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 188 is a conversation with Brad Bonavida and Rosy Khalife from Nexus Labs, as well as James Coleman from Princeton University.
Episode 188 is a conversation with Brad Bonavida and Rosy Khalife from Nexus Labs, as well as James Coleman from Princeton University. 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 (3:36)
Sign off (31:54)
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 Boeing 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 on the podcast.
Brad Bonavida: Welcome everybody back to the Nexus podcast. I'm Brad filling in for James today. I've got with me. Rosie, uh, from Nexus, as well as [00:01:00] James Coleman from Princeton. Um, so why don't we start, uh, we're gonna start with at the Nexus and explain some things that are going on at Nexus. But first, James, if you don't mind giving a quick introduction of who you are and how you found yourself here.
James Coleman: Um, first of all, thank you very much for having me on your, uh, podcast. I really appreciate the opportunity to be here and chat with you on this fine Halloween, uh, day. Um, as Brad said, I'm James Coleman. I, I work at Princeton University where I have a dual role. I am both a PhD candidate where I do research around smart buildings and ai and all of its sort of requisite technologies.
And I also work within our facilities. Team as a BAS analyst, um, helping run special projects on our campus.
Brad Bonavida: Right. And James, you, uh, won the NEXIA award for, um, up and coming leader in the industry. Uh, so congrats on that. Did you get your award? 'cause I mailed it to you 'cause you left right before the award ceremony.
Did it come.
James Coleman: Yes it did. Thank you so much for mailing that. And apologies, the flights to Philly are [00:02:00] infrequent, so thank you for mailing in.
Brad Bonavida: Yeah, no worries. Well, congrats. Um, yeah, that was kind of a, a democratic decision by a lot of people. So a lot of people think that you're gonna do a lot more good things in this industry, which is cool.
James Coleman: That's very kind. Thank you for the honor. I appreciate it.
Brad Bonavida: Yeah, of course. Um, I should probably mention in case there's anybody watching this on YouTube, that it is Halloween, so that is why Rosie, are you a princess?
Rosy Khalife: So Queen I. Thanks Brad. That's very kind of you to, I love that you just adjusted it. So my son is the dyno from Super Mario, and so then I'm like the princess in Super Mario.
That was what I was going for. Um, but you know, I had to keep it work appropriate, so we're wearing a sweater, but the crown is like, I wasn't gonna come in a whole princess costume. That'd be a lot for everyone. So that's me. And then Brad, are you, what are, what's, what would we name you if we had to name? I don't
Brad Bonavida: know.
This actually isn't what I'm gonna like tonight. Oh. Uh, my wife and I are gonna be, uh, [00:03:00] vendors. So we have like suits that like we have, uh. Like a little case that's suspenders that we have popcorn in that we can hand out to people like the vendor at like IT baseball game. You know that you can buy stuff.
Rosy Khalife: That is amazing.
Brad Bonavida: Yeah. But right now did Ellis do that? Yeah, I didn't, she did that. I was very, now I'm a hamburger hat. It's just a hat I have, it's a hamburger, so I love it. It's great. Okay. Let
Rosy Khalife: you look kind of like a je Is it a jester? What's that thing? The, the, the Joker kind of guy? Is that what it's called?
Yeah. Yeah. That's what James
Brad Bonavida: and I were just talking about. It's very gesture esque.
Rosy Khalife: Yeah, it's great.
Brad Bonavida: We'll kick it off here with at the Nexus, which is a couple things that are happening in the Nexus Labs community, um, and post Nexus Con. The big thing that we've been focusing on and talking about is the launch of Nexus Cast, uh, rose, do you want to tell people what Nexus Cast is?
Rosy Khalife: Of course. Yeah. We're super excited about it. It's really came out of us trying to find a way to bring the nexus con experience [00:04:00] into the virtual world. So it, nexus con happens once a year, but what about all the other months of the year? How do we kind of connect and stay in touch with each other? And so, um, in December, we're doing one on the state of connected ot.
So we'll be focusing on cybersecurity and networking. That's happening on December 10th. Um, that'll be a half day. So really easy to engage in it. You know, you'll take it from the comfort of your home or office, but you'll be able to connect with folks from the industry. We'll be bringing building owners together with vendors.
There'll be demos, there'll be deep dives and sessions and breakouts. Like it's a really cool kind of way of bringing the nexus con experience. Into this space that we are all in and a a lot easier for us to do, rather than, you know, having to travel. Um, so we're excited about that. Just to give some logistics.
If you are a pro member, uh, it's free for you to join that. So that's a no brainer, right? For all of our pro members, there's hundreds of you. Hopefully we'll see a bunch of you, uh, registering for it. And then if you're not a [00:05:00] pro member, you can obviously become one and then just register for free or you can just register.
And tickets are 149 bucks. So. Super, uh, low commitment, and then if you're a building owner, it's free to attend. So we've had a bunch of people sign up, we're excited about it. Please more of you sign up so that we can understand kind of who's coming and gauge the, the, the interest, but also, you know, make sure the content fits the folks that are gonna be in the room.
Brad Bonavida: Yeah. And, and we've been kind of dialing in on that content from just the state of connected OT to, I think we're gonna try to wrap it specifically around device management as a topic. Um, so we're, we're, uh, we're playing around with the name, um. What is it? Oh, getting control of your building's OT devices.
So we just came out with a device management article a couple weeks ago that has six steps to device management. Um, and we're looking, we're gonna look to get building owners who can provide case studies of how their. [00:06:00]How they're approaching one of those six steps. So ideally we'll have six different case studies, each one from a different step of what it takes to do device management of operational technologies correctly in your buildings.
Um, so we will be reaching out to building owners that we think are, um, ahead of the curve here, but also this is a, uh, shout out that if you are a building owner and you are doing something with device management or have device management problems, um. Talk to us and maybe you can share where you're at and um, yeah, you could be one of the presenters at Nexus Con, so reach out.
Okay. With that said, uh, let's get into it with James, our guest here. Um, I already congratulated you on the award. It's a pretty big deal. We only give out, like, I think we gave out what, five awards and there's, you know, 400 plus people there. Um, but James, can you just start by talking about your experience at Nexus Con?
James Coleman: Well, I mean, I had a, a great time just like I did last year. I, I think, I mean there were a of very [00:07:00] engaging panels. I. Um, sort of like a lecture style content. This year was especially strong. Like there was some really good informational like sessions that, that proceeded panels, which I thought were really, really effective at framing conversations for an audience who may not be at the same place as the panelists.
I thought that was really good. Um, and then just like last year, I thought that the, frankly my favorite part was just seeing the whole community in one place. I mean, it's pretty, pretty exciting and fun. So thank you for organizing that guys.
Brad Bonavida: Yeah, of course. Um, so you were part of a, a session. This, your session was called FM Meets AI Case Studies in Improving Operational Efficiency.
Um, you had a couple of case studies that happened before your part, so I know there was a really good one with Willow and Northern Arizona University. Talking about how they were using AI and AI agents, um, to improve operations for their facility management. And then you were on a panel. Um, [00:08:00] so do you wanna just like, kind of describe what your panel was and how it went?
James Coleman: Yeah. So the panel that I was on was maybe a little leadingly framed as basically where does the hesitancy in AI adoption in FM come from and what can be done to overcome it. Um, and we had, uh, a pretty lively, spicy to use your, the sort of themes you guys were working on, um, conversation both beforehand in an email chain as well as on the actual, actual stage, um, about how yes, facilities, teams could be doing more to, um, make themselves like AI ready or sort of experiment more.
But also I think that the industry as a whole can do more. To sell FM teams on the actual value proposition of the AI tools that they're building. And so that was sort of the framing of the conversation we had.
Brad Bonavida: If this was forced into a debate style and you had to stand on one side of the room or the other, are you on the [00:09:00] side of the room?
That's like AI for operational efficiency. AI for FMS is a real thing today that can happen and be effective. Or are you on the side of. It's kind of a pipe dream and it's like gonna come in the future, but there's not a lot of good tools out there. Which side are you closer to?
James Coleman: I, I, I hate having to take a side to be quite honest.
Um, because I don't think it's quite that binary. So the way that I look at it is that I know for a fact that there are a ton of useful existing tools for FM teams that fall under a broad umbrella of. AI for facility management and operational efficiency. I know that they exist. Um, are they being, uh, you know, packaged and communicated the right way to the right people at the right times?
Right now that I'm not so sure about.
Brad Bonavida: Okay. You, you, you dodged my question, but I'll, I'll allow it. You did, I think. Yeah. Big time. No, it makes [00:10:00] sense. I mean, do you have any examples of like, is Princeton using AI for operational efficiency in any places that's notable?
James Coleman: I have an answer to this, but I would first like use, this is not a Dodge 'cause I have an answer, but do you Brad, want to.
Clarify what you mean by ai or would we like to do that maybe collectively? That's okay, that's the question. Good point. That's totally valid. Good, good
Brad Bonavida: point. That's level set. Yeah. So I actually, I, I, I have something to say on how I've come across that James, and I'd be happy to hear your opinion on it as well as Rosie.
So this is just the beginning, but I want both of you to add to it. Um, when we were making the agenda for Nexus Con. We have, we have an energy management track and a facility management track, and the world is obviously not that black and white. Those jobs often blend together. You're o often doing things for your facility that are going to help improve operations and save energy.
So it's, it's kind of conceptual, right? Um, so in the energy management track, we had an AI for energy management session, and then we're talking about, uh, AI for [00:11:00] facility management. What I found very interesting is that the AI for energy management session was. Pretty easy to get case studies for that would fit that and the AI for facility management session, it was kind of like pulling teeth to get real case studies from building owners of what's happening there.
And I think that that's really telling, because my, my theory is that when you say energy management, you can put an AI agent, which an agent in, in, in my definition is. Artificial intelligence that's doing a job without, with little or no human intervention. It's kind of working on its own. And in energy management, you can make an an AI tool.
Um. Do something like change the set point of a chiller or a rooftop unit, very specific, one output based on a ton of inputs, and that, uh, can work pretty effectively. But in facility management, we're talking about like improving operations, making equipment last longer, and that's very [00:12:00] abstract. Like how do you.
Tell an AI to make your chiller like last more years or work better, or is it? And also there's always gonna be a human in the loop. 'cause there's people actually turning wrenches and plugging in wires to make these things happen. So that's why I saw the hesitancy in this particular, uh, session. Yeah, I don't know if James, you wanna add to that or Rosie, um, on what you think about these two topics.
Rosy Khalife: I'll just add one thing and James, I'd love to hear your perspective on this too. It also, Brad, I think you hit on something really important. It's like the people doing these jobs that are fms, there's also a lot of like physical labor that they're putting in, that they're doing. And so, um, I could see if I was in a role like that, there is a reluctancy.
Of how AI can, can sort of collaborate with me given that I am like in a actual boiler room, right. Of a building and like that's where I work. And so it's like this abstract thought of like, wait, how is AI gonna now help me do my job? Like they're not here, I'm here. You know? [00:13:00] And so I'm sure there's some of that that's coming up.
That is nuanced around, you know, reluctancy to change and also a fear around like, can it actually help me? And like, will it do it the right the right way? Would it do it how I would do it? Is it actually saving me time or is it just adding to my workload like. Sort of like, step outta the way, I'll just handle it.
So I'm sure all those things are coming up for, for folks that are in these types of roles. Um, but yeah, James, I'd love to get your, your thoughts on that.
James Coleman: Yeah. Um, I, I think just sort of start at the beginning with how Brad pointed out there's a, there's a challenging, um, overlap between. Energy management and facility management, especially at an organization like Princeton where one group is responsible for both.
So when an alarm comes in that um, you know, a set point is not being hit in a room, is that a facility management issue? Is that a energy management issue? Are we over cooling? Is [00:14:00] that why? We're not hitting the set point, and that's therefore spending a bunch of energy we shouldn't be or something broken.
And so even at that triage level, it's sort of a little unclear, at least in an organization like Princeton. Um, whether we're looking at tools for, uh, an operational efficiency or facility management, I think they're often gonna be the same thing. Um, and to that point, like the original question was, um, are we doing anything currently which.
I would classify as falling under the, uh, AI umbrella in specifically like definitely a facility management, not an operational efficiency framing. And I think, um, I think that, um, anybody who's running A FDD, um, particularly the diagnostic part of that acronym is kind of definitionally running. Um, ai.
Right, right. It has to be. Mm-hmm. So, and like as an organization, we've been doing that for, I, I mean, way before I ever got here, [00:15:00] but I'm pretty sure almost 15 or 20 years, maybe longer. Mm-hmm. Somebody would've to correct me, but like a long time. So does that mean we've been doing AI for 20 years? Uh, maybe not by the current definition, but I think in a sort of academic.
Dictionary type sense. Absolutely. We have been. So that's, I mean, just one sort of narrow explanation of how I think Yes, we absolutely are. Are we using chatbots? No.
Brad Bonavida: Right. Yeah. No, I think that makes perfect sense. Um, and, and just back on the energy management spot, I don't want to belittle the, the work it takes on that side either.
So like shout out to, I think it was Faci AI and Wendy's who gave a case study as well as. Brainbox and Amazon and Microsoft. Like, that's a cool topic in its own, but how do you bring it into this segment of facility management? Like, yeah, you're right James. I think that FDD is like kind of, it, it's a gray area of what, what, where do you draw the line of what's AI and what's not?
Um, you did, I do recall something. I thought was really cool that you said on the panel, and I'm gonna butcher it. [00:16:00] So like, let me try to explain what I heard and, and you can tell me how you actually said it, but you were giving advice to the audience of like how to watch out for, um, like AI that's just pizazz and not actually real.
And you were talking about how. Um, like to avoid a vendor lock-in situation with AI where you're stuck with one vendor, you should be able to ask the vendor if their, their solution can be run on other agents. So like you could take their solution as I understood it, and run it on open AI or anthropic or whatever, I don't know, environment that you use for AI as a way to prove that they're really in it for the benefit of AI and not just trying to lock you into something.
Is that kind of what you were getting at or can you explain that again? Sure
James Coleman: thing. So, um, by the way, the, by the way that the process works is, uh, creates risk for lock-in, which is that one of the reasons that, um, deploying any [00:17:00] kind of machine learning or, or AI tool to a complicated building or campus, um, whether that's for FM purposes or energy efficiencies, the bootstrapping costs to get some new tool into your building are often quite high.
And often, like one of the things that has to happen, regardless of what kind of model we're talking about, is that you need, um, some kind of model of your building and, and I'm talking about like tagging or, um, or something a little bit more, uh, knowledge graph base. But you need, there's a base set of information, which when I look at a lot of the solutions that are on the market, require each and every vendor to implement them.
Before they can roll out their solution, which then means you've put a bunch of s sun costs into a particular, uh, company because it requires a specific kind of modeling to happen first. And that's obviously a, a, a big opportunity for lockin. And so what I may, what I may have said that on the panel is that one [00:18:00] of the things that I want to see more of going forward is, is, um, to make sure that we can, um, quickly evaluate.
Any kind of AI tool on a complicated building that the vendors are able to bootstrap or bolt on their solution into, um, a open standard data model, which I own as a building owner, not the vendor. Got
Brad Bonavida: it. Yeah, that makes sense. So it's more about, it's not so much the using the, um, somebody else's AI tool as it is using your data layer to.
Perform the tasks at hand.
James Coleman: Yeah, yeah, exactly. Yeah. To make sure that my switching costs, if it doesn't go well are as low as possible, so I get the best fit at the right time.
Brad Bonavida: Yeah. Okay. Aside from that, do you have other pieces of advice for facility management groups who are like maybe getting pressure to adopt AI and aren't sure where to go?
Like what? What would you recommend to a group [00:19:00] that's curious about this topic and maybe not very far ahead?
James Coleman: It's an excellent question and I think, forgive me for maybe being a little nihilistic, but a lot of the pitches that I hear are not owner or context specific people come mm-hmm. To a building owner, and this is, this predates me being at Princeton, but like back when I was a specifying engineer, I'd sit on like mid meetings and I hear pitches, which were about the product, not the problem.
Yes. And so, right. And so if I'm an FM team who's very busy, right? Like sitting in pitch meetings is not the highlight of my day. It doesn't get me home any faster, right? That's true. And so, um, one of the things that I think owners can do to make sure that they get the most outta some of these like increasingly requisite meetings to talk with vendors about ai, uh, is have.
The vendor is not likely to come to you with your problems. So [00:20:00] have those ready to go and when they start talking in generic terms, you flip the script and say, these are my specific problems. What tool do you have that can actually address my pain points?
Rosy Khalife: Yeah. Interesting. I think, James, you're hitting on something that often happens with a lot of, uh, companies that have a software or do a, you know, have a solution or services or whatever they're trying to sell something.
Yeah. And it's a total Rosie's rant, which James, you don't know. That is probably, but um. That is a funny thing where I notice things in our industry that are rant worthy and what you're hitting on is totally one of those things where they talk too much about their own product, right? 'cause that's what they know and that's what they're really passionate about.
But especially, I think it depends on who you're talking to. And so in this case, we're talking about fms, who are. Super busy. They need to be in different locations. They're getting pings on their phones, they're getting alerts, like there's just so much going on in their day to day, and for them to sit there [00:21:00] and hear about your awesome platform that you know, creates this amazing dashboard.
So on and so forth. Like it's just, it's not real enough for them. And so I think that's great for anyone that's listening that it maybe is trying to, to sell to fms, that is such an important point, especially for fms, where they really need you to hone in on what their problem is for and show up knowing that instead of showing up and asking questions, like you gotta know who you're talking to.
Brad Bonavida: Yeah,
James Coleman: completely agree. Well said.
Brad Bonavida: Yep. Yeah. Yep. I think we see that a lot, Rosie, in the form of like demos. Where, and again, you, you said like the person's very passionate about this product that they've spent a bunch of time on, but that, that come turns into a demo where maybe they've got a. A sidebar that's got 15 different features within that platform, and they're like clicking through it and they're like, we can also do alert, we can also do scheduling.
We can also do this. And it's like, I, we didn't ask you if you could do that. [00:22:00] Like you need to be addressing what problem or pain point that you're trying to solve. And it, and it's maybe even exacerbate worse with, uh, AI because they just want to go there and say that they've put effort into AI rather than showing how that's helping the pain point completely.
Yeah. Totally. Cool. Um, maybe if we were to zoom out a little bit, like we're talking about AI hesitancy, but James, you're in just a unique place because you've kind of been at the. Um, the in-between of like the academic side of building automation and the industry side. And I think that you've probably, before AI even came along, seen this like push and pull of like, here's this great, shiny, incredible thing.
But wait, why is no one actually adopting this and putting it in their buildings to make them better? So, I don't know. Kind of just open from there. Like what can you say about the academic versus the industry side and what the hesitancy of adoption has been for for years beyond just [00:23:00] AI agents specifically
James Coleman: because I made a kind of drastic life change four years ago when I left.
Being a Smart buildings consultant professionally in New York City to. Go work on a PhD at Princeton to try to think through that problem specifically. 'cause when I was working in New York, working with like some of the biggest building owners in the US with the folks with the most resources, I was often seeing that either.
They were unable to justify the tools that they wanted for financial reasons or for complexity reasons, or they that they couldn't buy things that I knew existed because I had been reading academic papers for a long time. But, um, but like hadn't made it to market in a scalable way. And so part of the reason I left my fairly normal, uh, engineering life was to try to get to the bottom of that question.
Um, and to, I have a frustration with each side of this challenge. Um. Starting with the sort of market side, the industry side, one of the things that frustrates me [00:24:00] right now about the current landscape is that when people talk about ai, and I, I shouldn't paint with a broad brush, but it's all I hear, so it's challenging not to, but like when I, when people say AI and I ask 'em about their tech stack, or I just look it up nine times outta 10, recently what they're talking about are wrappers around large language models.
Brad Bonavida: Hmm.
James Coleman: Which are excellent if you want a chat bot. But I do engineering problems both professionally and in an academic sense. Like those are like the wrong fit. It's a, it's a round peg in a square hole. It, it, that is like fundamentally they're sort of un uh, awkward bedfellows. Like they don't work well naturally together and.
What is frustrating when I hear these people, when I hear vendors, whomever talking about, oh, I've got this great new LLM based feature. I'm like, what about all of the prove like value proven machine [00:25:00] learning techniques that we have been developing in academia for like 20 years now? Like, why aren't you trying to commercialize those things, which I know save 20 or 30% operational energy?
Like, what, what? Can't we start with those and then you can gimme chat bots. Yeah. And so that's really frustrating on, on the one side. Yeah. And then to sort of quickly wrap up the other side of this, there is some objectively incredible, technically impressive research that comes out of national labs, universities in the States and elsewhere, um, showing, like transparently showing how to.
Uh, use sophisticated algorithms to save energy in a provable way. Like there's nothing, there's no better case study than an academic peer reviewed paper, right? There's no bias. It's just they, they found this information and they, they're presenting it to you and they're telling you how they did it. Um, I think that's like a, uh, wonderful, un largely untapped resource in industry.
And I think part of the challenge here is that, um, [00:26:00] academics, unfortunately, because doing. FM or energy efficiency research outside of a lab is expensive and money is hard to come by in academia, especially right now that they, um, often present their findings as if the problem has been solved against, for example, a simulation or a one-off building.
And they do not consider the sort of scalability challenges, which is what vendors have to figure out all day long. It's a hard problem. And so the sort of the bridge that I am personally hoping to. Um, build between the two sides here is help, um, vendors see the wealth of public domain inform, uh, uh, algorithms, methods that exist and been published, um, freely help them be solving real world problems with existing stuff that's already been published, and then help academics get a little more comfortable with the real world aspect of deploying things at scale.
Brad Bonavida: [00:27:00] Interesting. I think you're kind of, this is reminding me about the point that you were making of FDD kind of being AI for facility management that's been happening for decades. I think, uh, maybe some credit to Nate Bennis who was on the panel with you from University of Nebraska. I think he was the one who said like.
We've kind of got physics down more or less in, and like HVAC control or energy efficiency in your HVAC is physics. So does it need AI or is like the level of linear log regression, linear regressions and rules-based automations that have been built over decades, like good and it just needs to be implemented Well, and I think that's kind of what you're hitting at.
It's like we've, we've come up with a lot of stuff that works and it's maybe being. Uh, maybe AI is more of a distraction than anything else for this particular industry. I, I don't know, maybe that's a bold statement, but,
James Coleman: well, yeah, I, I don't, I may, maybe distraction is harsh, but, um, definitely, well, definitely it's stealing the oxygen from [00:28:00] a lot of like, sort of proven technologies that have real paybacks and, you know, coming from a specifying engineering world where nothing could happen unless you could prove the ROI and then.
So knowing that's sort of the state of affairs. Then some vendor comes to me and goes, Hey, I've got a chat bot. Like, do you have an ROI on that chat bot? No. Then how, what do you want me to do?
Brad Bonavida: Yeah,
James Coleman: right.
Brad Bonavida: Uh, exactly. Alright, um, so I want to transition back to Nexus cast just for a second because I know that you are a, an OT guy, a tech guy as well, who knows the networking side of this.
So. Anything that you would like to see in this is basically Nexus Conva virtual, but focused on that OT track. Like what would you like to see talked about or addressed within this event?
James Coleman: I think that for a lot of, especially like medium sized organizations right now, device management is increasingly becoming a topic that they have to figure out.
Um, and this is being driven, I [00:29:00] think, I'm sure you guys have talked about this somewhere, but like a lot of this is being driven by, um, increasing requirement for converged networks. And that's one thing we know about, uh, sort of traditional network managers, that they do not like random devices on a network with unknown firmware, uh, releases and doing bad stuff.
So, um, yeah, talking about what device management prac best practices exist out there, I think would be invaluable for the industry as a whole.
Brad Bonavida: Awesome. Okay, cool. Um, I, any, any last comments on this Rosie James, before we, before we wrap up and go into car carve outs?
Rosy Khalife: Well, I guess I have a question for James just to, for, for our listeners that don't know you as well as we know you, like, maybe you could share with us the other things that, that you're excited about right now, aside from, from some of the stuff you've already shared with us.
Because I know you're, you do a lot,
James Coleman: two things, which I think everybody should be following, and if they're not, this is my prompt for them to do that. So there's two big initiatives coming, sort [00:30:00] of, um, sponsored by Ashra right now, which are reaching maturity, and I think you're gonna hit. The industry in the coming months and maybe, you know, year to year and a half.
Um, the first one is, uh, the semantic interoperability modeling standard that Ashley's been working on for like almost five years. Uh. It might be six years now. It's just about finished, it's public review and that's gonna become, um, published. And I think, uh, this is a reaction to the, what I think you guys used to call the ontology wars back when that was more of a, a sort of raging topic.
Uh, so the official sort of ashray sponsored version of this is coming out and, um. Much in the way that Backnet was transformational for the industry, I think that this might be as well. So I think, I hope that people will take a look at that. Um, and then related to this, there's a, a, a second ashra sponsored initiative called, um, controlled description language, CDO, which is a kind of standard.[00:31:00]
Programming framework for building systems and something I've been experimenting with both of these in my lab, which is, you know, why I'm bringing them up, um, here. But that's another really potentially transformational technology for our industry. I heard somebody maybe in one of the most recent episodes of the podcast talking about how, uh, the ability to program controllers, maybe it was Terry Hur, in a standard language, would be a game changer.
'cause one of the main ways we get locked out of our own gear. When the programming tools are proprietary, and so CDL is actually's reaction to that challenge and I think it's a big pain point that their smart people are working hard to address and I hope people will check it out.
Brad Bonavida: Nice. That's great. We will, uh, get links for those, both of those items to put in the show notes for someone who's interested in learning more about either of them.
Okay, cool. Uh, this has been great. I think we'll wrap it with some carve outs here, um, which is just something that's fun and more personal that we've been going through. Uh, Rosie do you wanna start with your carve out?
Rosy Khalife: [00:32:00] Sure. Um, so I was thinking about what is my carve out? The thing that came to mind is, so I have all these nice kitchen knives that.
You, you mostly my husband gets me and they're all like really high quality Japanese knives and whatnot, but a knife is only as good as how sharp it is. And so you need to learn how to sharpen these knives or else they're sort of useless and they become dull. And so I am learning how to sharpen knives.
I have this whole like kit. Yes, and it's a whole thing, but it's really hard actually. And like you could mess up the knife if you do a shitty job. And it like the last, there's so many different, you know. Pieces of it, but the last step is like a really nice leather that you, that's like the final step of the process and you do it with water and it's like this whole thing anyways.
Um, so I have not, I have not become a master. It is actually harder than I thought, so. To that end, I'm sure someone listening probably does sharpen their knives or something and has a, an easier [00:33:00] method than this kit that I got. That is really, I think this kit is like what people use to start a business to sharpen knives.
Like I'm, I'm not gonna be that person at the farmer's market. So, um, I'm all ears if anyone has advice on what I should use instead, but yes, that's what's happening for me. That's great.
James Coleman: Awesome.
Rosy Khalife: James, what do you got? Thanks, Brad.
James Coleman: Um, well, I mean, just 'cause people may, may have seen my hand here. I, uh, very excited to be, um, potentially having a functioning wrist for the first time in six years after a, a skateboarding injury.
Oh, that took me out for a while. Uh, so that's exciting. Um, but then. Some, a book or I recently read, which like I've been thinking about constantly since I finished it, which I, I've heard other people do this bit on your show. So I recently finished, uh, the biography of Leanest Vols. He was a guy who wrote Linux.
Rosy Khalife: Okay. And he cool 'cause an
James Coleman: autobiography, which he's just like a funny guy. Like one, it's a cool story [00:34:00] about how, how the name of the book is called. Just for Fun, the story of an accidental revolutionary came out in 2001, found it in like a used bookstore by accident. And it is just such a faci. He's a fascinating guy in the story of Linux, which now the whole modern world depends on it and sort of, this one guy did it in a dorm room in his like, I think late twenties.
And it's just like a cool, he's a cool, cool, cool story, cool person. Uh, check it out. If you're interested in open source software, that's great. Or yeah, so tech innovation.
Brad Bonavida: I love it. We, we barely even scratched the surface of open source, which is something that you are well in the weeds about. So maybe next time, yeah, I do.
I did wanna ask though, the skateboarding accident, were you doing something cool or was it one of those ones where you were just hanging out that it happened? Like what, what's the story there?
James Coleman: Sure. Thank you. Uh, all right, well, so here's, this is the embarrassing one. This wrist, I broke this summer doing nothing.
Cool.
Rosy Khalife: Oh my goodness. I hit a crack at full speed into what is going [00:35:00] on. I, timing, skateboarding, timing also.
James Coleman: Oh yeah. Sort of, I mean, just going faster. Oh, you're like, you're like
Rosy Khalife: a pro skateboard, like you're,
James Coleman: well, no, if I was, I'd break less wrists,
Rosy Khalife: I guess. Yeah, that's fair. You like it though, but it's a hobby.
Of course. I do enjoy it. Yeah. Okay.
James Coleman: Yes, exactly. But this one, I, I, I mean, vaguely cool. I was in a, a bowl in Manhattan, um, on the, on the Hudson River. Uh, just locked up doing a trick that's not super hard, but fell quite far backwards and. Yeah. Snap is Dang. Yeah. Ouch. Yeah.
Brad Bonavida: Alright. My, since it's, uh, since it's Halloween, um, I, my wife and I love horror movies, so my two horror movie recommendations, which aren't like from the depths, like if you've seen movies, you've heard of these both.
But we watched, uh. The Silence of the Lambs like two nights ago, which is such a classic. It's so good. I if you like real horror movies, it's really good. Oh my goodness. It's really well done. I've
Rosy Khalife: watched it once and it's taunts timeless. It haunts me still to this day.
Brad Bonavida: Yeah, it's good. I
Rosy Khalife: can't, I can't do horror, so watch that.
If you And
Brad Bonavida: it, okay, so for the [00:36:00] lighter, if that's too dark for you. Beetlejuice one is so good. And then classic. They actually did a really good job with Beetlejuice two as well. Interesting. Like they could have flopped, I think. Um, but it's actually really good. Is it as good as the first? Probably not, but it's really close.
It's also very good. So. Interesting. Okay. With that, we will wrap it, but thank you James. We really appreciate you being on the pod and we'll see you at Nexus Cast. Thank you Rosie, and have a good Friday everybody.
Rosy Khalife: Thank you guys. Thanks guys. Thanks for listening in. Bye.
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.
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 188 is a conversation with Brad Bonavida and Rosy Khalife from Nexus Labs, as well as James Coleman from Princeton University.
Episode 188 is a conversation with Brad Bonavida and Rosy Khalife from Nexus Labs, as well as James Coleman from Princeton University. 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 (3:36)
Sign off (31:54)
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 Boeing 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 on the podcast.
Brad Bonavida: Welcome everybody back to the Nexus podcast. I'm Brad filling in for James today. I've got with me. Rosie, uh, from Nexus, as well as [00:01:00] James Coleman from Princeton. Um, so why don't we start, uh, we're gonna start with at the Nexus and explain some things that are going on at Nexus. But first, James, if you don't mind giving a quick introduction of who you are and how you found yourself here.
James Coleman: Um, first of all, thank you very much for having me on your, uh, podcast. I really appreciate the opportunity to be here and chat with you on this fine Halloween, uh, day. Um, as Brad said, I'm James Coleman. I, I work at Princeton University where I have a dual role. I am both a PhD candidate where I do research around smart buildings and ai and all of its sort of requisite technologies.
And I also work within our facilities. Team as a BAS analyst, um, helping run special projects on our campus.
Brad Bonavida: Right. And James, you, uh, won the NEXIA award for, um, up and coming leader in the industry. Uh, so congrats on that. Did you get your award? 'cause I mailed it to you 'cause you left right before the award ceremony.
Did it come.
James Coleman: Yes it did. Thank you so much for mailing that. And apologies, the flights to Philly are [00:02:00] infrequent, so thank you for mailing in.
Brad Bonavida: Yeah, no worries. Well, congrats. Um, yeah, that was kind of a, a democratic decision by a lot of people. So a lot of people think that you're gonna do a lot more good things in this industry, which is cool.
James Coleman: That's very kind. Thank you for the honor. I appreciate it.
Brad Bonavida: Yeah, of course. Um, I should probably mention in case there's anybody watching this on YouTube, that it is Halloween, so that is why Rosie, are you a princess?
Rosy Khalife: So Queen I. Thanks Brad. That's very kind of you to, I love that you just adjusted it. So my son is the dyno from Super Mario, and so then I'm like the princess in Super Mario.
That was what I was going for. Um, but you know, I had to keep it work appropriate, so we're wearing a sweater, but the crown is like, I wasn't gonna come in a whole princess costume. That'd be a lot for everyone. So that's me. And then Brad, are you, what are, what's, what would we name you if we had to name? I don't
Brad Bonavida: know.
This actually isn't what I'm gonna like tonight. Oh. Uh, my wife and I are gonna be, uh, [00:03:00] vendors. So we have like suits that like we have, uh. Like a little case that's suspenders that we have popcorn in that we can hand out to people like the vendor at like IT baseball game. You know that you can buy stuff.
Rosy Khalife: That is amazing.
Brad Bonavida: Yeah. But right now did Ellis do that? Yeah, I didn't, she did that. I was very, now I'm a hamburger hat. It's just a hat I have, it's a hamburger, so I love it. It's great. Okay. Let
Rosy Khalife: you look kind of like a je Is it a jester? What's that thing? The, the, the Joker kind of guy? Is that what it's called?
Yeah. Yeah. That's what James
Brad Bonavida: and I were just talking about. It's very gesture esque.
Rosy Khalife: Yeah, it's great.
Brad Bonavida: We'll kick it off here with at the Nexus, which is a couple things that are happening in the Nexus Labs community, um, and post Nexus Con. The big thing that we've been focusing on and talking about is the launch of Nexus Cast, uh, rose, do you want to tell people what Nexus Cast is?
Rosy Khalife: Of course. Yeah. We're super excited about it. It's really came out of us trying to find a way to bring the nexus con experience [00:04:00] into the virtual world. So it, nexus con happens once a year, but what about all the other months of the year? How do we kind of connect and stay in touch with each other? And so, um, in December, we're doing one on the state of connected ot.
So we'll be focusing on cybersecurity and networking. That's happening on December 10th. Um, that'll be a half day. So really easy to engage in it. You know, you'll take it from the comfort of your home or office, but you'll be able to connect with folks from the industry. We'll be bringing building owners together with vendors.
There'll be demos, there'll be deep dives and sessions and breakouts. Like it's a really cool kind of way of bringing the nexus con experience. Into this space that we are all in and a a lot easier for us to do, rather than, you know, having to travel. Um, so we're excited about that. Just to give some logistics.
If you are a pro member, uh, it's free for you to join that. So that's a no brainer, right? For all of our pro members, there's hundreds of you. Hopefully we'll see a bunch of you, uh, registering for it. And then if you're not a [00:05:00] pro member, you can obviously become one and then just register for free or you can just register.
And tickets are 149 bucks. So. Super, uh, low commitment, and then if you're a building owner, it's free to attend. So we've had a bunch of people sign up, we're excited about it. Please more of you sign up so that we can understand kind of who's coming and gauge the, the, the interest, but also, you know, make sure the content fits the folks that are gonna be in the room.
Brad Bonavida: Yeah. And, and we've been kind of dialing in on that content from just the state of connected OT to, I think we're gonna try to wrap it specifically around device management as a topic. Um, so we're, we're, uh, we're playing around with the name, um. What is it? Oh, getting control of your building's OT devices.
So we just came out with a device management article a couple weeks ago that has six steps to device management. Um, and we're looking, we're gonna look to get building owners who can provide case studies of how their. [00:06:00]How they're approaching one of those six steps. So ideally we'll have six different case studies, each one from a different step of what it takes to do device management of operational technologies correctly in your buildings.
Um, so we will be reaching out to building owners that we think are, um, ahead of the curve here, but also this is a, uh, shout out that if you are a building owner and you are doing something with device management or have device management problems, um. Talk to us and maybe you can share where you're at and um, yeah, you could be one of the presenters at Nexus Con, so reach out.
Okay. With that said, uh, let's get into it with James, our guest here. Um, I already congratulated you on the award. It's a pretty big deal. We only give out, like, I think we gave out what, five awards and there's, you know, 400 plus people there. Um, but James, can you just start by talking about your experience at Nexus Con?
James Coleman: Well, I mean, I had a, a great time just like I did last year. I, I think, I mean there were a of very [00:07:00] engaging panels. I. Um, sort of like a lecture style content. This year was especially strong. Like there was some really good informational like sessions that, that proceeded panels, which I thought were really, really effective at framing conversations for an audience who may not be at the same place as the panelists.
I thought that was really good. Um, and then just like last year, I thought that the, frankly my favorite part was just seeing the whole community in one place. I mean, it's pretty, pretty exciting and fun. So thank you for organizing that guys.
Brad Bonavida: Yeah, of course. Um, so you were part of a, a session. This, your session was called FM Meets AI Case Studies in Improving Operational Efficiency.
Um, you had a couple of case studies that happened before your part, so I know there was a really good one with Willow and Northern Arizona University. Talking about how they were using AI and AI agents, um, to improve operations for their facility management. And then you were on a panel. Um, [00:08:00] so do you wanna just like, kind of describe what your panel was and how it went?
James Coleman: Yeah. So the panel that I was on was maybe a little leadingly framed as basically where does the hesitancy in AI adoption in FM come from and what can be done to overcome it. Um, and we had, uh, a pretty lively, spicy to use your, the sort of themes you guys were working on, um, conversation both beforehand in an email chain as well as on the actual, actual stage, um, about how yes, facilities, teams could be doing more to, um, make themselves like AI ready or sort of experiment more.
But also I think that the industry as a whole can do more. To sell FM teams on the actual value proposition of the AI tools that they're building. And so that was sort of the framing of the conversation we had.
Brad Bonavida: If this was forced into a debate style and you had to stand on one side of the room or the other, are you on the [00:09:00] side of the room?
That's like AI for operational efficiency. AI for FMS is a real thing today that can happen and be effective. Or are you on the side of. It's kind of a pipe dream and it's like gonna come in the future, but there's not a lot of good tools out there. Which side are you closer to?
James Coleman: I, I, I hate having to take a side to be quite honest.
Um, because I don't think it's quite that binary. So the way that I look at it is that I know for a fact that there are a ton of useful existing tools for FM teams that fall under a broad umbrella of. AI for facility management and operational efficiency. I know that they exist. Um, are they being, uh, you know, packaged and communicated the right way to the right people at the right times?
Right now that I'm not so sure about.
Brad Bonavida: Okay. You, you, you dodged my question, but I'll, I'll allow it. You did, I think. Yeah. Big time. No, it makes [00:10:00] sense. I mean, do you have any examples of like, is Princeton using AI for operational efficiency in any places that's notable?
James Coleman: I have an answer to this, but I would first like use, this is not a Dodge 'cause I have an answer, but do you Brad, want to.
Clarify what you mean by ai or would we like to do that maybe collectively? That's okay, that's the question. Good point. That's totally valid. Good, good
Brad Bonavida: point. That's level set. Yeah. So I actually, I, I, I have something to say on how I've come across that James, and I'd be happy to hear your opinion on it as well as Rosie.
So this is just the beginning, but I want both of you to add to it. Um, when we were making the agenda for Nexus Con. We have, we have an energy management track and a facility management track, and the world is obviously not that black and white. Those jobs often blend together. You're o often doing things for your facility that are going to help improve operations and save energy.
So it's, it's kind of conceptual, right? Um, so in the energy management track, we had an AI for energy management session, and then we're talking about, uh, AI for [00:11:00] facility management. What I found very interesting is that the AI for energy management session was. Pretty easy to get case studies for that would fit that and the AI for facility management session, it was kind of like pulling teeth to get real case studies from building owners of what's happening there.
And I think that that's really telling, because my, my theory is that when you say energy management, you can put an AI agent, which an agent in, in, in my definition is. Artificial intelligence that's doing a job without, with little or no human intervention. It's kind of working on its own. And in energy management, you can make an an AI tool.
Um. Do something like change the set point of a chiller or a rooftop unit, very specific, one output based on a ton of inputs, and that, uh, can work pretty effectively. But in facility management, we're talking about like improving operations, making equipment last longer, and that's very [00:12:00] abstract. Like how do you.
Tell an AI to make your chiller like last more years or work better, or is it? And also there's always gonna be a human in the loop. 'cause there's people actually turning wrenches and plugging in wires to make these things happen. So that's why I saw the hesitancy in this particular, uh, session. Yeah, I don't know if James, you wanna add to that or Rosie, um, on what you think about these two topics.
Rosy Khalife: I'll just add one thing and James, I'd love to hear your perspective on this too. It also, Brad, I think you hit on something really important. It's like the people doing these jobs that are fms, there's also a lot of like physical labor that they're putting in, that they're doing. And so, um, I could see if I was in a role like that, there is a reluctancy.
Of how AI can, can sort of collaborate with me given that I am like in a actual boiler room, right. Of a building and like that's where I work. And so it's like this abstract thought of like, wait, how is AI gonna now help me do my job? Like they're not here, I'm here. You know? [00:13:00] And so I'm sure there's some of that that's coming up.
That is nuanced around, you know, reluctancy to change and also a fear around like, can it actually help me? And like, will it do it the right the right way? Would it do it how I would do it? Is it actually saving me time or is it just adding to my workload like. Sort of like, step outta the way, I'll just handle it.
So I'm sure all those things are coming up for, for folks that are in these types of roles. Um, but yeah, James, I'd love to get your, your thoughts on that.
James Coleman: Yeah. Um, I, I think just sort of start at the beginning with how Brad pointed out there's a, there's a challenging, um, overlap between. Energy management and facility management, especially at an organization like Princeton where one group is responsible for both.
So when an alarm comes in that um, you know, a set point is not being hit in a room, is that a facility management issue? Is that a energy management issue? Are we over cooling? Is [00:14:00] that why? We're not hitting the set point, and that's therefore spending a bunch of energy we shouldn't be or something broken.
And so even at that triage level, it's sort of a little unclear, at least in an organization like Princeton. Um, whether we're looking at tools for, uh, an operational efficiency or facility management, I think they're often gonna be the same thing. Um, and to that point, like the original question was, um, are we doing anything currently which.
I would classify as falling under the, uh, AI umbrella in specifically like definitely a facility management, not an operational efficiency framing. And I think, um, I think that, um, anybody who's running A FDD, um, particularly the diagnostic part of that acronym is kind of definitionally running. Um, ai.
Right, right. It has to be. Mm-hmm. So, and like as an organization, we've been doing that for, I, I mean, way before I ever got here, [00:15:00] but I'm pretty sure almost 15 or 20 years, maybe longer. Mm-hmm. Somebody would've to correct me, but like a long time. So does that mean we've been doing AI for 20 years? Uh, maybe not by the current definition, but I think in a sort of academic.
Dictionary type sense. Absolutely. We have been. So that's, I mean, just one sort of narrow explanation of how I think Yes, we absolutely are. Are we using chatbots? No.
Brad Bonavida: Right. Yeah. No, I think that makes perfect sense. Um, and, and just back on the energy management spot, I don't want to belittle the, the work it takes on that side either.
So like shout out to, I think it was Faci AI and Wendy's who gave a case study as well as. Brainbox and Amazon and Microsoft. Like, that's a cool topic in its own, but how do you bring it into this segment of facility management? Like, yeah, you're right James. I think that FDD is like kind of, it, it's a gray area of what, what, where do you draw the line of what's AI and what's not?
Um, you did, I do recall something. I thought was really cool that you said on the panel, and I'm gonna butcher it. [00:16:00] So like, let me try to explain what I heard and, and you can tell me how you actually said it, but you were giving advice to the audience of like how to watch out for, um, like AI that's just pizazz and not actually real.
And you were talking about how. Um, like to avoid a vendor lock-in situation with AI where you're stuck with one vendor, you should be able to ask the vendor if their, their solution can be run on other agents. So like you could take their solution as I understood it, and run it on open AI or anthropic or whatever, I don't know, environment that you use for AI as a way to prove that they're really in it for the benefit of AI and not just trying to lock you into something.
Is that kind of what you were getting at or can you explain that again? Sure
James Coleman: thing. So, um, by the way, the, by the way that the process works is, uh, creates risk for lock-in, which is that one of the reasons that, um, deploying any [00:17:00] kind of machine learning or, or AI tool to a complicated building or campus, um, whether that's for FM purposes or energy efficiencies, the bootstrapping costs to get some new tool into your building are often quite high.
And often, like one of the things that has to happen, regardless of what kind of model we're talking about, is that you need, um, some kind of model of your building and, and I'm talking about like tagging or, um, or something a little bit more, uh, knowledge graph base. But you need, there's a base set of information, which when I look at a lot of the solutions that are on the market, require each and every vendor to implement them.
Before they can roll out their solution, which then means you've put a bunch of s sun costs into a particular, uh, company because it requires a specific kind of modeling to happen first. And that's obviously a, a, a big opportunity for lockin. And so what I may, what I may have said that on the panel is that one [00:18:00] of the things that I want to see more of going forward is, is, um, to make sure that we can, um, quickly evaluate.
Any kind of AI tool on a complicated building that the vendors are able to bootstrap or bolt on their solution into, um, a open standard data model, which I own as a building owner, not the vendor. Got
Brad Bonavida: it. Yeah, that makes sense. So it's more about, it's not so much the using the, um, somebody else's AI tool as it is using your data layer to.
Perform the tasks at hand.
James Coleman: Yeah, yeah, exactly. Yeah. To make sure that my switching costs, if it doesn't go well are as low as possible, so I get the best fit at the right time.
Brad Bonavida: Yeah. Okay. Aside from that, do you have other pieces of advice for facility management groups who are like maybe getting pressure to adopt AI and aren't sure where to go?
Like what? What would you recommend to a group [00:19:00] that's curious about this topic and maybe not very far ahead?
James Coleman: It's an excellent question and I think, forgive me for maybe being a little nihilistic, but a lot of the pitches that I hear are not owner or context specific people come mm-hmm. To a building owner, and this is, this predates me being at Princeton, but like back when I was a specifying engineer, I'd sit on like mid meetings and I hear pitches, which were about the product, not the problem.
Yes. And so, right. And so if I'm an FM team who's very busy, right? Like sitting in pitch meetings is not the highlight of my day. It doesn't get me home any faster, right? That's true. And so, um, one of the things that I think owners can do to make sure that they get the most outta some of these like increasingly requisite meetings to talk with vendors about ai, uh, is have.
The vendor is not likely to come to you with your problems. So [00:20:00] have those ready to go and when they start talking in generic terms, you flip the script and say, these are my specific problems. What tool do you have that can actually address my pain points?
Rosy Khalife: Yeah. Interesting. I think, James, you're hitting on something that often happens with a lot of, uh, companies that have a software or do a, you know, have a solution or services or whatever they're trying to sell something.
Yeah. And it's a total Rosie's rant, which James, you don't know. That is probably, but um. That is a funny thing where I notice things in our industry that are rant worthy and what you're hitting on is totally one of those things where they talk too much about their own product, right? 'cause that's what they know and that's what they're really passionate about.
But especially, I think it depends on who you're talking to. And so in this case, we're talking about fms, who are. Super busy. They need to be in different locations. They're getting pings on their phones, they're getting alerts, like there's just so much going on in their day to day, and for them to sit there [00:21:00] and hear about your awesome platform that you know, creates this amazing dashboard.
So on and so forth. Like it's just, it's not real enough for them. And so I think that's great for anyone that's listening that it maybe is trying to, to sell to fms, that is such an important point, especially for fms, where they really need you to hone in on what their problem is for and show up knowing that instead of showing up and asking questions, like you gotta know who you're talking to.
Brad Bonavida: Yeah,
James Coleman: completely agree. Well said.
Brad Bonavida: Yep. Yeah. Yep. I think we see that a lot, Rosie, in the form of like demos. Where, and again, you, you said like the person's very passionate about this product that they've spent a bunch of time on, but that, that come turns into a demo where maybe they've got a. A sidebar that's got 15 different features within that platform, and they're like clicking through it and they're like, we can also do alert, we can also do scheduling.
We can also do this. And it's like, I, we didn't ask you if you could do that. [00:22:00] Like you need to be addressing what problem or pain point that you're trying to solve. And it, and it's maybe even exacerbate worse with, uh, AI because they just want to go there and say that they've put effort into AI rather than showing how that's helping the pain point completely.
Yeah. Totally. Cool. Um, maybe if we were to zoom out a little bit, like we're talking about AI hesitancy, but James, you're in just a unique place because you've kind of been at the. Um, the in-between of like the academic side of building automation and the industry side. And I think that you've probably, before AI even came along, seen this like push and pull of like, here's this great, shiny, incredible thing.
But wait, why is no one actually adopting this and putting it in their buildings to make them better? So, I don't know. Kind of just open from there. Like what can you say about the academic versus the industry side and what the hesitancy of adoption has been for for years beyond just [00:23:00] AI agents specifically
James Coleman: because I made a kind of drastic life change four years ago when I left.
Being a Smart buildings consultant professionally in New York City to. Go work on a PhD at Princeton to try to think through that problem specifically. 'cause when I was working in New York, working with like some of the biggest building owners in the US with the folks with the most resources, I was often seeing that either.
They were unable to justify the tools that they wanted for financial reasons or for complexity reasons, or they that they couldn't buy things that I knew existed because I had been reading academic papers for a long time. But, um, but like hadn't made it to market in a scalable way. And so part of the reason I left my fairly normal, uh, engineering life was to try to get to the bottom of that question.
Um, and to, I have a frustration with each side of this challenge. Um. Starting with the sort of market side, the industry side, one of the things that frustrates me [00:24:00] right now about the current landscape is that when people talk about ai, and I, I shouldn't paint with a broad brush, but it's all I hear, so it's challenging not to, but like when I, when people say AI and I ask 'em about their tech stack, or I just look it up nine times outta 10, recently what they're talking about are wrappers around large language models.
Brad Bonavida: Hmm.
James Coleman: Which are excellent if you want a chat bot. But I do engineering problems both professionally and in an academic sense. Like those are like the wrong fit. It's a, it's a round peg in a square hole. It, it, that is like fundamentally they're sort of un uh, awkward bedfellows. Like they don't work well naturally together and.
What is frustrating when I hear these people, when I hear vendors, whomever talking about, oh, I've got this great new LLM based feature. I'm like, what about all of the prove like value proven machine [00:25:00] learning techniques that we have been developing in academia for like 20 years now? Like, why aren't you trying to commercialize those things, which I know save 20 or 30% operational energy?
Like, what, what? Can't we start with those and then you can gimme chat bots. Yeah. And so that's really frustrating on, on the one side. Yeah. And then to sort of quickly wrap up the other side of this, there is some objectively incredible, technically impressive research that comes out of national labs, universities in the States and elsewhere, um, showing, like transparently showing how to.
Uh, use sophisticated algorithms to save energy in a provable way. Like there's nothing, there's no better case study than an academic peer reviewed paper, right? There's no bias. It's just they, they found this information and they, they're presenting it to you and they're telling you how they did it. Um, I think that's like a, uh, wonderful, un largely untapped resource in industry.
And I think part of the challenge here is that, um, [00:26:00] academics, unfortunately, because doing. FM or energy efficiency research outside of a lab is expensive and money is hard to come by in academia, especially right now that they, um, often present their findings as if the problem has been solved against, for example, a simulation or a one-off building.
And they do not consider the sort of scalability challenges, which is what vendors have to figure out all day long. It's a hard problem. And so the sort of the bridge that I am personally hoping to. Um, build between the two sides here is help, um, vendors see the wealth of public domain inform, uh, uh, algorithms, methods that exist and been published, um, freely help them be solving real world problems with existing stuff that's already been published, and then help academics get a little more comfortable with the real world aspect of deploying things at scale.
Brad Bonavida: [00:27:00] Interesting. I think you're kind of, this is reminding me about the point that you were making of FDD kind of being AI for facility management that's been happening for decades. I think, uh, maybe some credit to Nate Bennis who was on the panel with you from University of Nebraska. I think he was the one who said like.
We've kind of got physics down more or less in, and like HVAC control or energy efficiency in your HVAC is physics. So does it need AI or is like the level of linear log regression, linear regressions and rules-based automations that have been built over decades, like good and it just needs to be implemented Well, and I think that's kind of what you're hitting at.
It's like we've, we've come up with a lot of stuff that works and it's maybe being. Uh, maybe AI is more of a distraction than anything else for this particular industry. I, I don't know, maybe that's a bold statement, but,
James Coleman: well, yeah, I, I don't, I may, maybe distraction is harsh, but, um, definitely, well, definitely it's stealing the oxygen from [00:28:00] a lot of like, sort of proven technologies that have real paybacks and, you know, coming from a specifying engineering world where nothing could happen unless you could prove the ROI and then.
So knowing that's sort of the state of affairs. Then some vendor comes to me and goes, Hey, I've got a chat bot. Like, do you have an ROI on that chat bot? No. Then how, what do you want me to do?
Brad Bonavida: Yeah,
James Coleman: right.
Brad Bonavida: Uh, exactly. Alright, um, so I want to transition back to Nexus cast just for a second because I know that you are a, an OT guy, a tech guy as well, who knows the networking side of this.
So. Anything that you would like to see in this is basically Nexus Conva virtual, but focused on that OT track. Like what would you like to see talked about or addressed within this event?
James Coleman: I think that for a lot of, especially like medium sized organizations right now, device management is increasingly becoming a topic that they have to figure out.
Um, and this is being driven, I [00:29:00] think, I'm sure you guys have talked about this somewhere, but like a lot of this is being driven by, um, increasing requirement for converged networks. And that's one thing we know about, uh, sort of traditional network managers, that they do not like random devices on a network with unknown firmware, uh, releases and doing bad stuff.
So, um, yeah, talking about what device management prac best practices exist out there, I think would be invaluable for the industry as a whole.
Brad Bonavida: Awesome. Okay, cool. Um, I, any, any last comments on this Rosie James, before we, before we wrap up and go into car carve outs?
Rosy Khalife: Well, I guess I have a question for James just to, for, for our listeners that don't know you as well as we know you, like, maybe you could share with us the other things that, that you're excited about right now, aside from, from some of the stuff you've already shared with us.
Because I know you're, you do a lot,
James Coleman: two things, which I think everybody should be following, and if they're not, this is my prompt for them to do that. So there's two big initiatives coming, sort [00:30:00] of, um, sponsored by Ashra right now, which are reaching maturity, and I think you're gonna hit. The industry in the coming months and maybe, you know, year to year and a half.
Um, the first one is, uh, the semantic interoperability modeling standard that Ashley's been working on for like almost five years. Uh. It might be six years now. It's just about finished, it's public review and that's gonna become, um, published. And I think, uh, this is a reaction to the, what I think you guys used to call the ontology wars back when that was more of a, a sort of raging topic.
Uh, so the official sort of ashray sponsored version of this is coming out and, um. Much in the way that Backnet was transformational for the industry, I think that this might be as well. So I think, I hope that people will take a look at that. Um, and then related to this, there's a, a, a second ashra sponsored initiative called, um, controlled description language, CDO, which is a kind of standard.[00:31:00]
Programming framework for building systems and something I've been experimenting with both of these in my lab, which is, you know, why I'm bringing them up, um, here. But that's another really potentially transformational technology for our industry. I heard somebody maybe in one of the most recent episodes of the podcast talking about how, uh, the ability to program controllers, maybe it was Terry Hur, in a standard language, would be a game changer.
'cause one of the main ways we get locked out of our own gear. When the programming tools are proprietary, and so CDL is actually's reaction to that challenge and I think it's a big pain point that their smart people are working hard to address and I hope people will check it out.
Brad Bonavida: Nice. That's great. We will, uh, get links for those, both of those items to put in the show notes for someone who's interested in learning more about either of them.
Okay, cool. Uh, this has been great. I think we'll wrap it with some carve outs here, um, which is just something that's fun and more personal that we've been going through. Uh, Rosie do you wanna start with your carve out?
Rosy Khalife: [00:32:00] Sure. Um, so I was thinking about what is my carve out? The thing that came to mind is, so I have all these nice kitchen knives that.
You, you mostly my husband gets me and they're all like really high quality Japanese knives and whatnot, but a knife is only as good as how sharp it is. And so you need to learn how to sharpen these knives or else they're sort of useless and they become dull. And so I am learning how to sharpen knives.
I have this whole like kit. Yes, and it's a whole thing, but it's really hard actually. And like you could mess up the knife if you do a shitty job. And it like the last, there's so many different, you know. Pieces of it, but the last step is like a really nice leather that you, that's like the final step of the process and you do it with water and it's like this whole thing anyways.
Um, so I have not, I have not become a master. It is actually harder than I thought, so. To that end, I'm sure someone listening probably does sharpen their knives or something and has a, an easier [00:33:00] method than this kit that I got. That is really, I think this kit is like what people use to start a business to sharpen knives.
Like I'm, I'm not gonna be that person at the farmer's market. So, um, I'm all ears if anyone has advice on what I should use instead, but yes, that's what's happening for me. That's great.
James Coleman: Awesome.
Rosy Khalife: James, what do you got? Thanks, Brad.
James Coleman: Um, well, I mean, just 'cause people may, may have seen my hand here. I, uh, very excited to be, um, potentially having a functioning wrist for the first time in six years after a, a skateboarding injury.
Oh, that took me out for a while. Uh, so that's exciting. Um, but then. Some, a book or I recently read, which like I've been thinking about constantly since I finished it, which I, I've heard other people do this bit on your show. So I recently finished, uh, the biography of Leanest Vols. He was a guy who wrote Linux.
Rosy Khalife: Okay. And he cool 'cause an
James Coleman: autobiography, which he's just like a funny guy. Like one, it's a cool story [00:34:00] about how, how the name of the book is called. Just for Fun, the story of an accidental revolutionary came out in 2001, found it in like a used bookstore by accident. And it is just such a faci. He's a fascinating guy in the story of Linux, which now the whole modern world depends on it and sort of, this one guy did it in a dorm room in his like, I think late twenties.
And it's just like a cool, he's a cool, cool, cool story, cool person. Uh, check it out. If you're interested in open source software, that's great. Or yeah, so tech innovation.
Brad Bonavida: I love it. We, we barely even scratched the surface of open source, which is something that you are well in the weeds about. So maybe next time, yeah, I do.
I did wanna ask though, the skateboarding accident, were you doing something cool or was it one of those ones where you were just hanging out that it happened? Like what, what's the story there?
James Coleman: Sure. Thank you. Uh, all right, well, so here's, this is the embarrassing one. This wrist, I broke this summer doing nothing.
Cool.
Rosy Khalife: Oh my goodness. I hit a crack at full speed into what is going [00:35:00] on. I, timing, skateboarding, timing also.
James Coleman: Oh yeah. Sort of, I mean, just going faster. Oh, you're like, you're like
Rosy Khalife: a pro skateboard, like you're,
James Coleman: well, no, if I was, I'd break less wrists,
Rosy Khalife: I guess. Yeah, that's fair. You like it though, but it's a hobby.
Of course. I do enjoy it. Yeah. Okay.
James Coleman: Yes, exactly. But this one, I, I, I mean, vaguely cool. I was in a, a bowl in Manhattan, um, on the, on the Hudson River. Uh, just locked up doing a trick that's not super hard, but fell quite far backwards and. Yeah. Snap is Dang. Yeah. Ouch. Yeah.
Brad Bonavida: Alright. My, since it's, uh, since it's Halloween, um, I, my wife and I love horror movies, so my two horror movie recommendations, which aren't like from the depths, like if you've seen movies, you've heard of these both.
But we watched, uh. The Silence of the Lambs like two nights ago, which is such a classic. It's so good. I if you like real horror movies, it's really good. Oh my goodness. It's really well done. I've
Rosy Khalife: watched it once and it's taunts timeless. It haunts me still to this day.
Brad Bonavida: Yeah, it's good. I
Rosy Khalife: can't, I can't do horror, so watch that.
If you And
Brad Bonavida: it, okay, so for the [00:36:00] lighter, if that's too dark for you. Beetlejuice one is so good. And then classic. They actually did a really good job with Beetlejuice two as well. Interesting. Like they could have flopped, I think. Um, but it's actually really good. Is it as good as the first? Probably not, but it's really close.
It's also very good. So. Interesting. Okay. With that, we will wrap it, but thank you James. We really appreciate you being on the pod and we'll see you at Nexus Cast. Thank you Rosie, and have a good Friday everybody.
Rosy Khalife: Thank you guys. Thanks guys. Thanks for listening in. Bye.
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.
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.