.
You can join Nexus Pro to get a weekly-ish deep dive, access to the Nexus Vendor Landscape, and invites to exclusive events with a community of smart buildings nerds.
Sign-up (or refer a friend!) to the Nexus Newsletter.
Learn more about the Smart Building Strategist Course and the Nexus Courses Platform.
Check out the Nexus Labs Marketplace.
Learn more about Nexus Partnership Opportunities.
Episode 183 is a conversation with James Dice and Brad Bonavida from Nexus Labs, as well as Devan Tracy from Lockheed Martin.
Episode 183 is a conversation with James Dice and Brad Bonavida from Nexus Labs, as well as Devan Tracy from Lockheed Martin. In this episode of the Nexus Podcast, the Nexus Labs team breaks down the top stories relevant to energy managers, facility managers, IT/OT managers, and workplace managers.
Introduction (0:50)
At the Nexus (3:23)
Digitizing Operations & Maintenance (07:30)
Integrating, Connecting, and Securing Devices (16:40)
Smart Building Champions (25:28)
Sign off (32:55)
Music credits: There Is A Reality by Common Tiger—licensed under an Music Vine Limited Pro Standard License ID: S706971-16073.
Note: transcript was created using an imperfect machine learning tool and lightly edited by a human (so you can get the gist). Please forgive errors!
James Dice: [00:00:00] Hey friends, if you like the Nexus Podcast, the best way to continue the learning is to join our community. There are three ways to do that. First, you can join the Nexus Pro membership. It's our global community of smart billing professionals. We have monthly events, paywall, deep dive content, and a private chat room, and it's just $35 a month.
Second, you can upgrade from the pro membership to our courses offering. It's headlined by our flagship course, the Smart Building Strategist, and we're building a catalog of courses taught by world leading experts on each topic under the smart buildings umbrella. Third, and finally, our marketplace is how we connect leading vendors with buyers looking for their solutions.
The links are below in the show notes. And now let's go onto the podcast.
Welcome back to the Nexus podcast. It's James here. I have Brad. We're missing Rosie this week. Hello. Hello. Rosie's on vacation. Uh, and then we have a special guest who I'll introduce [00:01:00] in just a second. Um. This podcast is designed to give you a preview of everything going on, um, in our content at our events, uh, and obviously what we're doing right now at at Nexus.
So, um, if you want the full experience though, this is just a preview, go over to our website and sign up for the newsletter and that will give you everything we publish each week along with everything going on with our events. Um, our special guest today is, uh, Devin Tracy, smart Buildings lead at Lockheed Martin.
Devin, do you wanna introduce yourself?
Devan Tracy: Sure. Hi everybody. Thanks for having me, James. And Brad, great to be here. I have been with Lockheed for. 12 years. My background is mechanical engineering, got a PE in hvac and I started the Smart Buildings program I think back in 2018. We modeled hours after Google. I had the opportunity to meet Darryl Smith back in the day and [00:02:00] we've been going ever since, starting with FDD and expanding out to other industrial internet of things technologies.
So I love Nexus and. Huge supporter of this community,
James Dice: and you guys have thanks to yourself and all your hard work. Um, and your team, obviously, um, you guys have one of the most mature smart buildings programs out there, so we've been thinking a lot about where all of the building owners in our community, where they sit in terms of maturity of their programs.
And what we learned next last year at, at Nexus Con was that a lot of building owners showed up. We had over 62 organizations. A lot of building owners showed up and were kind of like deer in headlights a little bit, like hadn't really got their programs off the ground and I think yours, they can look at and say like, well, where can we be in eight years?
And it's probably fun for you to look back on the journey.
Devan Tracy: It is. Yeah. It started out with just me and we proved [00:03:00]the value pretty quickly and. Then the need for a team to actually sustain the program was, it was realized pretty quickly that this, you can't just make wave a magic wand and, and hope that the technology will fix everything.
And so that's been crucial to our success.
James Dice: All right. Well, we're excited to get into some AI stuff with you in just a second. Let's talk about what's been going on at Nexus. Um, we've been on vacation a little bit, so all of us took a couple weeks off. Uh, we're kind of resting up for a big push ahead of Nexus Con this year.
Um, Brad, you want to, you want to, uh, talk about what's been going on?
Brad Bonavida: Yeah, I mean, apologies for the delay in the podcast publishing here. I know like when I listen to podcasts a lot and then someone doesn't produce on their regular schedule, you're like, what's going on? So yeah, we were enjoying our July, but we're back in the saddle now getting ready for Nexus Con.
Um, aside from that, we do have some. Pretty exciting new partnerships. We've been working with some [00:04:00] companies that, uh, are really leaders in the space that we're excited about. Just to rattle off a couple, um, we're working with McDonald Miller, which is a big mechanical, electrical, plumbing and HVAC controls provider in the, uh, Pacific Northwest.
So they're kind of our first, like we're calling it technology enabled services partner. They're really like leading. They're a leading provider of adding these technologies like FDD into the services that they're offering to, uh, their clients, which is really cool. Um, we're partnering with ar, which is a massive, uh, consultant, but they've got a really cool smart buildings division that's doing amazing stuff, helping big.
Uh, commercial real estate owners get to their outcomes by, you know, creating a data layer and getting all their data in one place and putting smart applications above that. And then lastly, uh, J two Innovations as well, which is a Siemens company and they are doing HVAC controls in a way that kind of no one else is in.
I mean, maybe James, you can add something [00:05:00] there. I know you just talked to Scott, but the way that they are, um, enabling the entire technology stack at the HVAC control layer is pretty unique.
James Dice: Yep. Yep. Um, we will talk about them at some point in more detail. I think it's really cool how they templatize.
Control system. So
Brad Bonavida: yeah,
James Dice: the old world of controls, and it was funny talking to Scott a couple different times, but, um, he's been doing this for 30 years, but it, it is funny to talk about how this is like cutting edge, but, um, a typical, the old controls world is that you might program a VAV to talk to an air handler, and a, a programmer might do that.
Like they're doing it from scratch, right? And it might be this like custom expensive solution and they're, they're developing the ability to just say, well, you know what? That's a VAV with hot water reheat. It's talking to this air handler and the system just basically, uh. Does programming, graphics, alarms, f, d, D points, all that based [00:06:00] off of predetermined templates because as we know, the systems are the same across buildings.
There's just a little bit of variation. And so using the ability for templates and wizards to, and we did a podcast with them like going through that a couple years ago, that people can go check out, um. Yeah. Um, I think it's really cool to have McDonald Miller as a partner. Um, it's our, our first sort of regional contractor to come in and say, like, I'm, I want to tell smart building stories, which is really, really cool to see.
Um, the, the investment in smart buildings from you. We, we expected to see it more and more, but, um, it's not a startup with a venture backed marketing department that is doing this. It's, it's them basically saying like, here's, here's our role in, in the ecosystem, and I think that's really cool.
Brad Bonavida: We, we just watched them demonstrate how they kind of pitch their FDD solution to a new, uh, building [00:07:00] owner.
And it's spectacular. Like just the way that they've integrated FDD into all their services. So there's a fault and like they're taking care of it before anybody even knows that it's an issue is it's pretty incredible. They're cutting edge on that for sure.
James Dice: Awesome. Alright, let's jump. What we've published in the last few weeks.
While, while we took some time off, we didn't stop publishing. So we have a, a good couple of essays if you were also on summer vacation. Definitely come check these essays out. Brad, you wanna start with the
Brad Bonavida: AI one? Yeah, sure. So this was for our, our digitizing operations and maintenance audience. So like facility managers, those type of people.
Um, it's funny, I was thinking back to. Nexus Con 2024 and we, we sort of avoided the term AI 'cause there was so much hype around it, like everyone was just saying it and no one really knew what they were talking about. And the amount, it's changed now that we have like, so many real world examples. Um, so that's kind of what inspired this piece is like.[00:08:00]
Demystify how people are using AI for facility management to somebody who might not be like deep in the weeds with it every day.
James Dice: When we, when you say ai, you're talking about what has been built on top of the generative AI wave. I think some of the people that commented on this piece we're, we're telling us to basically remember that we've been using.
Regression modeling and expert systems like FDD for decades. It's not like we're saying like, this is now ai. Right. We're just saying like, here's the ways in which generative AI is being, you know, totally integrated in with products.
Brad Bonavida: Yeah, I know exactly what you're talking about. There was some good points brought up about Yeah.
How regression models have been used in the past and I, you know, I made a big infographic about like the different versions of this, and that comment made me think you could actually do like a fork of this infographic is where like. AI in terms of like LLMs and Chacha PT came into the industry, but there's this other fork where like [00:09:00] regression models have been doing similar things for decades.
Yeah. So totally recognize that valid point. Um, when, when I did a deep dive, this is not like a industry standard, but my brain thought there's, you know, three versions that I saw in terms of like how FMS are applying. Generative AI and agentic AI today. Again, aside from the regression models, so just really quick, you know, version one is like your standard just using Chate or Gemini as your support throughout your job, which is effective, but it doesn't know anything about your building.
It's just answering questions. It can't make decisions. Um, and then we are seeing a lot of. Technology vendors who are adding retrieval, augmented generation on top of that. So think about like an LLM, like Chacha, PT or Gemini or whatever. But it also knows all this data that's in your building at your dating data layer.
Like I don't know what, what the set points were yesterday and what the temperatures were yesterday and what the fault codes were yesterday. So the information it's giving you back is much more [00:10:00] rich with. What is relevant for your building? And then on last, I put my version three as when AG Agentic AI comes in and I see AG agentic ai as simply put, it's starting to make decisions for you.
So if it knows something's wrong with the building, it can actually react to that and either make a command or, you know, um, integrate or force a workflow to, to occur so that someone has to go out and do some maintenance. Um, so check out the article. Uh, lemme know what you think, but. That's one of the big reasons we wanted to get Devin in here is 'cause um, Devin, I know you were gonna, you know, at Nexus Con you're gonna talk a little bit how Lockheed's starting to, you know, get into this and using AI for operations and maintenance.
So what did you think did, does this article apply to the way you guys are doing it? How are you guys doing it? Um, what can you tell us?
Devan Tracy: Absolutely I, I liked how you broke up the different versions and I would say at Lockheed we're between version one and two. Right now, we're on the simpler end, but I think there's a lot of value [00:11:00] to be harvested there, so I don't think there's any rush for us to try to jump to version four when we haven't even scratched the surface.
On version two, I would say. I agree with your vision, Brad, kind of where the industry's going with self-healing buildings, you know, way finding for assets. I remember seeing, I think Brainbox did a, a video where they had a technician driving to an air handling unit out on the roof and while they were driving they pinged Aria and or asking, you know, what, what year, what model, what's wrong with it?
What alarms. What filter do I need to bring up? And so see that in the future. But I also see so many gaps in the underlying data set, and I think that's one of the other articles we're gonna get to. So again, that's why we're really focusing on the basics. I'll share a couple examples. So. We're going through a huge risk and [00:12:00] resiliency exercise right now for maintenance, and we're really honing in on what are our critical assets and back to basics.
What is our spare parts plan, what are the criticalities? And so we are using our internal LL and to help fill in some of those gaps to help free a draft troubleshooting guide to help create a draft. PM plan. What spare parts do we need? How often do we need to maintain that asset? So that's one example.
Second example I would say is querying building code. So, you know, for example, we'll put in a prompt, you know, I'm a facilities engineer, I need to utilize UFC when designing specs per facility. What type of piping should I use according to UFC? And also include the reference, right? A lot of trust, but verify types of activities.
I would also say we are using [00:13:00] our internal LLM to summarize survey results that we get. You know, those. Basically iPads you would see out in the airports. We have similar setup and there's just too many responses to go through manually, so that's a really low hanging fruit example, writing job recs, I mean, that applies to any industry writing scripts for videos that we're creating.
I should say drafting, not writing right. And then a lot of bluff, like the bottom, bottom line up front for an executive who has no idea what Laura Wan is, let's, let's utilize that to help quickly summarize that in a couple bullets. So again, a lot of value at the simple level. Mm-hmm.
James Dice: Yeah. Can you go back to the surveys?
Who's taking the surveys?
Devan Tracy: Employees. Building. Uh, yeah.
James Dice: Occupants.
Devan Tracy: We, we put them in occupants. Thank you. We're putting them in cafeterias, break rooms, bathrooms.
James Dice: And then what do you do with that data? Um. How does that sort of feed into the program? [00:14:00]
Devan Tracy: So one of our interns actually just created a chat bot to automatically process the data weekly, summarize some of the key complaints, and that goes back to the site lead at each facility.
For them to react to. And we specifically said we're, we can't just put these surveys out there if we're not planning to do anything with them, because people will stop responding obviously. So that's the current process.
James Dice: Cool. Interesting. Yeah. Yeah. Really cool. So what does it look like as you guys try to like go to, to what Brad is calling stage two or, or getting into that rag rag world?
Devan Tracy: Definitely filling in the data gaps. I mean, our buildings are, I think 50% of them are 50 years or older, and we have not done a great job documenting standards, documenting [00:15:00] design specs, and so in order to get to what Brad was talking about, I think we're gonna need to stand up a team two. One, maybe two teams, one to work on what's the process from now moving forward for any new builds, and then how do we start to clean up our backlog in a strategic way?
'cause we're not gonna be able to. Everything. But you know, probably based on criticality,
Brad Bonavida: the, the scary part about that too is, you know, for anyone who's using LLMs and a lot knows that they're all probability based and they don't tell you that they don't know something, they just fill it in with their best guess.
So if you're concerned about the data from your rag model and now you're feeding that with an LLM, that's just gonna give you its best answer. It could. Be hard to uncover what's real and what's not real if you're already worried about the underlying data. 'cause it's going to try to answer the question.
It's not gonna say, I don't know. We don't have that information on that unit, which is kind of scary. [00:16:00] Yeah,
Devan Tracy: and we actually had a deep dive in person last week. All the controls engineers for our business area got together and we were like, okay, this is the standard moving forward and we're gonna agree and, and hold the vendors to it anytime we do a BMS upgrade, et cetera.
And so that was pretty cool. That was our first kind of data cleansing exercise as a beyond just my team. That extended out to the controls engineers.
James Dice: So that's a great segue. I feel like we should go into the next article here. Um, this is an article that I wrote. Um. Published in June, I think it was, um, before there, there's, in my brain, it's before vacation and after vacation.
This was definitely before vacation. Um, and it was, it's called AI Won't Fix Your Buildings Data, how to actually make it useful and. We started out on this when we started, put this in our editorial calendar and started planning out this piece. It was originally supposed to be called Dirty Jobs. [00:17:00] Um, the, you know, the dirty job of cleaning up, uh, dirty data in your building.
It's kind of what you're describing. Um, and when I did the interviews, I sort of, um. You know, we interviewed three or four for four experts as part of this, and what, what each of them kept saying is like, AI is not actually gonna solve this problem here. And it makes a lot of sense. It's like, it's, it, it doesn't know what, um, that sensor that's mislabeled is.
AI is never gonna be able to figure that out. It might be able to take a well named point and add tags to it, but it's not gonna be able to. Do something with that piece of data, with no label or with no modeling or with no relationships added to the data. Um, it definitely can't, and like you're saying, Devin, in the 50-year-old buildings with a lack of digital systems, it definitely can't do anything there.
Right? And so the articles about sort of building this. [00:18:00] Data model of your building up over time. And it's actually kind of, it's unfortunate from where we're at in industry, but it's kind of the building owner's responsibility. Because right now the vendors aren't necessarily taking that responsibility across projects.
And sometimes, as you're saying, your meeting that happened last, last week or whenever, the control system's not being installed with. The data model coming with it, right? It's like the contractors are not taking that as one of their pieces of responsibility, and so this article is about, you know, what vendors are doing, what building owners need to be doing to like basically stand up for themselves and make it so that they have this clean set of data and, and more specifically the data model that then enables AI applications to sit on top of it.
Right. It sounds like a lot of what you've been thinking about right now, Devin.
Devan Tracy: Yes, spot on. I would. What is gonna incentivize vendors to do that, [00:19:00] in your opinion?
James Dice: I think nothing, and I think what's gonna solve the problem is that it's just simply built into the workflows of like, I can't. Set this control system up without one of the outputs being a model that, you know, fully explains the data here.
And so in the article we, we quoted Scott from J two Innovations. That's how the system works. When you set up the control system, it automatically. Sets up haystack tags for your data, right? 75 F does the same thing. Um, I think if we dug harder, we could probably find other control systems that, that do that as well.
Um, we also quoted and cited, um, normal framework. So Steven Dawson, Haggerty talked about how you can't, you can't set up the data layer or you can't set up the integration, um, gateway without. Also modeling and ta and tagging at the same time. [00:20:00] So, so just like baking those into the deployment workflows. Um, FDD is the same way, right?
You can't set up FDD without modeling the data, but what we're trying to do is basically incentivize the players to do it further and further down the stack, right? Um, so that there's not so much effort, uh, higher up the stack.
Brad Bonavida: It's, it's such a good segue from the last article, like if you just started at the like, oh, let's just push AI into this, and you like, the hot take is if you don't have your data layer set up, or you know, you don't have an ontology that makes sense.
Like AI is just gonna expose all of that because you're just gonna be asking questions about data that doesn't make any sense. So, yeah. The other one that came to mind was onboard data. When you were listing companies who were doing this, they got a pretty cool tool of like put your, whatever data you have into their system and it'll, you know, work through that to create haystack and Google DB with probabilities of like how, what each point should be named in those ontologies, which is pretty cool.[00:21:00]
James Dice: We've reached a point in our company's growth where Brad has seen vendors demos and I haven't, so I actually don't. I, I don't know about that yet. So, uh, I'm excited to learn about it. Yeah.
Brad Bonavida: Yeah. We'll have to talk about
James Dice: it. We'll
Brad Bonavida: check it out.
James Dice: Um, so, so Devin, and when you think about like, making this happen in all of the buildings that you guys are renovating slash managing and not planning on renovating, how, how are you thinking about this like.
Of building out that model over time.
Devan Tracy: Yeah, I mean, just to hear that some vendors are, are starting to do this is is huge because it is, it is an unsustainable amount of work to play catch up on, like you were saying. And so to be able to just say, you know, I want Project Haystack or Brick, or whatever, and build in that whole model from the beginning.
Just to have vendors understand what that means and, and do it, I think is a, a big first step. [00:22:00] Yeah. We've had haystack in our requirements for several years, and oftentimes we've had to explain to them what that is. Yeah. And that's kind of the base, the most basic level of what you're talking about.
James Dice: Yeah.
So then it's like, do they know what it is, step one. And is their system capable of exposing those tags? Step two. Um. Yeah, fascinating. One of the things that came up in the interviews that I learned just by writing this article is that there's this gray area that Andrew Rogers called, um, digital commissioning.
And so you have, a lot of times you'll have your commissioning agent if you've chosen to have a commissioning agent on your project. Um. You have an MSI usually, uh, if you've chosen to have an MSI on your project, but even if you pay for both of those scopes of work. You still might have this middle ground where the commissioning agent is thinking from a sequence of operations [00:23:00] and mechanical, is this mechanical system operating properly?
And the MSI is thinking about it, are these systems that are supposed to be talking to each other, talking to each other properly? Okay, great. But we still miss this middle ground where, which is. How is the data modeled and is that data model set up properly? And so Andrew brought up like, we actually need more scope.
Like it's not necessarily just about convincing building owners to do the commissioning, and it's not just about convincing them to have an MSI. It's also about like maybe there's this like extra piece of scope that needs to happen so that when you hand over the building, it has a data model that comes with it as well.
Yeah,
Brad Bonavida: that's, that's so true. Like when I think about traditional commissioning, I think about, uh, the commissioning agent going underneath the elevator and seeing that when the leak detector gets a little bit of water on it, the screen highlights that the leak detector's on, and that tells you nothing about any of the point [00:24:00] names and how they'd be identified in other systems or anything like that.
So it's a real gap still, for sure.
James Dice: Yep. Yep. We can do it. It's not hopeless.
Brad Bonavida: So I, I have one more question for Devin on that before we move on. Yeah. Okay. I'm just curious, like if you, if uh, uh, one of your technology vendors or a new technology vendor came to you today and pitched that they have this. LLM with rag built into their system.
Is that appealing to you today or are you like, man, we have too much to do at our data layer that we're like, I don't believe that it's going to work for us yet. We need to do more at the data layer.
Devan Tracy: We honestly don't think we could trust it, right? Mm-hmm. Because, like you were saying, it's gonna start making decisions based on so many assumptions and if, if it makes one bad decision.
That thing is gonna be shut off. Right. Especially if it's in a politically critical area or, [00:25:00] uh, one of our production areas with critical specs. So I think to get to that point, we have a lot of of work to do on our end. Mm-hmm.
James Dice: Mm-hmm. Interesting. Fascinating. Cool. All right, let's go. I wish we
Devan Tracy: could make wave the magic wand.
Yeah.
James Dice: Yes. Yes. Um, let's go to our last article briefly, Brad. So you published, um. We did a Building Owner meetup in June on space utilization technology. Um, and Brad, you then wrote an article about the top five takeaways from that discussion. So take it away.
Brad Bonavida: Yeah, sure. So building owner meetups. If you've listened to the podcast, you've heard me talk about 'em before, but we get a big Zoom meeting room full of just building owners to talk about a very particular topic, and then they do breakout rooms.
So you've got two building owners who may have never met before who find out that they're working on the. Exactly the same problem in our exchanging notes. It's pretty cool. Um, we're doing one on [00:26:00] August 20th, I believe, on AI for facilities and operations. So basically what we've been talking about here.
Um, so watch out for the invite for that. But the one in June was about space utilization. And Devin, we were excited to have you as part of this podcast because you were a big part of that conversation. And when I went back through my notes about what everyone was talking about, my first biggest takeaway was that.
People or building owners who are doing space utilization? Well, every single one started by leveraging kind of what they already had. It's not like they had nothing and they're like, fill every room with sensors. Then we'll have space utilization. They kind of like, uh, almost MacGyver the beginning of it with all these different ways that they could get data about where people are using spaces.
So I thought maybe you could talk for a bit about. I mean, please let us know what you thought about the Building Owner meetup, but specifically the examples that you guys brought up with what you're doing with your smart restrooms, I thought was, um, you know, very pertinent to leveraging what's there.
James Dice: Yeah. Toilet flushes as a proxy for occupancy is something I hadn't heard [00:27:00] before.
Devan Tracy: Yeah. So I thought that that meetup was, there was a lot of healthy discussion, right? Everyone wasn't like, oh, let's go centralize everything before we even know what to do with the data. There were, there were a lot of opinions on both sides and.
I did bring up the, the flushes and the, the hand washing data points that we already have, though might as well use that. We've been using badge data for a while. I mean, before that we literally spent sent space planners around to count. People in seats. So we've, at least we've matured beyond that badging data, at least for space planning is typically all you need.
But I do think we have more opportunity to do set point adjustments based on the occupancy. But also, why don't we, if we're already monitoring carbon dioxide, does that get us close enough to the needle? So I think we really need to assess how big are the gaps before we just go about centralizing [00:28:00]everything.
'cause. Maybe we'd rather spend that money on vibration sensors on, you know, all the critical pumps versus do we really care about knowing to that granular detail?
James Dice: Yeah. Um, I just want to ask a nerdy question, which is, how do you digitize toilet flushes?
Devan Tracy: I don't, and I'm not gonna say the vendor, but because I don't think I could say it, but it's, it's all just built into the flush valve.
James Dice: Oh, okay. Got it. We had to retrofit
Devan Tracy: the flush valve and we're not going to do that retroactively. But we did say any new bathroom, we're gonna install these. But also for leak de, it has leak, leak detection capabilities as well. Got it.
Brad Bonavida: Okay. Is that connected to we don't we care about
Devan Tracy: when people are flushing the toilets?
Brad Bonavida: Is that actively connected to your maintenance plan? So like you have janitors who are getting alarms that this bathroom's been used X amount of times.
Devan Tracy: Yes, and the idea is [00:29:00] to use that to help prioritize their time and maybe, you know, skip a bathroom if there was only one person in it all day.
Brad Bonavida: Cool that that's O one use case of space utilization is smarter maintenance for your bathrooms.
And I went back through that building owner meetup and I came up with 19 that I put on the infographic in this article, which is crazy like. You talk about space utilization and some people kind of don't know what that means. It's 'cause it can mean so many different things with what you're gonna do for that data.
One other really creative one that I heard during that building owner meetup that I hadn't heard before is, uh, providing the fire department with real time occupancy during an emergency. So you can imagine like something happens, the fire marshal shows up and you can hand that person a tablet that shows like, this is where people are in the building.
I mean, you could imagine how. How much that's going to help them in their job, for sure. Oh, yeah,
James Dice: yeah, yeah. You created a really cool people that are, that are [00:30:00] thinking about Spatialization should definitely check out Brad's infographic from, from that article. It, it displays all 19 of those in a, in a cool way.
Brad Bonavida: And I want someone to show, tell me one that's not on here. That's my challenge. I want to hear, I wanna hear what the 20th one is,
Devan Tracy: but again, for that example, it's like, can we just use. Our phones in the emergency situation. Yeah. Not that the company's gonna track every minute, but in an emergency, have the permission.
To track that way.
James Dice: I think this is a big, like tension point right now in this, you know, we've been writing about workplace tech for six months now, and when I talk to the building owners, it's all about what can I use without sensors? And you talk to the vendors, it's all about why you need new sensors.
And it's just like we're, we're seeing both sides and it's like, well wait a second, you guys aren't exactly talking about the same thing here. You know? That's a great point. So that's another, another way in which we are sitting in the middle.
Brad Bonavida: May, maybe, uh, maybe that's a, a [00:31:00] idea for sensor providers to also be providing consultative services for how they can use what's already there for certain parts of, of what they're doing.
'cause I really like that point. That's, that's definitely what we're seeing.
James Dice: It speaks to the importance of the data layer. Right? Because like, I think I wrote an article about this in 2021. All the different ways you could get occupancy data. In this case, toilet data that has a proxy for occupancy data.
But basically you're, you're, you need some way to basically aggregate all of that and figure out what the best signal for occupancy is for a given space because Devin, you guys might have different buildings that don't have those flush valves installed, right. Um, right. Yes. Uh, and, and you might have some other, you know, badge data over there, or you might go to a different building and it's got a different badge system or whatever, you know.
Um. You need to build ability to aggregate all of that together. Kind of speaks back [00:32:00] to the data model, I guess, in a way.
Brad Bonavida: Yeah.
James Dice: All right. Any, anything else on that one?
Brad Bonavida: There was, it's a whole nother topic, but there was also a really good discussion around the tension between landlords and tenants when it comes to space utilization.
'cause they both want the data, but it's also kind of a, a negotiation. Uh, it helps each one of them with negotiation if they have it and the other doesn't. Right. So you're about. To renew your lease if the landlord knows you're not using the space. They come into that negotiation differently, and if the tenant knows they need more space, they come into that differently.
So it's like they can both benefit from it, but they don't really wanna share it yet. It's a pretty interesting dynamic there for a lot of the commercial real estate land.
James Dice: Yeah, and we put them, I think we put them in the same breakout room that the occupiers and, and landlords.
Brad Bonavida: Yeah. Yeah. That's good.
James Dice: All right, cool.
Let's move to carve outs and close out this conversation. Um, so carve outs are something that's just more fun from your personal life. Could be a book, podcast, [00:33:00] new album, have all been shared. Um, I'll go first. So I was in Hawaii and I read this book, um, called The Overstory by Richard Powers and it's fiction.
And I don't read a lot of fiction. Um. Generally when I'm reading, I'm like trying to teach myself how to run a company for the first time. And um, this time I was like on the, in the plane or on the beach just reading this, this novel and I could not recommend this novel enough. It's about trees and, but it's about humans' relationship with trees.
And it goes all the way from like, and it, it's, it's not historical fiction, but it's based on. Things that have actually happened, like it, it covers the Pacific Northwest Timber Wars, Brad, which you may have read about, that happened in the late eighties and nineties, like before you were born. Um, people were fighting over these old growth trees in the Pacific Northwest.
And, um, it tells this story [00:34:00] over decades. You know, maybe it's 500 pages or whatever, but like. It's kind of like telling the story of humans from the perspective of trees while telling the story of trees from the perspective of humans all at once, and it, it mixes in all these different people's stories all together.
It's just fascinating and he is written other books as well and I'm gonna move on to those. There's one about the ocean and yeah, so The Overstory by Richard Powers, definitely check it out if you want. That sounds great. A change of pace. Yes,
Devan Tracy: I'll add it to my list. Yeah, I like it. And I'll also
James Dice: plug reading a book and then talking to chat GPT about that book.
So I finished the book. A flight home from Hawaii is six hours. I finished the book like with like five hours to go, and I'm not kidding you guys. I bought internet on the plane so that I could talk to Chachi bt about trees and I literally talked to Chachi Bt for like four of the remaining five hours about trees.
So I know. Wow. I [00:35:00] know all about the timber wars. I know, I know so much. So let's ask me later. Devin, you wanna go next?
Devan Tracy: Sure. That's great. I also asked Chet GBT to remind me what some of my book was about, because I've been reading this one page at a time, basically. Because I don't want it to end. Oh. But it's, it's called Joyful by Ingrid Lee, and the subtitle is The Surprising Power of Ordinary Things to Create Extraordinary Happiness.
Rosy Khalife: Hmm.
Devan Tracy: I also love reading nonfiction as James, you were saying, and I love studying happiness and joy, and she just reminds, she's actually an interior designer, so a lot of the book is how do you design spaces to inspire joy, whether that's. The colors that you choose or the, the shapes of objects, but then also things like looking at the clouds for a few minutes.
Or my favorite is watching the sunset every night, free art, and I'm gonna [00:36:00] do that tonight for my birthday. So I love, I love just these gentle reminders of finding joy every day.
James Dice: It's your birthday today.
Devan Tracy: It's my birthday today. Oh my gosh. Happy
James Dice: birthday. You're spending your birthday with us. That's awesome.
Brad Bonavida: Wow. That's a, I like how you slipped that in there, right at the end of the podcast. Yeah. And it is in fact your birthday fact. Your birthday. Yeah. That's funny.
James Dice: Well, happy birthday Devin. Yeah, happy. Thank you. We'll spare the listeners, Brad and I singing voice and just tell you happy birthday. No one wants that.
Brad Bonavida: Yeah.
James Dice: Uh, my birthday was on Saturday, so I'm a fellow.
Brad Bonavida: Oh, dang. I didn't know that. He belated. Happy birthday, James. Thank you.
James Dice: Yes, thank you. Um, all right, Brad, do you wanna go? Yeah,
Brad Bonavida: I got a fun fact. So I was also in Hawaii. I was not with James, different island. I didn't know to Hawaii before I went to Kauai.
Um, which it was spectacular. We had a great time, but if you'd never been to Kauai, the fun fact about Kauai is that there's chickens everywhere, like literally everywhere. And so I did a [00:37:00] small amount of research to figure out what the story was there. And apparently like in the late eighties, there was a pretty bad hurricane.
Released all the chickens from their coops and uh, they started to cross breeded and breed with like feral birds. But really the big story is there's no predators there. A lot of the other islands, they introduced mongoose to take care of rat populations, rat infestations, but they don't have the mongoose on Kauai.
So you, I mean, we were like. Five miles into a jungle hike and there's like a rooster that like walks across the trail. It's wild. They're, they're everywhere. Um, so you can always tell somebody who just got to the island too, 'cause like, they'll be there and they'll be like, whoa, look at all those chickens.
And like, you definitely only say that for the first like six hours you're there until you realize so many literally everywhere. But it's, it's kind of cool. It's like on all of their souvenir stuff and stuff, they've embraced the chicken as kind of their, uh, their mascot.
James Dice: Yep. Yep. I was on Oahu and there's also lots of chickens there, but they do have, [00:38:00] they do have the Mongos.
Yes. Yeah.
Brad Bonavida: Alright,
James Dice: all this has been a fun convo. Thanks for joining us on your birthday, Devin. Yeah,
Brad Bonavida: thanks Devin. Appreciate all the insights.
Rosy Khalife: Okay, friends. Thank you for listening to this episode. As we continue to grow our global community of change makers, we need your help. For the next couple of months, we're challenging our listeners to share a link to their favorite Nexus episode on LinkedIn with a short post about why you listen. It would really, really help us out.
Make sure to tag us in the post so we can see it. Have a good one.
Head over to Nexus Connect and see what’s new in the community. Don’t forget to check out the latest member-only events.
Go to Nexus ConnectJoin Nexus Pro and get full access including invite-only member gatherings, access to the community chatroom Nexus Connect, networking opportunities, and deep dive essays.
Sign Up
Nexus Labs exists to move the smart buildings industry forward, without the fluff.
On Wednesday, we send out the renowned (and free) Nexus Newsletter.
This is a great piece!
I agree.