Sam Ramadori, CEO of BrainBox AI, helps us explore the evolving landscape of AI and its practical applications, especially in energy management and sustainability in built environments.
Sam Ramadori, CEO of BrainBox AI, helps us explore the evolving landscape of AI and its practical applications, especially in energy management and sustainability in built environments. He says AI has the potential to significantly reduce energy waste.
BrainBox AI uses deep learning, cloud-based computing, algorithms and a proprietary process to support a 24/7 self-operating building that requires no human intervention and enables maximum energy efficiency.
---------
SAM: It's like driving your car: instead of going 50 miles an hour on the highway for several hours, which is the efficient thing to do, Imagine you drove that car and you go zero to 80, and then you bring it back down to 20, and then you go back to 72 and then you bring it back. If you did that, your gas consumption would be a disaster. Right. And that's what systems do today because they don't know what's coming.
SAM: Tt's exciting times, uh, to be in a firm like yours or ours because what used to be an impossible task to justify the investment to A CFO is now getting easier every week because they're seeing examples we had that connected to this, and we were able to collect that data in efficient manner, and so on and so forth.
SAM: That's where I see AI being able to help us leapfrog; it's going to not necessarily replace people as much as it's going to make us so much better at decision making by presenting us all of the relevant options in order to make a change.
SAM: Buildings account for a huge amount of emissions globally, period, right? Huge percentage. And then number two, there's studies out there—some of them might be a little bit dated—but on average, your average building wastes about 30% of energy. Right? Literally wastes 30%. Maybe things have gotten better in the last 10 years, let's say, and it's 20%. It's still a huge number. AI will allow us to go and tackle those problems so that buildings—we don’t have the average waste of 20%—we have the average waste of 2%.
---------
00:19 - In the News
12:54 - About Sam
19:42 - Today’s AI capabilities
39:04 - Tech integration
45:55 - Rapid Fire
---------
Find Richard on LinkedIn
More about Accruent
Eric: Hello and welcome to Beyond Built. I'm Eric Cook.
[00:00:02] Richard: I am Richard Leurig.
[00:00:03] Eric: And on today's episode, we'll be talking with Sam Ramadori, CEO of Brainbox AI, about, you guessed it, AI. what it is, what it isn't, and how to truly use it effectively today. Sam, it's great to have you here.
[00:00:16] Sam: Good to be here. Thank you.
[00:00:18] Eric: Alright, listeners and viewers, as we like to start every episode with our segment called In The News. Richard, our in the news segment today is all about, again, surprise, AI. There was a recent article in CNBC where they took some excerpts from from Bill Gates and his interview with Jimmy Fallon and I kind of wanna discuss that with you today.
[00:00:38] Richard: That sounds great.
[00:00:39] Eric: So the first thing I want to ask you,
[00:00:41] what Bill's saying, hyperbole, saying that people are going to be replaced by AI in the next 10 years, or do you think that that's really true?
[00:00:51] Richard: I actually think that many positions and roles will be replaced or will change in nature. I'm not sure [00:01:00] entirely that humans will be replaced in the next 10 years. I do think the 10 year horizon is a little longer than what I believe the pace of change will be.
[00:01:10] AI currently, and certainly with Agentic AI coming out and advancing over the next few years, you're gonna see a lot more autonomous interactions and autonomous work by AI and, and what it can do.
[00:01:23] I'm not sure it can entirely replace some of the thought leadership at this point, but the other side of it is there are certain positions in the world where physical interaction with an asset or a device is absolutely required, and to a certain degree, those aren't going away anytime soon in my belief.
[00:01:43] A car for example, can operate and tell you a lot more and try to self-repair and self-heal a building. And we'll be talking some about autonomous, I've talked about this before, and, and Sam is an expert in, in buildings and autonomous, autonomous work related [00:02:00] to assets that run buildings and factories and plants and things like that.
[00:02:03] But to a certain degree, somebody still has to go in and actually put the HVAC unit in the building and somebody has to turn a wrench at some point. And so the nature of some positions is going to change, but overall, I believe there will be certain positions, and this is where it becomes very, very difficult to look at it.
[00:02:23] There'll be certain positions where somebody's sitting at a desk. Doing a certain amount of work the AI system will actually just replace what that person is looking at or doing over time. And then that'll change fundamentally the nature later on of what a SaaS application or what data interacting doing that position, it's almost like the AI is gonna learn what it needs to do to do the job, and then it's gonna learn what kind of systems it needs, and it'll invent its own systems to do the job as long as it has access to the data.
[00:02:54] Eric: Dovetailing off of that then, what skills do you think are gonna be most valuable for professionals in our [00:03:00] industry and in people who manage real estate built environments, industrial estates, things like that, in this AI driven future.
[00:03:09] Richard: Well, I mean I still think architects we're gonna are gonna be augmented by ai, but they're still going to need to know and understand how to do architecture. Engineers are gonna need to understand how to do engineering. But there will be a lot less, there'll be a lot more presented to them and they will be doing more and more quality checks and reviews and, and kind of inputting what do they want to see out of the certain asset or what are they building?
[00:03:33] I think I used to say, being a technologist myself, that people need to know more about technology. We need more people in science and engineering positions, and I still think we need more people in science, engineering and positions. But for example, our software engineers are not gonna be sitting down, heads down coding so much in the future.
[00:03:54] The code will be generated for us and there'll be a lot more, a lot less interaction on [00:04:00] hands-on keyboard coding, something. And it'll be a lot more of def definition of what it is that we want built. At least that's what I believe. I'm sure Sam will give us, uh, his insights as well today.
[00:04:12] Eric: The last question I have for you, Richard, before we move on, is what ethical considerations do you think we should keep in mind as we implement AI, especially in built environments?
[00:04:21] Richard: Well, the problem with anything that's taken out the human factor, and by the way, the more you get into agentic AI and AI becoming more and more self realizing, you know, basically understanding more about how humans interact and operate. I think the problem is a direct AI scenario could actually, for example, you may, you may have a certain need to have certain vendors and certain suppliers doing things for you.
[00:04:48] And I'm thinking specifically of minority and women owned businesses and things like that, where an AI might just say, I'm gonna go with the cheapest. I'm gonna go with, you know, whatever I need to do. Without really thinking about the [00:05:00] human factors of how we're trying to, uh, level the playing field, for example, across, um, societal norms and societal things that we're trying to do.
[00:05:08] But I do believe over time AI will even learn those things that we're trying to do and we'll bring those to bear. The other thing is that, like any huge evolution or revolution of, of things like technology or the industrial age or things like that, there's going to be human impact.
[00:05:26] And so I think the ethical consideration, I don't know if it would be an ethical consideration or a human consideration, about how we treat our employees, how we treat people through this change, um, people still need to be employed. We're not in the era of people just sitting at home watching AI run the world.
[00:05:45] And so how we treat individuals, how we train individuals, educate them, basically prepare them for the next generation of where the world is going. I think that is, in a sense, an ethical consideration. I. It's important that we [00:06:00] understand that. And then the last thing is, um, and I've talked about AI making its own decisions,
[00:06:06] We all know certain forms of AI now hallucinate or try to answer questions that aren't actually, it doesn't know the answer to, and things like that. And that's a big concern when you get into areas that Accruent is heavily into mining, oil and gas chemical. You know, what you don't want is AI making a decision at a chemical plant that could actually cause a big issue.
[00:06:26] A big issue for the environment, a big issue for, the people that live around that plant and so on. And so again, I'm not sure that's really an ethical dilemma as much as it is. Something that we need to watch out for and ensure that we still have quality controls around. And essentially the same things you build around humans, which is the box that they live in and the guardrails that they operate a chemical factory, for example, under, you have to give those same guardrails to ai, in my opinion. So.
[00:06:56] Eric: When I, and I think that you may end up seeing people whose [00:07:00] entire job roles is about helping ais define what those guardrails are. Because you know, you can tell a human, don't dump this in that river. But you can, if you told an AI that they may say, okay, well I'm gonna go dump it in the other river then, right?
[00:07:14] You may have to create a different set of principles where a human would go, okay, don't dump anything in any river. Got it. So I think there's some of that too. So Sam, I wanna bring in you into this as well, what I wanted to ask you is how do you see this integration of AI into our daily lives impacting, energy management and sustainability as we move forward?
[00:07:37] Sam: Great to hear the discussion. I think these types of discussions are becoming more and more relevant every week that goes by, right? The speed at which they're releasing more and more capable large language models, uh, only invites, the need for, to have these types of discussions.
[00:07:51] I think I remember hearing, uh, someone describe, you know, major technology adoptions over the last, call it even 200 years. And there's always that, [00:08:00] you know, we're not, as humans, we're not really good at predicting something that is such a disruptive change. So the question about, okay, what happens when AI starts autonomously taking over capabilities and how does that relate to the human's role and will it be replacing that? I mean, what's happening in the built environment? And the question around energy efficiency is a great prime example. So you start with fact number one. Buildings account for a huge amount of, of emissions globally, period, right? Huge percentage. And then number two. There's studies out there. Some of them might be a little bit dated, but on average, your average building wastes about 30% of energy. Right? Literally wastes 30%. Maybe things have gotten better in the last 10 years, let's say, and it's 20%. It's still a huge number.
[00:08:49] Okay, good. Not good, bad, but what do you do about it? Right? And like the problem with the built environment is we're talking about how many millions of buildings in the US alone, right? In [00:09:00] North America. Uh, that's number one. Each building is unique. The situations within the buildings are unique. How they're used is unique.
[00:09:06] There's a reason for that waste. It's not like we, you know, people don't like to throw money out into the air at the thus far. I don't know many people like that. And the problem is, is that it's uneconomical to go and service a rebuilding out there from a big office tower. Fine, you might invest some money there.
[00:09:23] You might have good teams on the ground, but your mom and pop store out in the middle of suburbia, in the middle of, uh, the Midwest. Uh, you're gonna send a technician there every day, right? Like, it's just so that that waste naturally builds up just 'cause of the way things are. And as of today. We can only spend that much money to send people in a van to go take care of these buildings. And they can only go so often. They can only spend so much time there. Like the economics of it all makes it such that it's limited and therefore 20, 30% of, of on average of energy is wasted in buildings. To me, when we started talking about what AI could [00:10:00] do, it's like, it's like it improves the Pareto as existing systems, the way we respond to them, allows you to hit, I don't know, the top 70% of problems. Well, to me, the, the AI and the capability it brings only means that we can go down, past 70, get to 80, get to 90, maybe we'll let go of the last 3% of energy waste. But AI will allow us to go and tackle those problems so that buildings, we don't have the average waste of 20%.
[00:10:25] We have the average waste of 2%. So that's the way I, I kind of view the reach of what AI could do. Richard alluded to it when he talked about AI coding, right? Like right now you wanna build a piece of software there better be a good market for it. But as coding becomes much more rapid, much more efficient, it only means you'll be developing more and more software for those edge use cases, making those edge use cases more efficient and you keep going down the Parato curve.
[00:10:51] That's kind of the way I view the world.
[00:10:52] Eric: I think that's a great insight, and I think, you know, based on this whole article and it's precedent, it's [00:11:00] precedence about ai taking over a lot of the things that we don't even think about it will take over. But I like the iterative approach that you're talking about and how we're it's going to help us get better.
[00:11:11] There's a, a newsletter that I read every day called AI for Good, and it talks about the breakthroughs that AI is helping make in, in medical in energy management, in all sorts of humanitarian areas as well. It's an area that, that I'm really interested in. So let's, um, pivot a little bit, 'cause this interview is about you and I want to introduce you to to everyone.
[00:11:33] So first of all, welcome to the show. I want to get into, uh, your background a little bit. So could you share with us a little bit about your journey and your role and the history of Brainbox ai?
[00:11:44] Sam: I think it's a bit different. Richard, you defined yourself as a technologist. Was that the right term? Was that the term you used?
[00:11:51] Richard: Uh,
[00:11:51] Sam: I think
[00:11:52] Richard: Yeah, I guess so.
[00:11:53] Sam: Sorry, I, I, I, I forgot you just said it a few minutes ago. My background's actually pretty different. I, uh, [00:12:00] started many moons ago as a lawyer and decided, that wasn't for me.
[00:12:03] Quickly shifted to more of the business side and, and, mainly was in private equity investing for the bulk of my career. And that allows you to touch many different industries, are are the firms I worked with. The typical investments we worked on were around traditional industries. So manufacturing, uh, you know, materials, uh, business to business services.
[00:12:27] That experience led me to a moment in time, crazy decision you have to cover with your partner at home. When you go home and say, I saw something really crazy and I'm gonna leave my day job and go. And go, uh, uh, help support this adventure.
[00:12:42] And that's, that's what, uh, led me to, to Brainbox. So our CTO today, Jean Mo, who came up with the idea behind Brainbox, which is how can we use autonomous AI in the, uh, heating and cooling systems of a building to make them just step change more efficient in a [00:13:00] step change scalable kind of way. And for me, the moment of that, uh, that aha moment for me when I met Jean and the small team that was starting to work on it was, you know, I had been around many types of manufacturing facilities, complex, large steel plants, uh, cement casting, all those kind of things. I kind of knew internally how they operate, like you alluded to before, the, the, the person running a chemical plant, right. You don't give a date to person on the job, the responsibility to run a chemical facility.
[00:13:31] Right. You usually, it's experienced 25 year, uh, person that's in that role. And they're doing the best they can. And I used to remember when we, you know, we were thinking about a project to optimize one of these large facilities with the large, large complex equipment. We used to bring together a, a team of 10 people, a mixed engineers operating experience.
[00:13:51] We'd go to the CFO, get a, I don't know, several million dollar budget, and you're gonna interrupt the operations of it, like it's painful, long, expensive, all to go [00:14:00] save. 2% of energy, 3% of energy. And when I met, Simone uh, and, and the, the small team working on it and, and what they were seeing and then just initially touching was, yeah, we're just gonna be able to put a software overlay, an AI platform, which should be more specific onto existing operating systems and buildings.
[00:14:20] And we probably think we'll be able to get, I don't know, 10, 15, 20% of energy savings. And that's where, you know, for me, I was like, poof. You know, my mind exploded. I ran home, told my wife, my wife, my partner, I said, uh, we're, I'm gonna quit my day job and I'm gonna go do this. And so that's, that was the start of my journey on, on AI with a whole lot of learning.
[00:14:41] This goes back now, I guess, almost six years and a whole lot of learning. And of course the timing of my, of that decision coincided with, ai just, you know, back then it was deep learning models. The evolution then two, three years ago, gen AI. But I, I basically, in my mind, I grew up with the industry, as a young [00:15:00] child when I jumped in because that's the level of, of, uh, of understanding I was at back then.
[00:15:04] And this is where we're, we're seeing more and more of the challenges, bridging that AI capability with what's really happening in the field. Uh, and that's where, and I was able to contribute to the team and, and help, you know, drive the growth of, of Brave Box.
[00:15:17] So today, a much larger company, and part of the trained technologies family as of recently.
[00:15:23] Richard: What does Brainbox focus on right now?
[00:15:27] Sam: We have like, our, our core product that we started working on seven, eight years ago was really to apply autonomous artificial intelligence into, uh, buildings, heating and cooling systems. Most people will recognize the application of autonomous AI in the self-driving car. It's probably the most well known example.
[00:15:49] And autonomous AI is still a pretty rare, rarely applied technology. A very rarely applied technology in the industrial world, for sure. Like we, we struggle to [00:16:00] find, um, kinda many examples. And so it was a deep, technological lift to, to, to get it there and make it work. But that was the core premise is can we take a building management system today that is well programmed, uh, but it's inside the building.
[00:16:17] It's reading sensors from inside the building and making decisions, but while it's making decisions about how to heat and cool rooms and keep it comfortable, it knows nothing about the weather outside. The cost of the energy it's using, is the energy using kind of high emissions or not? Uh, what's the occupancy in the building, et cetera, right?
[00:16:35] Like it's, it's making these decisions devoid of any of this information. And this is where we felt autonomous AI could come in, just create far more tele, a situation with more data, far more intelligent decision making. And then, in the case of a self-driving car, everyone understands what you're trying to do, which is let's remove the driver from the equation and still make that car get you from point A to point B.
[00:16:59] In the case of a [00:17:00] building where autonomous AI is really impactful, is a building is, if you break it down, I dunno, a 30 story building probably has 400, 500, 600 pieces of HVAC equipment. So even if you did have more information, how, on what basis, how many people in a room do you need to put to be able to be making decisions on a grant scale where you're tweaking a lot, a little kind of VAV box, uh, pump, fan motor, all over the building in real time.
[00:17:25] And that's where, in our case, autonomous AI really created a capability we never had before.
[00:17:32] Richard: Are you finding when you deploy brainbox into a facility, you know, we've talked to, to customers who have, for example, deployed carbon monoxide sensors into conference rooms. And I've learned that if you keep people in a very small space for too long of a period of time, lo and behold, there's a lot of carbon monoxide and the team starts to underperform over a number of hours without a break, right?
[00:17:56] I love what you guys are doing, but within that, have you had [00:18:00] some learnings that are really outside of what you expected to learn when you went in to start to look at all of those great things you're doing, which is, what's the weather outside? What's the right level of temperature? How do we make the building management system perform better?
[00:18:15] Was there kind of an unintended find or get that you, you had in there?
[00:18:20] Sam: I think what it, what it is the big leap that applying AI in that type of situation. So I take the buildings data, I marry it with the external data, what, what do I get? And it's really not like a list of learnings. It's really a new capability that I could not have achieved otherwise. And that in our case, is basically when the AI is done learning a building and it learns that individual building.
[00:18:43] 'Cause as you know, full well, as soon as you go through the building, across the street, it's completely different. And so you can't, you can't just take an AI that's learn something from one building, move it to the next completely different situation. But when it's done learning that building, what it's able to do is forecast the thermal behavior in each [00:19:00] HVAC zone slash room over, we use the next six hours with over 95% accuracy. And that's a big deal. Like right there, you've changed everything. 'Cause today's systems are a thermostat on the wall that only knows what's happening now, feeding the central computer and making decisions. That thermostat doesn't know what's gonna happen in five minutes, let alone six hours.
[00:19:21] So now you've given it an almost perfect view of the future, and you could just imagine the way it has opened up, optimize the optimization options 'cause, you know, when that thermostat says it's too hot, now I'm sitting here, I'm not gonna wait four hours for you to cool down the room. Right? I'm sorry, I've paid, I'm paying good rent money here.
[00:19:39] I'd like it cool, like pretty soon. So what's the reaction of a system in that case? It, it has no choice but to aggressively cool that room. Turn on the 18 pieces of equipment that are as fast as possible, gonna cool down that room and right away, kinda like, you know, it's like driving your car and then what you do instead of going a 50 miles an hour on the highway for several hours, which is the [00:20:00] efficient thing to do. Imagine you drove that car and you go zero to 80, and then you bring it back down to 20, and then you go back to 72 and then you bring it back. If you did that, your gas consumption would be a disaster. Right. And that's, but that's what systems do today because they don't know what's coming.
[00:20:15] So, you know, so thinking about knowing the future suddenly opens up a huge, like an almost endless amount of different decision making than you could have made before. Roof's gonna be hot, three hours, I could start flowing air now doesn't have to be as cold, et cetera. So like, you could put together all kinds of combinations to get to a better outcome.
[00:20:35] And that's the real kind of outcome of using AI in a, in a system like that.
[00:20:41] Richard: Yeah, that's great.
[00:20:41] Sam: Yeah.
[00:20:42] Eric: I know we probably wanna talk more about AI in general, but I do have one more specific question about, about the sort of applications that you guys are using today. And it has more to do with how does it figure out what people are doing. We have some conference rooms that get used maybe [00:21:00] very regularly for one week, and then they don't get used at all for the, for two weeks, and then they get used a lot again. So is it picking up patterns of what people are doing as well and using other things like scheduling information or how is it may help making those decisions?
[00:21:14] Sam: The core of the, of that prediction I just mentioned comes with the core set of information. The building information is always the same coming out of the building, but on, on the outside world, you're marrying that those external factors I mentioned. And that already gives you a very good prediction, right?
[00:21:31] Everything's about, are you at 98% prediction? If we were at 60% prediction, probably not enough to be making those, those autonomous decisions a little bit. You're on the edge of making things uncomfortable. And we didn't expect it either, right? When we were first working on it, the premise was, ah, if we crossed 80% forecasting accuracy, I think we can segue it. And, and as usual, AI surprises you and it gets to 88 and then it gets to 92, and then it gets to 95, and you're like, okay, we're more than fine. That's with the factors, [00:22:00] less what you just mentioned, Eric. Okay. We don't need to know where everybody is.
[00:22:04] We, in the building or what they're doing. We still have a very accurate forecast that we go get savings. Of course, like any other kind of data play, more data the better, generally speaking. And I'm gonna jump ahead here, Eric, some of the work your team does around the solutions it provides has exactly incredibly attractive data if available.
[00:22:24] Now, everything around occupancy, I think as we know out in the real world, there's generally low. Data available. It's only just coming with products like you offer. In a retail setting, however, quite different. A lot of retailers have some form of people counting, shopper counting mechanism, or you can even use Google Popular times, which gives you an idea of, you know, busyness in the store.
[00:22:48] That's already a good, a good step up. It would be our dream for all office buildings to have some form of, at least I, I, I don't want any personal information, I don't wanna know who it is, but how many people are in an area that would be [00:23:00] our, our dream. But it would only, it would, I don't wanna say only, it would only add to those big, chunky savings that are gotten from the beginning, because the system today, as I mentioned, is in that current state where it, it kind of knows nothing other than what the thermostat is saying.
[00:23:14] If I can simplify it that way.
[00:23:16] Eric: That's why I wanted to ask the question. That's actually amazing.
[00:23:18] Sam: Yeah.
[00:23:19] Eric: So, in general, if we step back from specific applications of AI, um, what have you real seen around the evolution of AI and what has really changed for us as AI comes to the more public consciousness, rather than just us assuming it's a general thing, or people assuming it's just generative ai.
[00:23:41] 'Cause that's where a lot of people still stick, that AI is generative AI. What do you think is going on in the, in the evolution of the way that we view things in AI today?
[00:23:50] Sam: It's interesting question 'cause because brainbox, if I take our seven, eight years, it exactly had, not in the middle, but in the fifth year, let's say it had the, the advent of that [00:24:00] famous November where chat GPT came to the awareness of the planet and then it went like wildfire. You can imagine we had five years before that where.
[00:24:10] I, I don't know that, I don't wanna say we were out in the wild or in the darkness, but when you went up to the, to a customer, especially in traditional industries, many probably that you work with and, and certainly real estate, you know, it's not the most forward, uh, in terms of technology adoption. I guess if you're a telco and you're not at the edge of technology adoption, your business is in danger.
[00:24:30] But that's not the case in real estate, right? And so you're talking to a customer about these novel deep learning models and what they do and what what you can't do with them. It was challenging discussion. There was a tremendous amount of education and a tremendous amount of doubt at the beginning about what it could do.
[00:24:46] And, and is it real? That was the deep learning model period, reinforcement learning, and then our autonomous type application. They were working on the self-driving car back then and, and so on and so forth. Then came generative [00:25:00] AI. Which in essence is a different approach of AI and different outcome, which became more human.
[00:25:07] When we run an autonomous optimization on your building, the dream is that you don't even notice a thing. Okay. Like we had no dashboard at the beginning. We had a horrible dashboard for a while because that wasn't the core of the product. That wasn't the magic. Generative AI changed all that. It put it in not only folks like us that work with it every day, but it put it in everybody's hands, including our children.
[00:25:31] Right. And suddenly, oh my, like, wow, that, that wow moment. So that's, I agree with you, Eric. Everyone views AI through that lens because that's where the general public touched it in, in a real way. For folks like us that, uh, work with it, more, uh, intimately than for us. Generative AI created a new capability and a new purpose.
[00:25:50] So while our core autonomous system is meant to optimize the background, yes, there will be interactions with the building operators, but the real results are happening in the background. [00:26:00] Generative AI allowed us to, to deliver real value to the operator in its in, in his or her day-to-day work. And we went from our autonomous solution to, what we call a virtual engineer for your building, and the way we kind of colloquially refer to it as, if Iron Man has Jarvis, as, as his helper, uh, with Brainbox, you have Aria, that's the name we gave it. And, and basically you have that, that real live, platform, AI generative AI platform that you can query for data.
[00:26:35] That you can ask for reasoning on why something is going wrong, that you can ask it, well, if that's what's wrong, what parts do I need to fix it? What you know, et cetera, et cetera. Like a richness of information that your poor, traditional building operator doesn't have at their fingertip. Like they could waste two, three hours looking for something.
[00:26:53] Whereas with a generative AI tool that's plugged in and has all the data we have, you're answering questions in a matter of five minutes or eight [00:27:00] minutes, that would've taken you two hours without it. So for us, it's been that kind of evolution. So when AI evolves and there's something really useful, we tend not to do AI just 'cause it's queued and modern and popular. It has to have a real outcome, and that's why we went from autonomous AI to then developing a virtual engineer using LLMs and generative AI because the need is dire, right? It goes back to the first comment we asked, like, do I see AI replacing? Well in our world, there's so many problems out in the field to go solve that. You know, people need every tool they can get to go tackle those without waste.
[00:27:31] Eric: That's interesting 'cause I know that Richard, you talk about this quite a bit, or at least we have, we have in a lot of our interactions about how we are going to see the next step of large language models and generative AI help us interact with the system. So, very much as Sam was talking about why did that happen?
[00:27:50] How do we prevent it from happening again? How do I give you the right information so that you can make better decisions in human language, rather than just using data? Because a lot [00:28:00] of things have to do with the way people feel about an issue or, or how they are experiencing it.
[00:28:06] Richard: Yeah. A store manager or a person in a building that's a site manager, a facility manager in a larger building in a high rise, they don't like logging onto screens anymore. I. And you talked, Sam about the evolution of technology over time. You know, one of the big changes was the advent in 2007 of an iPhone or, you know, an Android.
[00:28:27] And a lot of, at the time the term was the consumerization of IT. And really what that translated to is the expectation that the systems you're using at work. Or even the websites you're using at home start to have a certain look and feel and a certain interaction and usability and user experience that you've come to know and expect.
[00:28:47] Prior to that, people were used to, you know, whatever they were used to from, from what they were being delivered. The data, like a block screen of something and all of a sudden now you're used to getting it and it looks right on your mobile [00:29:00] device and it interacts the way you want to interact and it has fewer clicks and fewer things that you need to do, and you're seeing data in a different way.
[00:29:09] I think AI is now taking that to the next level, which is, I don't wanna real, really interact with that device anymore in the same way, I don't want to see screens in the same way. I want to ask questions and get answers, and I want to say things, and I'm not talking like the traditional people always jump to the traditional bots that answer your phone calls that don't know how to respond when you say, no, I want to change my reservation. I will send you to, you know, cancellations. No, I wanna change, it's the age old joke, right? AI is taking that to the next level though, so that a store manager can say, I'm seeing a leak in my store, and the AI knows it's raining outside. It's not raining outside. It's, it might be a a, depending on where they're currently standing, or what picture they took, a leak in the roof of a pipe that is a known [00:30:00] pipe in the engineering drawing specifications. And it might call out a critical person to come fix the pipe. Whereas before somebody was calling back to a network operation center or interacting, the store manager was interacting if they wanted to because they were like, it's too tough to interact with this application.
[00:30:17] So they'd call back to their network operation center. The operation center would then open a work order on a system. Somebody eventually would reach them out to them. You might roll a truck inadvertently because really there's nothing wrong. Or you might roll the truck too slow when there's a real major problem in the roof.
[00:30:34] I think what we're getting to now is the point where people want to interact and just say, the store manager, let's say in this case, the leak is still here. When's somebody gonna really be here? And they don't want to go into a system and look at a time or whatever.
[00:30:48] They want the system to respond and they go, I need faster response. So those feelings that you're talking about, Eric, around how, how people are interacting with it, the expectation is changing [00:31:00] dramatically. As to how you interact and get responses from, from these type of, these type of things.
[00:31:05] And I don't, I don't know if that, that that has to have played in Sam to your business and how you've, when you've deployed.
[00:31:11] Sam: You're describing it. And you know, the light, the light bulbs are going off and it's, it's exactly it and it's quite powerful because, you know, at the end of the day, I guess what we viewed when we were developing the core technology in our industry, there's very limited use of data, I will say.
[00:31:26] Right? There's no big real estate firm that's gonna proactively go and create a data platform in the cloud, pulling data from all their systems in a, like, they just don't do it. Like, unless there's a clear ROI on the spot, they don't do it. What we felt our system did was deliver that ROI to bring all kinds of data to the cloud, right?
[00:31:46] Stuff that you wouldn't do, just like that. And then what that ends up creating Richard, in, in the situation you just described, is that once it's somehow paid for the first time, it's amazing how much value just [00:32:00] keeps getting spun out of it. And, and everything you just described obviously needs a connectivity to the building and some measurement system and the, I don't know, maybe the water meter and, and and, etc. We are finally in an era where there's value to do so. Not all industries have jumped on it at the same speed. Not all, but it's becoming in, um, inescapable, right? Like people are seeing other industries do it and see it. And so Richard, it's exciting times, uh, to be, you know, in a firm like yours or ours because what used to be an impossible task to justify the investment to A CFO is now getting easier every week because they're seeing examples of, oh, if we had that connected to this, and we were able to collect that data in efficient manner, and so on and so forth. Yeah, of course we could do A, B, C, and D. So I think we're entering, as you say, a whole new world with what this thing is delivering to us, if it can put it that way.
[00:32:50] Eric: Well, and I'll even connect that back to the, the article and why I was a little incredulous about, Bill Gates' assumption that, so many things were gonna be taken over by AI. 'Cause at the [00:33:00] end of the day, we do need humans to temper the, the information, but AI is gonna help us make leaps as you said, that we can't even today fathom.
[00:33:09] And it already has. I mean, the amount of time that I spend researching on any topic now is, is negligible. Because of AI because I can just set AI now, especially, you know, now that, um, there's, there's, uh, reasoning engines, uh, built into most of the, major large language models. I use perplexity a great deal and I can point Perplexity. I can ask one simple question and it will bring me back pages and pages of relevant information because it can infer from my question what my context was, what I was trying to achieve. And that's where I see AI being able to help us leapfrog is it's going to not necessarily replace people as much as it's going to make us so much better at decision making by presenting us all of the relevant options in order [00:34:00] to make a change.
[00:34:01] Richard: Well, think about the transformation that went through when Google Search and, and Yahoo and Bing search and and so on. Have uh, have, have developed. How many people spend money making sure they're on the first page? How many ads are bought? But when you get to a world, and I know they're all incorporating ai, but when you get to a world where it's an interaction where you're asking something. And it's going and searching and giving you back a more specific response or a summary of responses. It's gonna change a lot of what we do. You know, when we say, are you gonna go Google it? I mean, at some point that's gonna have a whole new connotation. It may or may not be through Google, but it's gonna have a whole new connotation, which is like, are you interacting and getting back the response you want? In the time that you expect it without having to go through pages and pages of information to try to figure out whether the, the picture you were asking for, the image you were asking for, something you were asking for is the right image or picture. I'm, I'm [00:35:00] curious, Sam, in that kind of evolutionary world in which, in which you're working, I can really see the, the application of what you're doing in a, let's say, a retail store inside of a mall.
[00:35:10] I can see it inside of a high rise building or corporate complex. What came to mind is, the large footprint areas the mall itself, uh, arenas that are, that are cooled in the middle of the summer here in Texas, so that the Dallas Mavericks can come play when it's 110 degrees out and they're trying to cool the American Airline Center down.
[00:35:33] These things apply really globally. I mean, there are a lot of areas where I think, while it's not day in and day out and people in conference rooms, what you're doing could actually apply in some of these larger footprint areas where, I'm guessing there is a lot of energy waste. A lot of sustainability issues. Have you thought through that or seen that, come to fruition or is that something you're thinking about?
[00:35:58] Sam: The classic right now given moment in [00:36:00] time is data centers, right? I mean, they're going up like crazy. So it's, it's interesting the way, the way you would split the 70 years of development is it took us three or four years to just get that core autonomous capability to work. Five years, some some would say. And then, then after that, the doors opened up because it's like, it's that core capability that was incredibly hard. Then fashioning that or, or shaping that capability around a, a hockey arena like you said, or I don't know, an airport or something. It, it, you're then now starting to play with the algorithms and what objectives you give them and what extra data that I presume that hockey arena must have all kinds of sensors with regards to the equipment making the ice and all that kind of stuff. Well, okay, let's let right away, let's plug into that. That would be the instinct. So you're getting to those more tactical because the core technology has been tested, tried, improved, so on and so forth. So yeah, we are in that area where there's too many opportunities for us to focus on, which is a, which is a good, nice [00:37:00] place to be.
[00:37:00] Richard: Great thing for you, right? Yeah.
[00:37:02] Sam: Exactly. Yeah.
[00:37:04] Eric: Let's move on to another question that you sort of foreshadowed a bit earlier, Sam, which is, how do you see AI integrating with other technologies moving forward? Because I, I see a world where we maybe have AIs that talk to other AIs and they, they are making autonomous decisions between them. So what sort of things are you seeing happening in the future?
[00:37:26] Sam: That also, uh, Eric, I would define as in, in our history, in our short history, the last maybe year and a half, two years has, I mean, three years ago, four years ago, we had an API to nothing. Well, I'm exaggerating. 'cause we were, we were pulling external information like weather information Yes.
[00:37:45] Through an API. But we had. Uh, that's just data sources, but to other operating systems, we had API to nothing. And the last two years has been an explosion. Again, too many to tackle at the same time. But, you have a solid CMMS platform. Well, our [00:38:00] two systems should be talking to each other, right?
[00:38:01] Like you have work order management that would benefit in a huge way from the data about how exactly that equipment was run or was running and what faults and what degradation of performance is seen, not just from day-to-day data collection. So there's, there's things like that, that have a natural marriage.
[00:38:17] But we're definitely in the zone where people are acting, asking us, especially with generative AI where you go to the customer and say, okay, this is what Aria, this virtual engineer, can do. And then immediately, but hold on, we have a database of, I don't know, spare parts, can you please connect? So that way Aria knows that when there's a problem right away, it knows if I have the spare part? Or whether we need to rush order it?
[00:38:39] Like now, those are, those are coming up constantly, those requests. Um, so we're in that era. Richard, you alluded to it before. We're in that area. We're just value multiplies if you're able to pull in. And so we are on a drive to make more and more connections to, uh, other folks, whether it's their AI, although that yet [00:39:00] hasn't happened yet. Uh, but certainly there are platforms and the data each of us are collecting as we're delivering our services. The lines, I think between the service offerings are gonna start blurring up.
[00:39:12] Richard: Yeah. Yeah, I agree with that.
[00:39:14] Eric: And I think as you pointed out, Richard, the, the interfaces and things like that aren't going to be as important then, because I was thinking about, Sam you said about how it took you guys a long time before you even had a dashboard and then it took, it was a while before it was a good dashboard.
[00:39:26] Right. But doesn't a dashboard become fairly irrelevant if, if you just have, if you just have something telling you what the actual data is and what the actual thing that you need to do is.
[00:39:37] Sam: Yeah, yeah. Do I say Rich here, Richard, I see you nodding your head as well. Must be living the same experience, right?
[00:39:43] Richard: Well, and I think what will really happen in the future, so right now it's already a different user interaction and usability. Things are having to evolve. I think it used to be Sam, probably when you started that, people said, I can't see it. Show me something cool I can look at. I think now that's changing.
[00:39:59] I [00:40:00] think as it evolves, it'll also be that the AI, Eric will start to decide, if I were a dashboard, what would a dashboard do? And it's gonna start to show you the data that it thinks you want to see. That a dashboard would show you. So there will be no development of a dashboard. It'll be, if I was a dashboard, what would a dashboard show you that would be relevant to, to what you're looking for?
[00:40:23] And that's when we kind of take it to the next level. I mean, that's when we start to say, is Sam really here? Or how do we know it's not his large language model talking to us right now. Right. And is it there a little dot on the top right end of the of the screen you know, that'll tell us that? Right? Because I think what'll happen is you'll, you'll get to a point where what we believe are our products today, like A-C-M-M-S product or a leasing product, will start to be manifested by AI to say, if you're a leasing administrator on a building, what would you want to see if you were administering a lease?
[00:40:56] You'd wanna see lease subtraction, you'd wanna see the terms of the lease. You'd [00:41:00] wanna see certain things, and it'll start just rather than a, a software engineer trying to develop it with ai. It'll actually be developing it in real time. To fit your needs as Eric Cook. And that's where I believe that's gonna happen a lot faster than people believe.
[00:41:18] I'll end with this. People keep telling me it's 10 years out and I don't know. I'd love to hear your thoughts on this, Sam. People think it's 10 years out. I think it's three to five years out. It's, it's close. It's closer than people realize.
[00:41:30] Sam: Think about just the last year of generative AI. Your context window upon which you used to be able to put in questions was quite limited before. Now we're getting to million token context. You can literally ask a question that's framed within 18 paragraphs now and lobbying, right?
[00:41:46] You just not not be able to do that. Then the memory about the user. That's something that's evolving quickly as well. So when Sam logs on today for an hour and I close and I go back on tomorrow, today, it doesn't know what happened yesterday, it's gone. You know what [00:42:00] my conversation that's changing rapidly, all it takes is those two or three things and then, you alluded to it before, Richard. It's starting to code itself or starting to create itself. So you see it happening, this, this AI coding is happening right in front of our eyes in these months. So that dashboard that you were alluding to, you know who was gonna create it, how? Well it's in the background.
[00:42:22] It's now being put, it's coming up within seconds, make a request, or it figures out, like you say, every day Sam logs on and he asks the same four questions. So. Let's not waste our time and let's just feed it to him as soon as he logs in. Right. They were not far away from that at all.
[00:42:35] Eric: We're also honing in on more personalization where those roles are concerned as well. I'll give you a great example. There's been a couple of products out for the last few years that have been able to build your personal digital twin, right? Which will watch a video of you and you record like a five minute conversation.
[00:42:52] It gives you a script, you record it, and now it can use your tone of voice, it can move your mouth, right? It can approximate an avatar of you, [00:43:00] but what it wasn't able to do until recently was it wasn't able to also get your personality out of that because it was just reading the words and learning how to pronounce them.
[00:43:09] But now they're even going further. I mean a lot of people have noticed recently that Chat GPT, and it's always done this a little bit, but have noticed that Chat GPT, you're in an entirely new conversation and it remembers something you said from another conversation. 'Cause as you said, those tokenization are getting wider and wider and wider.
[00:43:27] And I think that's a great thing because especially when we think about how people interact with these systems, if the systems can learn to interact in the way the people want them to as well, then adoption goes through the roof so much faster.
[00:43:41] Sam: Yeah, I agreed.
[00:43:44] Eric: Thanks for that conversation, Sam, that was really interesting. We're gonna move on to our rapid fire segment where I'm gonna ask you a few just fun questions just to get your, uh, take on some things and hope you're, uh, ready for them now. So,
[00:43:56] Sam: Let's go.
[00:43:57] Eric: First of all, what is your favorite innovation and [00:44:00] why?
[00:44:01] Sam: I think more recently I'm quite excited and it is not our field about AI being used for things like new materials innovation, drug discovery, new chemicals. To Richard's point about how long it's gonna take to see, uh, outcomes there, I think we're gonna see some very interesting outcomes in terms of crazy new materials and more, far more rapid kind of drug discovery. I'm pretty excited about that.
[00:44:25] Eric: I love that. I love that thing. Things that we as humans couldn't even conceive of.
[00:44:29] Sam: Mm-hmm. Mm-hmm.
[00:44:30] Eric: So on the opposite side of that coin, what is one innovation that you just wish would go away?
[00:44:38] Sam: Look, I'm not a, I'm not a big fan of, uh, already social media was, had challenging elements to it. We all use it. Uh, has challenging elements to it. Unfortunately, I think the challenging elements combined with AI only supercharged those challenges or is supercharging those challenges. So if we could find a way to slow that down, that, that would be [00:45:00] appreciated.
[00:45:01] Eric: I, I, I understand that. I think that humans may be the limiting factor there.
[00:45:07] Sam: On that one. Yep.
[00:45:09] Eric: What is one thing that organizations are not doing today that you think they should be doing?
[00:45:14]
[00:45:14] Eric: I.
[00:45:15] I mean, in general, uh, as I mentioned before, plenty of industries that are not the fastest adopters of technology. If I was within a business, I would remove the ROI criteria of all the work around taking your data, and making it useful. And that means standardization. That means one data platform that means properly set up data, whatever, all that whole platform.
[00:45:39] Sam: If I was a company, I would be investing it with even a zero ROI justification immediately. That's a bit of a controversial statement, but you know, as for everything we talked about, world's changing really fast. Don't be left behind because your data and your whole structure is inaccessible.
[00:45:55] Eric: I think that's amazing. Richard, he's hired in our sales team, right? Right, right. Job offers out the [00:46:00] door. What is one thing that we wouldn't know about you by looking at your LinkedIn profile?
[00:46:06] Sam: Well, anything personal really, I am, uh. a big foodie. Long, long before the term foodie, you know, was, uh, was created, I don't know how many years back, but, uh, I grew up in the restaurant business and, uh, we are, as a family, fanatical about food.
[00:46:22] Eric: Okay, so I've gotta ask a follow up. I don't usually ask follow ups in this section, but what's the best meal you've ever eaten?
[00:46:27] Sam: Uh, well, we're Italian by background, so it is gonna be somewhere in Italy, uh, undoubtedly. And it's gonna be some nondescript little restaurant in the middle of nowhere town that is on no list of food ratings or anything like that. I mean, I could think of several dozen meals like that that were just spectacular.
[00:46:45] Eric: Now I've gotta turn it around. Richard, what's the best meal that you ever ate?
[00:46:49] Richard: Any New Mexican food. I'm from Albuquerque, all the little hole in the walls in Albuquerque and around New Mexico in the smallest of smallest towns. [00:47:00] It's beautiful. New Mexican food is unique. I suggest you go to Albuquerque or Santa Fe and have some.
[00:47:05] Eric: Alright, the next time I'm in the US I'm gonna make a side trip. Gonna have to, you're gonna have to gimme some names of some places.
[00:47:11] Richard: And I'm gonna get money from New Mexico for this advertisement as well.
[00:47:15] Eric: Excellent, excellent. Thank you for, thank you to the Board of Tourism. So Sam, thanks again for your time today. It's been re a real pleasure to have you. It's been a, such an interesting conversation. We're just so grateful for you, for you today.
[00:47:28] Sam: No, it, it's, it's great discussion
[00:47:29] Eric: so Richard, how did you feel about our conversation with Sam today?
[00:47:35] Richard: It was fascinating. It's a topic I love. I love understanding more about it. Sam brought a lot to the table. What his company is doing is incredible and, uh, incredible, not only for his customers, but incredible for society generally. So, I just found this great.
[00:47:50] Eric: Listening to Sam, there were several things that I picked out that I'm going to adopt as part of my, my ethos and my introspection as I look at [00:48:00] how we apply ai. 'Cause I think that just like everybody. Else, I get caught up in the actual tools that are available, rather than the possibilities that exist in a wider application.
[00:48:11] Do you think that our application of AI as not just as a software company, but as you know, humans and as we move forward, is going to, uh, cause some bumpiness in the future? Or do you think that we are going to find better ways of adopting than we have in previous technological revolutions?
[00:48:31] Richard: I think it's interesting because I think. We always talk about the rough times through these, these movements in society, these movements in technology or advances in things. Humans are amazingly resilient. They find ways to adapt, to change and to overcome those, those obstacles. So there will be bumps in the road.
[00:48:52] There will be barriers that are, have to be overcome. But I, I have a lot of faith that what will come in the end is [00:49:00] actually a much better place that we're in. But let's make no d no bones about it. There will be some bumps here in the road.
[00:49:08] Eric: Yeah, absolutely. And I think that, you know, as a final thought on this as we think about especially energy management and sustainability, I love the perspective of Brain Box, that there are two sides to the coin that can both be exploited at the same time, right? Not just about making people more comfortable or providing better environments for people, but it can also be about money at the same time.
[00:49:31] And I think that's where quick decisions get made too, which is really great. And I don't know what your thoughts on that are?
[00:49:39] Richard: I think it's the blending of most things that occur, we tend to look at them and say, we're gonna save energy. We're gonna have better sustainability. There must be some pain felt. And I actually think the one thing that AI really does, and he described it very well, is I can run a building more efficiently using less energy, but actually [00:50:00] make that meeting that you're getting ready to go into more comfortable. And do both at the same time.
[00:50:04] And I think that's the thing that people need to, to realize is that we can actually have it both ways. We can improve our lives, but actually have certain things like AI and, and other technologies helping us with sustainability, energy management, cost effectiveness, productivity improvements. I think it's a matter of just making sure when we build them, that we build them with both of those things in mind.
[00:50:29] I.
[00:50:30] Eric: Thanks again to Sam and to you Richard, for your, uh, participation today. I know that people are gonna find this conversation fascinating and I can't wait to see the conversations that we start around this. So thanks again.
[00:50:41] Sam: Thank you for having me.