Episode Transcript
[00:00:03] Speaker A: Welcome to AINA Insights where prominent leaders and influencers shaping the industrial and industrial technology sector discuss topics that are critical for executives, boards and investors. INA Insights is brought to you by INA AI, a firm focused on working with industrial companies to make them unrivaled segment of ONE leaders. To learn more about INA AI, please visit our website at www.ina.AI.
[00:00:40] Speaker B: Welcome everybody to another episode of Titanium Economy podcast series by aina. Today we have Vijay Shankaran from Johnson Controls with us. He is the Chief Digital and Information Officer. Johnson controls is a 140-year-old company focused on smart sustainable building with operations in over 150 countries.
Johnson control focuses on automation, fire security and thermal solutions. You're focused on solutions for airports, data centers, hospitals, essentially critical building applications.
Now Pooja, you've actually spent time at multiple countries companies before this.
You joined Johnson controls in 2021 as chief technology Officer leading the digital mandate and since 2024 you've now taken over the digital and the customer facing mandate to look at all of innovation and development. Before this you were CIO at TD Ameritrade, you spent a decade plus at Ford in executive positions and you're also on the board of Feraldo which is a water and a food focused company. Vijay, it's a pleasure to have you on our podcast today and thank you for making the time.
[00:01:54] Speaker C: Thank you.
[00:01:55] Speaker B: Let's just start with understanding a bit about Johnson Controls and your mandate.
Can you give a quick sense of what Johnson Control does today and essentially what this strategic sharpening means when it comes to the digital agenda?
[00:02:09] Speaker C: Yeah, absolutely.
Johnson Controls first of all is 140-year-old company and has been serving different facilities of all shapes and sizes over the course of that 140 years. And, and most recently the focus has been around thermal management. And when I talk about thermal management it's really in the breadth of how do we help mission critical environments, data centers optimize their cooling inside of their facilities and then add the control systems on top of that to be more precise in terms of that optimization. And then the digital and the AI to help those control systems and those H Vac units optimize around energy and preventative failures and things of that nature. And so we feel that what we are really embarking upon is to address the most advanced and most precise reliability and real time criticality type applications that exist across all kinds of facilities. So it's pretty cool to be able to see the byproduct of the different elements of the innovation that add up into that.
[00:03:32] Speaker B: I Think one thing, you mentioned this building and you said it's 140-year-old company, started with thermal management, more equipment and hardware, now transitioning to software solutions, and it's in 100 plus countries right now.
This is as much as a people problem than anything else at this scale. Could you talk to us about how did you sort of work through these challenges of taking a equipment hardware company to think more like a solution software company?
[00:04:07] Speaker C: Yeah, well, I think first of all, it's really important to kind of just be clear that it's the total package that brings the value proposition to life. And I think there might have been a point where we pivoted a little bit towards more of a software centric perspective. But really through that process learning and recognizing that the thing that the customers value the most is the innovation on our highly performant cooling solutions in the marketplace. And so, you know, when we, when we took a step back and said, hey, what problems are we trying to, to solve, right. On behalf of our customers, cooling is an enormous problem. And you know, whether it's, you know, we're here in Chicago and Michigan and you see the heat outside here, and there are other parts of the country and world where it's sweltering even more. You think about a hospital in Houston, Texas, in the middle of the summer, if they're H vac unit or chiller goes down, they have to relocate the ICU in the hospital. That's a big problem. And so the mindset is now that alone is critical to that facility.
But without digital, you can't do it precisely. You can't know the early warning signals that basically would indicate that there's potentially a failure in that system.
You can optimize the temperature in real time based upon when different fronts are coming through or what the outdoor conditions are. And so when you look at thermal management, it's actually a very sophisticated kind of end to end system.
Everything from the chiller itself to all of the vavs inside the building, as well as the know, the air handling units to the actual chiller plant itself, that needs to be optimized. Right? And you know, in the past these have all been standalone mechanical, you know, mechanical dynamic units. And now the, the big opportunity with digital data and AI is on top of that being able to drive those level of optimizations. And so the evolution of our talent and our culture has been to teach people to think from more of a systems lens, which is what is the complete package of all of those capabilities that really provides the end customer with the Most value your typical chiller service technicians or your product engineers or the installers.
They're not typically digitally savvy individuals. And so how do we teach them about the value proposition that digital brings to the table to add more energy efficiency, to add more predictive capabilities. And so what we've done is we've introduced a whole suite of different digital solutions on top of it. Everything from connecting all of our devices and getting the telemetry on our chillers and controls units and helping our service technicians with a to look up solutions in the manuals all the way up through how do we connect all of the endpoints in the building to optimize the facility itself using AI? And some of those capabilities we've acquired and some of those capabilities we've built ourselves.
[00:07:45] Speaker B: Got it.
Just drawing the same thread from digital and moving to your OpenBlue platform. Now, OpenBlue has not transformed into an AI powered ecosystem and a platform for Johnson Controls.
You have Predator maintenance, you've got autonomous control, and you've got a whole set of applications running to sort of help the facility managers. I think in your own experience, where are you actually seeing measurable ROI impact and which places have been challenging going into it? That hey, we would expect a lot of impact, but we haven't seen it. So in experience, where is impact?
[00:08:26] Speaker C: Yeah, I mean, what I would say is I've been at Johnson Controls now for over five years and I think when I came here the thesis was that digitization in buildings would follow the same patterns of digitization in other industries. Right. Whether we've seen that in automotive, we've seen that in healthcare, we've seen it in lots of different places. Right.
And I think what I've learned over the last five years is that it's more complicated to digitize buildings than other different forms of businesses. And largely it a has to do with the complexity of the different types of equipment inside the buildings. You have 50 different types of OEMs providing devices inside the buildings that may just be for one building. Right. And then you have a portfolio of buildings.
Some buildings are owned, some buildings are leased, some buildings have third parties operating them. So there's a whole set of business dynamics that make digitization of buildings that much harder. Right.
That being said, what we've really tried to do now is to focus our energy on building solutions that can provide tangible outcomes for mission critical and large real estate type customers and portfolios. And the value propositions come in a couple of different dimensions. One is energy efficiency. So how do you deliver 10 to 20% energy efficiency annually by optimizing cooling inside of a building. That's a really important value proposition.
Second is predictive monitoring. So there are lots of faults coming off of different control systems inside of buildings. So how do you apply AI and data in real time so that an operator knows that they need to submit a maintenance ticket or service ticket quickly so that they can address an issue before it becomes a larger issue.
And then the third thing is, so that's really around the reliability and precision controls and the third thing is really around decarbonization and optimization of your emissions and buildings in order to comply with whatever regulations that may exist in different places around the world. In places like New York City where you have Local Law 97 that has certain requirements around carbon emissions. Most of these buildings don't have a sophisticated tracking mechanism to understand what their emissions are and what their different equipment and building are emitting. And so that's another value proposition. So what we've really tried to do is not try to be all things to all people, but really have a very deliberate value proposition around our strategic vectors which are around mission critical and sustainability and decarbonization and now really expanding into data centers as that space takes off as well.
[00:11:35] Speaker B: Let's talk about data centers for a bit. Right? It's your fastest growing segment and like you said, mission critical, extreme uptime and cooling requirements.
Could you talk to a bit more about how is JCI evolving to sort of address this segment in the best way possible?
[00:11:54] Speaker C: JCI is in a very unique place in terms of its positioning around thermal management in data centers. You know, I had the, the, the fortune of seeing this evolve over the last several years and was here when we acquired Silent air back in 2022.
And what we bring to the table is sort of that end to end management of the thermal, thermal management and thermal chain within the data center. So everything from you know, what is the air cooled chiller and cooling required inside of the data center to what do you get, what do you need in terms of CDUs that sit close to the rack. And now with the acquisition of Alloy, how do you do it directly on the cold plate itself inside of the rack?
I think seeing the opportunity to string these all together as one end to end package and then providing the services on top of that for data center customers to have that level of reliability and uptime is a differentiating value proposition in the marketplace. It is also an enormously high demand segment in that when we acquired Silenaire at the time I think it was people wondered, hey, all of the revenue is concentrated in some of the major hyperscalers.
And I think in a way maybe that played out very positively for the company in the sense that those are the fastest growing players in the data center space. And we've been able to leverage what we've learned from the solutions we've developed to them to serve other hyperscaler customers as well. But it's not just the hyperscalers, it's not just the data center operators. There's lots of new data centers that are popping up all over the world, whether that's in India or in the Middle east or in Europe. And I think by understanding the problems that different operators face in different types of data centers, we're able to create the right solution package that optimizes thermal management inside of that. And that's a pretty cool place and space to be in right now.
[00:14:21] Speaker B: That's interesting. I think just the data center portion and what you mentioned earlier, right, you're focused on developing solutions for key problems and you spoke about predictive maintenance. Can I tell our facility managers before something fails? And the second thing is about energy efficiency. And yet those are applications going into OpenBlue.
So just thinking ahead, what's next for OpenBlue? What is the next area that you're now focused on? Next application or the next solution?
[00:14:54] Speaker C: Well, I mean, it's interesting, because I've had recent conversations around this, is that we are right now in the process of launching our mission critical agentic solution for OpenBlue. And that is like the next generation of OpenBlue. And so what that essentially does is it transitions OpenBlue from being a set of dashboards, a set of insights, to becoming completely agentic in the way that it works in the sense that whether it's telling the building to basically optimize certain conditions or consuming the information from the building in real time with whatever questions that you may have about the faults that exist, comparing the relative performance of different facilities as part of your portfolio, potentially even comparing yourself to other similar Portfolios in the OpenBlue customer set, bouncing yourself off of different building standards, like ASHRAE, for example.
What we're doing now with OpenBlue is really creating that Persona centric relationship for a person who operates some aspect of a building with whatever data exists about the building. And so all of the data is brought in standardized, curated. We used AI to normalize that, put it into, you know, a standard building entity data store topology. And then we built a number of agents sitting on top of that basically to that are very vertically focused to. We're going to start with, you know, hospitals specifically in healthcare and mission critical environments. And then beyond that we'll go to more vertical applications. So really what's new is that we've taken what we've learned over the last five years about, you know, the differences in different types of buildings and we're built bringing a much more agentic approach to all of that and really customizing those agents around verticalized applications.
[00:17:06] Speaker B: That's very interesting.
You mentioned about building systems.
I want you to dive a bit deeper into the foundational aspect of building Open blue. Right. I mean especially Johnson Controls is across multiple digital systems. You've got Azure, you've got Snowflake, you've got a bunch of ERPs to work with.
That's quite a challenging environment of bringing data together. Can you just talk about how did you sort of work around that? What does it really take to sort of bring that data together both internally and externally?
[00:17:36] Speaker C: Yeah, I mean I think that's, you know, I've always been very passionate over the years about data. I mean I started out as a data geek out of college and I would say that's still fundamentally at my roots. And I think if I look at how we're leveraging data to enable our enterprise, right. You know, with the ERP complexity that we have and the commercial systems that we have, we're a global company that we've grown up through acquisitions in the past, right. It would be, hey, I've got all of these ERPs and I've got to figure out a way to bring that data together. Right now with agentic AI you can use your data lake, your data layer as a mediator to create agentic workflows. That makes everything much simpler to the end consumers.
A couple of great examples of this are we've launched a new workflow called Warren for our H Vac commercial teams and basically they'll get these 300 page RFPs from customers. They need to figure out what they need to bid spec price and get it back. They need to get special custom quotes on things like air handling units and things like that. It was a very manually onerous process, sometimes weeks to fulfill.
Now with agentic AI, AI does the processing of what needs to be specked out. AI basically creates the opportunity automatically. In Salesforce, an AI agent will automatically put that into our configure price and quote system and give the seller different options on configurations and things like that.
And so the agents are basically mediating that End to end workflow, and the Personas are interacting directly with the agents. They don't actually have any knowledge of what the backend systems are, what the underlying persistency layers are, so on and so forth. We are basically abstracting the complexity for them. And what that does is it brings an incredible amount of efficiency and streamlining to those processes. So over time, what our hope is is through an agentic architecture, all of the functions for a particular domain, whether that's data center or just traditional H Vac or Fire business, is that you can generate proposals, you can understand the engineering specs, you can understand the installation specs, and actually manage your product project directly through an AI workflow. And it doesn't matter what ERPs are underneath that or what Salesforce systems are underneath that. The AI is your engagement layer. And the same is true that we're using the exact same paradigm for Open Blue as well, which is in the future, people want to interact with agents and they want to get the information that they want. Okay, they don't want to, they want to generate a dashboard, they want to do it on their terms through the AI. They don't want you to give them a dashboard. Right. So a lot of it is architecture. And how do you connect all of this underlying complexity in a way and streamline it as well so that you can deliver that workflow centric experience through AI?
[00:21:09] Speaker B: One of the consequences of having more AI, more digital in buildings is they're more connected.
There's a lot more information flowing around.
And given the critical nature of the infrastructure that you work with, security risks are always a concern.
Can you talk about what are the security risks people and customers are most worried about today, and how are you addressing them?
[00:21:38] Speaker C: Yeah, I mean, I think this is a great question, especially in this new age of the vulnerability explosion that has been reported broadly.
First of all, having vulnerabilities on software is not a new thing. I've been in software for the bulk of my life and vulnerabilities have always existed.
Now what's changed is the speed at which vulnerabilities are being discovered and how quickly they can be turned into exploitable vulnerabilities to exploit infrastructure, and also the ability to exploit vulnerabilities that may be decades old. Right. And so what does that mean in the context of a building? Buildings have lots of OT systems, they have old OT systems, new OT systems, and to the extent that those systems aren't connected to the outside world. Right.
They're largely secured. Unless somebody were to introduce an issue when they're plugging in to do maintenance, something like that. But that also implies that there's some connectivity to the outside world. The reality is that we have to assume that everything is connected to the outside world because people want to pull data, they want to pull service telemetry, people want to do remote diagnostics. It's just a byproduct of the world that we live in. And so when we build systems like OpenBlue, like our Metasys control system, number one, we have to make sure that our software doesn't have any vulnerabilities that could be exploited. Right. And so we are going through some very thorough processes right now to make sure that we scan everything. We have been scanning everything for years now, but also just even making that more robust in this new era. The second thing is we acquired a company in 2022 called Tempered. And Tempered was a zero trust edge device. And what basically Tempered does is that it plugs in into your local OT network inside of a building or facility, and it creates a zero trust overlay network on top of that. So what that means is that all the data going in and out is completely encrypted.
And the devices that sit behind this device basically have a cloak on them so that you can't see what those devices are. And only the people that are specifically named users have access to those devices. And so it also reduces the attack surface. And so that's something that we're deploying as part of OpenBlue. We're deploying as part of our connected chiller solutions as well as our connected control solutions, and more pervasively deploying that all across the ecosystem. And we believe that for building customers, that's a way that they can de risk vulnerabilities in their building ecosystem.
[00:24:47] Speaker B: That's great. That acquisition is pretty interesting.
[00:24:50] Speaker C: Yeah, very.
[00:24:51] Speaker B: In the physical world of how to sort of plug the security issues. Well, we have one final question for you.
Just looking at your career arc, you've basically led transformation, digital transformations at automotive companies, at financial services, and now building solutions in your own view.
Where do these companies and sector stand in their AI journey and what will separate leaders from laggards?
[00:25:19] Speaker C: Yep, that's a great question. And this is a question that I spend a lot of time thinking about. Right.
So first of all, I think we're all in the first few innings of the AI journey, Right.
And to some degree, I think we're in the natural ebb and flow of what happens with innovation, which is everybody gets real excited about something and then Everybody goes through the discontent that's potentially generated through the complexity of solutions that are out there, the cost implications, so on and so forth. And so the first thing I'll say is that we're all learning on this, on this journey, right?
The second thing I'll say is that what will separate the winners from the laggards, right? Is what I would say has always been true about adoption of technology, right?
Those people who think about their, their processes, their solutions end to end, you know, and then apply AI workflows in the context of the processes are the ones that are going to win. The ones that are out there developing lots of agents and basically everybody's doing a little piece of their own workflow with agents are going to have even more fragmentation than maybe they had when they did that for applications because you're going to have agents gone wild, basically.
The second thing is I know a problem, an opportunity that we've seen for the last 30 years, and that's data.
The people who are going to win here are going to be ones that can really leverage a strategic data advantage in the context of these AI workflows to add value in the context of their processes and to their customers.
You know, AI is in some sort of a panacea to solving, you know, the data problems of the world, right? I mean, it's garbage in, garbage out to some degree, right? And most of what, you know, these large frontier models have been trained on is stuff, you know, and so is the stuff accurate? Is the stuff inaccurate? Is the stuff curated? You know, when you're working in mission critical environments, when you're, when you're fixing thermal management issues, right?
You can't have 90% good enough, you know, you have to give a service technician deterministically, you know, the right information so that when they go into, you know, repair a CDU or an air cooled chiller, they actually know how to solve that particular problem. And so only the information that's been curated as such too, that the AI can leverage in order to provide that solution to that problem and is valid in that context. And so we have to be extremely precise in terms of the types of information that get distributed through our AI workflows. And so that's where we put in a lot of effort into the AI, the data curation for AI, right? So where is the data coming from that supports all of our AI workflows? And so again, that's, I think, going to be a second key ingredient. The final one, I'll say is the talent, you know, and the ability to think differently. Right.
Because AI does require you to think very differently in the sense that just taking an existing process, mapping out all the steps and saying, okay, here, let's go apply AI to it doesn't necessarily give you. Build you a better approach. Right. You have to think about it more in terms of inputs and outputs. What can AI handle? You have to almost adopt the mindset that I'm going to start with, the mindset that AI can do everything in this process. Right. And what's the simplest set of steps I'm going to take? And I'm going to push AI to do that. And then there's going to be things AI can't do yet. Okay. But if you go at it the other way, where you're like, okay, every step I'm going to do a different agent, it's going to be somewhat inefficient and you're going to be managing all these agents. So thinking differently and finding the talent to help you think differently is also going to be a key differentiator of those winners.
[00:29:48] Speaker B: No, that's actually a very interesting point. Well, Vijay, we're at the end. We really appreciate that you made the time to sort of join us today.
[00:29:55] Speaker C: Yeah, thank you so much for having me. It was amazing to share insights on the industry and you know, what Johnson Controls is doing in innovating in thermal solutions.
[00:30:04] Speaker B: Oh, perfect. Thank you very much, sir.
[00:30:05] Speaker C: Yeah, thank you.
[00:30:11] Speaker A: Thanks for listening to INA Insights. Please visit Aina AI for more podcasts, publications and events on developments shaping the industrial and industrial technology sector.
[00:30:29] Speaker B: Sa.
