How Trane Technologies Uses AI to Make Buildings More Autonomous

0:00We have a very simple approach to using AI. We call it the three-pillar approach. The first pillar is AI for growth. The second pillar is AI for employee experience. And the third pillar is AI for business transformation.

0:17Peter High: Welcome to Technovation. I'm Peter High, president of Metastrategy. My guest today is Riaz Rayhan. Riaz is the Senior Vice President and Chief Digital Officer of Train Technologies, a company that is a climate innovator that generated more than$21 billion in revenue in 2025. Through its Train and Thermo King brands, the company provides heating, ventilation, and air conditioning systems, building controls, and transport refrigeration solutions designed to improve energy efficiency and reduce emissions across buildings, homes, and transportation. As Chief Digital Officer, Riaz leads the company's digital strategy and works to deepen its technology, data, artificial intelligence, and connected product capabilities, helping customers improve energy and asset performance, lower operating costs, and reduce greenhouse gas emissions.

1:02Peter High: Riaz joined Train Technologies in 2023 after holding senior leadership roles spanning digital transformation, AI, the Internet of Things, product development, and general management. In this conversation, we'll discuss responsible AI, the relationship between AI and energy consumption, the potential for intelligent buildings to support a more resilient electrical grid, and how trained technologies is using brain box AI to make HVAC systems more autonomous and efficient. Riaz, welcome to Technovation. It's great to speak with you today. It's a real pleasure to be here. Today's episode is brought to you by Tines 3B.

1:36Peter High: Organizations are pushing teams to move faster with AI. That's creating a wave of AI-generated code and unsanctioned automations, leading to rising costs, security risks, and huge IT burdens. Meet Tines 3B, a single, secure environment that enables employees to build with AI while giving IT and security complete control to run and govern that work. 3B enables effortless building. Describe what you need in plain English, and 3B builds a fully functioning workflow. It uses IT-authorized connectors, meaning credentials stay protected, and non-technical teams can innovate without taking on unseen risk.

2:09Peter High: Also, with full visibility across every workflow, IT gets total governance, while features like Auto-Tune and Auto-Fix ensure workflows improve and self-heal continuously over time. Employees get the freedom to build, while IT gets the control to govern. No shadow AI, no data leaks. Tines 3B moves IT and security from blockers to enablers, giving teams the speed to innovate without compromising control. Visit the Tines website at tines.com, that's T-I-N-E-S.com, to learn more and try 3B for yourself. And now on to the interview. It's my pleasure, certainly. Well, maybe take it from where I left off, if you would, Riaz, and talk a bit about train technologies in your own words, if you would.

2:49Yeah, sure, Peter. Train Technologies is a 130-year-old company, and we've been around a long time. And we started off in the HVAC business and refrigeration for both buildings as well as transportation. And over the years, we have developed a broad array of services. And over the past couple of decades, we've developed very strong competency in controls and digital services. And through that process, we've made great innovations in sustainability, in driving services that help reduce carbon in buildings and in transportation. And I have the privilege of leading our digital business at Green Technologies.

3:32Peter High: We'll talk about that role if you would, Riaz. No two chief digital officer roles are quite alike. I'd love to understand how it applies in your case. Sure. So we're a big company. As you said, we do well over$20 billion in revenue. We have over 45 ,000 colleagues. We do manufacturing in over 40 locations. About half of those are here in the United States and the other half are in different parts of the world. So we manufacture in region for region. And by the same token, we have a big investment in digital as well. My role as Chief Digital Officer is to build products and services that help our customers drive value.

4:13So we have a sizable digital business where we actually sell these products and services. Drainbox is a good example of that, but there's more to it. There's enterprise asset management, there's telematics, and there's a whole slew of services that we offer through DrainCloud. And we actually bring this to market. So we are a software company as well as being an HVAC manufacturer.

4:40Peter High: I wanted to ask you, Riaz, connected HVAC systems can generate highly detailed information about buildings and their occupants. And I'd love to understand how you use that data. There must be a treasure trove of it. And I'd love to understand some of the basics of how you draw that down and derive insights from it. Yeah, it's a great question, Peter. And I'd like you to think about this at three levels. The most basic level is in ensuring that that equipment is well-maintained. We call this connected mechanical service. And through that, we are tracking a number of key variables and metrics about how the machine is operating and trying to use that to predict when the machine might fail and if it does, what the failure might look like.

5:26Obviously, the goal here is to prevent that failure before it happens. but equally to make sure that the machine is well-maintained and we do a truck roll when a truck roll is required. So that's, I would call that kind of a level one, right? But if you take that to the next level, at the next level, we're trying to understand a lot more about the energy consumption of the machine. We're trying to optimize how that machine might consume energy, how that machine might give us analytics that could even help us design better machines in the future or use that as input to our engineering organization.

5:59And I'd like you to think about that as level two. But if you dig into the next level, which is level three, think of that as autonomous. So this is the autopilot. At this point, think about what we do as autopilot for HVAC systems that are large, complex, commercial HVAC systems that are running data centers, airports, hospitals, large buildings, large office towers, and the like. And this is when we put in an autonomous AI agent that actually looks at internal data. It looks at external data. It looks at structured and unstructured data. So it's looking at lots of different data feeds, occupancy levels, the equipment in the building, the shape of the building, the sun position, particulate matter in the building around the atmosphere around the building.

6:46It's also looking at weather forecasts. And it's crunching all this in a machine learning model and making predictions that allow the HVAC system to make tiny changes, just like an autopilot, to make sure that we are optimizing every kilowatt of energy being used to heat or cool that building. And through the process, without sacrificing comfort, you can save 15 % to 25 % of the energy spent on the building. So think of that as level three. So those are the three levels that we use the data for.

7:22Peter High: And talk now about the way in which you apply artificial intelligence to what you've described. I know you've been very thoughtful in terms of thinking about the evolution of artificial intelligence. You've talked about 2024 being a land grab for AI, 2025 a period of reassessment, 26 a year to become more thoughtful. Describe what being thoughtful in 2026 means and how it applies to what you've just described in terms of the data strategy that you have in place as well. So let me start by describing a little bit about what AI can do, which is unique to AI, right? So AI is very good, especially the machine learning or classical or traditional AI is really good at taking vast amounts of data and doing complex math on that data that gives us an unfair advantage in how we operate and manage HVAC for large and complex buildings.

8:18And that's the obvious value proposition, right? We acquired Brainbox AI, which is the leading AI company in the HVAC business in 2025. And over the past two years, I've been on a mission to really help buildings of all sizes optimize their energy consumption. So that's the value proposition, the promise of AI and HVAC, right? And because HVAC generally accounts for about 50%, sometimes more, but 50 % of the energy consumed by any building, driving the cost of running HVAC down has a very material impact on the electric bill of any building, right? So CFOs love this, COOs love this, building owners love it, operators love it.

9:00So that's the business case. That's the ROI. But let's talk about what it takes to drive that ROI. The first thing it takes, as you described, is data. And we are pulling a lot of data about how the building operates from the building in real time. And I believe as AI providers, we have a fiduciary responsibility to make sure that the data is safely collected, the data is safely used and the data is then destroyed when it is no longer necessary. So there's a lot of rules and guidelines that we follow. Most of them are self-imposed and many of them are industry standards to help us be responsible and ethical in the use of data for AI.

9:45What's the advantage of doing this? Number one, trust. We are a 130-year-old company, as I said, and we have trusting relationships with our customers. We want to maintain those. Number two, cybersecurity. We want to make sure that we leave absolutely no window for cybercriminals and bad actors to take advantage of HVAC systems. And number three, and perhaps most importantly, value. We want to make sure that we're delivering constant value to our customers because at the end of the day, our customers trust us to do things that are right.

10:20Peter High: In our past conversations, Riaz, you've talked about a three-pillar plan for AI. Could you take a moment and describe the three pillars for us here? We have a very simple approach to using AI. We call it the three-pillar approach. So think of our strategy as having three vertical pillars. The first pillar is AI for growth. We just talked about that. It's about using AI to drive savings for our customers, selling it as a product. and us acting as a software company that creates AI, but also sells it to our customers and drives value. So that's AI for growth. We want to grow. We want to be known as a growth company.

10:58And as a growth company, we have to find new channels to market and new ways to add value. So that's AI for growth. The second pillar is AI for employee experience. And this is a very interesting area because here you're trying to give employees and colleagues access to AI tools. And it doesn't matter what tool you use. There's many in the market. Some people like Gemini, others like Anthropic, others like ChatGPT, and yet others like something else. But whatever tool you use, make it an enterprise standard and allow and enable employees to use that to make their job more fun. And the goal here is to allow employees to make their jobs less tedious, give them the freedom to experiment and use AI safely, and through that process, drive innovation across the organization.

11:50And these are all micro-innovations, right? People are using AI in different ways. Some are using it to summarize notes. Others are using it to manage their calendar. And others are using it to write job descriptions. So all of those and hundreds of other use cases are welcome. However, there is a third pillar. And the third pillar is AI for business transformation. This is when we want to be very surgical about the use of AI. And we want to have specific targeted use cases where we can bring in the right kind of artificial intelligence, often with deep domain expertise, to solve problems that will lead to a hard, measurable ROI.

12:31Let me give you an example. We buy a lot of freight. We're a manufacturer. We move a lot of things around and we invest a lot of money in buying freight from third parties. AI is very good at freight optimization. It helps us get smarter at how and when we buy freight. It helps us optimize the entire freight procurement process from start to finish. And because it doesn't take vacation or get tired or get bored, it's constantly watching the markets and helping us optimize our spot purchase of freight. That is a great example of AI for business transformation because now we are driving specific, tangible benefit.

13:10The same applies to other areas. AI for contract intelligence, as an example. AI for inventory optimization. We have a lot of inventory in our supply chain. And using AI to do end-to-end demand and supply planning allows us to take a lot of inventory out of the supply chain. So these are some examples. But overall, Peter, those are our three pillars. It's AI for growth, AI for employee experience, and then AI for business transformation.

13:38Peter High: Yeah. AI is often criticized for its rapidly growing energy requirements. And you've got some thoughts. You've even alluded to it already in our conversation, but I wanted to double click a little bit further about the means by which AI can actually help solve the energy challenges that it's contributing to ultimately. That's right. I mean, we invested in Brainbox because we believe in AI for good. And what does Brainbox essentially do? It uses traditional quantitative AI. It uses advanced thermodynamic and quadratic equations and calculus to solve complex math problems that helps be an autopilot for HVAC.

14:20And through that, we are reducing the consumption of HVAC by up to 25, in some cases, 30%. That is a massive savings without sacrificing comfort, right? So here's a case of where artificial intelligence is actually helping reduce the amount of power required. So there's a lot of talk today in the market, and for good reason, about the amount of power AI consumes. So I want to make two points over here. When we use the word AI, we're kind of lumping all kinds of AI into one bucket, right? Generative AI as well as what I consider quantitative AI. Quantitative AI is a lot less resource intensive.

15:01It's generative AI, which tends to be a lot more resource intensive. But if you focus on quantitative AI for a second, especially in the HVAC world, HVAC accounts for about 10 % of the entire energy consumed by the planet. Now, as a thought experiment, I'd like you to imagine a world where we can apply brain box technology to that 10 % of what the world consumes. And that can drive significant benefit. It can take out about 2.5%, 3 % and just take out that consumption globally. And that number is so big, it is actually four times larger than all of the power consumed by all the data centers in the world, whether they are AI data centers or regular data centers.

15:51And I think that thought experiment shows the power of using AI for good. This is a classic and very tangible example of where AI can help reduce the energy consumed by HVAC. The other number to keep in mind is that 70 % of the buildings in the United States and North America have HVAC. In the Eastern Hemisphere, in India and China, that number is only 30%. And they are rapidly adding HVAC as the world becomes warmer. So HVAC consumption is only going to increase over time. So it's a good thing that Twin Technologies is thinking about how to reduce the power needed to run HVAC. And for all those reasons, I think AI can actually be a force for good.

16:36Peter High: That's really powerful data you have there, in addition to the methods you're describing as well, Riaz. Thank you. You alluded also to some of the cybersecurity issues one needs to gird themselves against. It says more HVAC equipment and building management systems become connected. They create a larger potential attack surface. Talk a bit about the methods you're using to counteract this and how you foresee ensuring that that doesn't become an issue. We take this very seriously, Peter. You know, I spend a lot of time in the tech world in software and hardware companies, including Cisco, where I was president of the IoT business.

17:18And one of the things you learn about IoT, Internet of Things, is that the attack always happens at the weakest link. The cybercriminals are very smart, and IoT devices are some of the most hackable on the planet. Why? Because by definition, it's the Internet of Things, not the Internet of Humans. And a human being is using a machine. A human being can detect very quickly if the machine has been hacked. If it's just a connected device, the device might take a long time to actually signal that it has been attacked. And for that reason, hackers love IoT devices. So it's even more important for us to be very, very careful about connected devices.

18:00So here are some of the things we do. Number one, we make sure that we are applying some of the latest frameworks and regulations around the world and not only matching them but exceeding them. And we call this secure by design. We have a chief information security officer who works in close concert with my team. We want to make sure that everything we design is secure by design. So this is about regulation compliance. It's about making sure that we're using the best technologies. The second is using the best capabilities, both software and hardware, to make sure that our infrastructure is built on the latest technology.

18:40This is a never-ending battle. So things are constantly evolving, and we want to make sure that we're using the best software to provide the highest level of security. And the third, perhaps the most important, is that we want to make sure that the data collected is used in a very responsible fashion. And I cannot overemphasize this, Peter, especially in a multi-cloud world where you have to have absolute trust from your customers to be able to use their data. Granted, it's about that HVAC, but even so, it's very important that we use the data for the purposes it was intended. And after that use, it is destroyed safely, right?

19:18So think of it as a value chain. There's no one piece that's more important than the other. I think they're all important and they all have to work in concert and they have to work 24 by 7.

19:28Peter High: Yeah, fascinating. Fascinating insights indeed, Riaz. Connected equipment generates far more information than people can analyze independently, and especially as somebody who has had major Internet of Things responsibilities in your past, as you just alluded to. How does AI represent the next natural stage of IoT, and where are you seeing the greatest value from converting that stream of data into decisions? You know, when we started the IoT revolution, if you go back and look, it started around 2010. That's when the term was coined and people started using it. By 2011, 2012, it was part of everyday vernacular in the business world, right?

20:07And then it hit a peak around 2015, 20. So IoT has been around for that long. And when we started this journey in IoT and different companies have had different journeys, one of the things we realized as practitioners is IoT creates so much data, it is humanly impossible to analyze all the data. It's like drinking from a fire hydrant, right, a firehouse. And it was just too much. So the irony of IoT in the early years was that most data generated by IoT was essentially wasted. It was digital exhaust. And nothing was done with it, right? At best, companies stored it. At worst, they didn't even store it.

20:51They collected it and destroyed it almost immediately. And there was almost no value from it. They were collecting it and they were learning how to collect it. And that by itself was useful. And then the AI revolution kind of arrived. So AI is almost this new magic tool that allows us to make sense of the terabytes and petabytes of data that IoT generates. As an example, in trained technologies, we collect over 5 billion data points per day. I let that sink in. It's not per year. It's not per month. It's per day. And by some estimates, we will be collecting over 20 billion data points per day by the end of this decade.

21:37That is the rate of growth of the amount of data being generated. And AI is the perfect toolkit to analyze the data, to make sense of it, to do pattern recognition, to do defect, you know, identification based on patterns. and, as I've described, to become an autopilot for complex HVAC systems. And, you know, the autopilot idea is an important idea. And if I could explore that for a second, you know, most people listening to this podcast take flights almost on a weekly, daily, monthly basis. And if you're flying from New York to Los Angeles, you know, after the plane takes off from New York, the pilot essentially engages autopilot.

22:21And the plane is at that point, the autopilot is really flying the plane, so it's flying itself. And what does an autopilot do? If you think about it, the plane is pointing in the right direction for less than 5 % of the time. So for 95 % of the flight, the plane is actually, the nose is pointing in the wrong direction. And the autopilot makes tiny adjustments to the flight-controlled surfaces and points it back in the right direction. These are very small adjustments. And it's making those adjustments based on wind direction, altitude, and a variety of other factors. And eventually, the plane reaches LAX and the pilot lands the plane.

22:56Without autopilot and without those tiny adjustments, the plane could well end up in San Francisco, you know, worst case scenario in Seattle. So autopilot has that kind of impact because it makes small changes on a regular basis. And that's really what AI can do in HVAC systems. Complex HVAC systems have hundreds of controls, which in turn have thousands of parameters that can be constantly monitored and tiny tweaks can be made to those parameters. And AI can do those tirelessly to land the plane, so to speak. The difference between a plane and a flight and HVAC is that a flight usually lasts for a few hours.

23:37HVAC is a 24 by 7 game. So you have the opportunity and the responsibility in our industry to operate a system which is literally mission critical to most buildings and it runs 24 by 7. Yeah.

23:51Peter High: In our past conversations and this one, Riaz, you've spoken with great excitement about the promise of millions of buildings operating as intelligent, flexible energy assets. And I wonder, how far are we from achieving that reality? As you pointed out a moment ago, it has major implications not only domestically but globally as so many more buildings around the world and important developing and developed countries coming online with more mature HVAC capabilities. Talk a bit about the pathway to achieving that. Peter, one of the most important consumers of electricity on the planet are buildings.

24:31buildings consume, even today, despite all the talk about data centers and despite all the talk about some of the other assets, buildings still consume the vast majority of electricity on the planet. And the numbers range, depending on country, from 50 % all the way to 70%, 80%. So the buildings are consuming the vast majority of power in the world. And therefore, buildings represent the biggest opportunity. Because as we move on this journey, we are on a mission at Rain Technologies to make buildings smarter and more resilient. You know, smart, resilient buildings. This is the future. So if you think about most buildings today, right, they have investments in either fossil fuel-powered boilers to heat the building.

25:21They have access to the grid, and the grid is supplying energy that is used for heating and cooling. Some buildings have geothermal. Other buildings have on-site generation. Some buildings have solar. So there's a plethora of options that buildings can choose from. Some buildings have ice tanks. And if you think about what's happening with the battery industry right now, the battery technology has improved to the point where a lot of buildings are using batteries for peak shaving, right? In the commercial sector, which is a sector I really want to focus on, peak rates for electricity are dramatically higher than non-peak rates.

26:02As an example, in the state of New York, in the city of New York, a kilowatt of power at non-peak would be as little as 15 to 20 cents per kilowatt. That same kilowatt, if purchased at peak, could be as high as$50 to$60. It's the same unit of energy, just being purchased at a different time. So think about the huge opportunity for buildings to use all of these different technologies that I just described, and use AI and software, a control center, if you will, to modulate and choose from these different sources at different times of the day to reduce the peak that the commercial building is paying for.

26:52Now, the other factor that most people don't realize, and I certainly didn't until we started working on this, is that once a building hits a certain peak, That is the rate that you pay for the next 60 to 90 days until you hit a new peak. And then that becomes the rate. So this is all the more reason that we focus on reducing the peak. So bringing the peak down has tremendous economic value. We are on a mission to help buildings do that, but eventually to help groups of buildings do that. So think about buildings as batteries. Now we have the opportunity to use software and AI to help buildings optimize how they use their power from different sources.

27:33And to do this at different times of day automatically, autonomously. So we are on this journey. And I think the first step in this journey is to use AI to actually bring down your consumption. But the second step in the journey is to combine that with battery storage, with on-site generation, if that happens to be available, to use thermal storage ice tanks and to use software to modulate these different sources to bring down peak even more regularly. And then eventually in some parts of the world, Europe is a great example. European countries have always been very good at pooling resources together.

28:07Think about groups of buildings doing this together. So it's not just one individual building, but it's groups of buildings. By the way, Peter, the same thing played out a hundred and some years ago with the electric revolution. Today, we don't even think about it. We just plug in an appliance into an outlet and we expect it to work. You don't even think about where the power comes from. Eventually, that would be the case with thermal heating and cooling. But we are still ways away from that future. However, we are on a path to get there. And the concept of a smart, resilient building can have material impact on how much energy the world needs and how little energy the world wastes.

28:48Peter High: Fascinating. Riaz, this conversation has certainly given us all a window into the breadth of knowledge that you have, the ways in which you are bringing a vast array of experiences to bear in this current role of yours. I wonder, in this fast-paced, fast-changing business climate, what do you read or listen to in order to stay current that you would recommend to others? I try to read as many different perspectives as possible. And I also listen to as many different perspectives as possible. For example, I like podcasts. I like to listen to your podcast. I like to listen to The Economist. I think they have a very nice podcast now.

29:29I listen to a number of other speakers who I think have very interesting podcasts. I also listen to a lot of scientists. Amil de Grasse Tyson has a very nice podcast these days. And I think it's interesting, just it activates different neurons in your brain. In terms of reading, I strongly recommend, you know, folks, again, read different publications from different parts of the world. I actually think having a well-rounded perspective is more valuable than just kind of reading, you know, one thing and kind of becoming an echo chamber to some extent. So I strongly recommend having that diverse reading.

30:00So I read a lot of, I read a lot of publications, you know, I read some of the obvious ones like the Harvard Business Review and some of the reviews that MIT puts out. But I also tried to read the Wall Street Journal, a lot of different newspapers, The Economist. And then I read publications from Asia as well. I think a lot of innovation is coming out of that part of the world, the Eastern Hemisphere now. And it's a nice balance and a nice complement to what we hear here in the West.

30:25Peter High: Well, given the breadth of perspectives you've shared, it's no surprise that you've got a broad intellectual diet. Thank you so much for sharing a bit about that, Riaz. And thank you more generally speaking for a fantastic conversation, emblematic of the remarkable work you and your colleagues are doing in trained technologies. I appreciate you sharing those with me today. And thank you, Peter. Love your podcast. Love your style. And thank you again for having me as a guest. I appreciate it. Thanks for listening to Technovation. If you enjoyed this conversation, subscribe and leave a rating. It helps more leaders find our podcast.

30:58Peter High: Thank you. And I look forward to seeing you next time.

31:06Thank you.