Willow On How AI Is Changing The Way Buildings Operate

[00:00:00] Regular listeners will know I always say that I only partner with companies that align with my values and what I'm trying to build here at Tech Talks Network. And NordLayer fits that perfectly by helping businesses stay secure without adding unnecessary complexity. And every company today relies on SaaS tools, whether it's collaboration platforms, CRMs or finance systems. The common denominator here is everything now happens in the browser.

[00:00:28] But here's the problem. Most browsers were never designed with business security in mind. This is one of the many reasons why NordLayer has introduced a new business browser that puts security, visibility and control directly where work happens. And it helps protect against things like phishing, malicious sites and unauthorized access, while also giving your teams a safer and more controlled way to work online.

[00:00:55] And the best part is it doesn't change how people work. It just makes what they're already doing more secure. So if you want to learn more, please head over to nordlayer.com slash browser and see how it fits into your workflow. But now it's time to get today's guest on.

[00:01:17] What if the buildings we rely on every day could actually respond in real time to the people inside them, the purpose they serve and the pressures that are increasingly placed on them? Well, my guest today is the CEO of a company called Willow. And we're going to talk about how AI and digital twins are moving far beyond dashboards and straight into a world of operational action.

[00:01:45] And I mean everywhere from hospitals and airports to campuses and commercial real estate. Today's conversation is about making the physical world smarter, more efficient and more responsive. And not only that, we've also got an inspiring origin story of how my guest got to solving these problems today. So if you found yourself wondering how do you turn buildings from just passive assets into systems that can actually think and act, you're in for a treat today. But enough for me.

[00:02:15] Let me introduce you to my guest right now. So a massive warm welcome to the show. Can you tell everyone listening a little about who you are and what you do? Thanks for having me, Neil. My name is Bert van Hoof. I'm the CEO of Willow. We are often called now a frontier company that's delivering operational AI to the build environment through digital twins. Long story around that, which I'm sure we'll unpack.

[00:02:41] But our platform connects data from buildings, infrastructure and operations into a unified digital twin that AI can now reason over and act upon. Right. So we help organizations run complex physical environments more efficiently, more sustainably and with better outcomes for the people who rely on them every day. So that's a key component is embedded in our mission statement, which we'll get to at some point in the conversation, I'm sure.

[00:03:07] But it's around to create a world where every building on the planet responds to the people, purpose and environments they serve. There's a lot packed into that, but that includes hospitals, airports, corporate education campuses, schools, commercial buildings, stadiums. Right. So all of these are not built to be buildings, but they have a purpose. And then the people who operate within them are part of fulfilling that purpose.

[00:03:32] And when we say environments, that's the environment at large, but also all the subcomponents in there. If you look at an education campus, you have classrooms, labs, athletic facilities. Right. So that's true for every environment like that. And so before Will, I was a partner at Microsoft where I spent 20 years working on everything from Windows Office solutions for health and life sciences. And for most of the last decade there in Azure, always in product and engineering.

[00:04:02] So I'm a technoid and I was part of the executive team that created Azure IoT and Azure Digital Twins. And that's where there's a bunch of history that led me to Willow, but worked with the whole ecosystem. So that's my quick intro. Wow, man, what a story there. What a journey you've been on. I'm right so far and still enjoying the current one as well. Yeah, and there's so much more that we're going to be looking at in a few moments about looking towards the future. But just going back to your origin story there.

[00:04:31] So you spent decades shaping smart buildings, digital infrastructure from creating Azure Digital Twins at Microsoft to leading Willow today. But going right back to the beginning there, what first convinced you that AI could fundamentally improve how the physical world operates? There's got to be a story there. Was there something that lit the spark in you? What was that moment? No, it's such a great question. Actually, there is a real story there. So I'll have to start with a little bit of history.

[00:05:00] When we all discovered the Internet of Things and when that revolution happened, we were definitely at the forefront of that. But we always talked about how it created binary value at the start, right? You went from dumb or silent physical assets with no data to having them being sensorized. And now you've got floods of data. And then it was always, well, now what? Right. So it's cool. It's cool.

[00:05:25] We got all of this, but it was pipelines and airplane engines, everything applied, so to speak. But it was very impactful in and of itself. So, but you went from simple connect and monitor where you then got basic anomaly detection and alerting to earlier versions of predictive maintenance using machine learning. So you started to see the early versions of AI kind of start to intersect.

[00:05:50] And elevators were a great example of that, of why predictive maintenance matters so much. When an elevator goes down and there's a safety issue, that may be a very quick fix. But in many cases, they have to be re-inspected and re-certified. So even though the fix can be almost instant, that elevator may be down for days until all these processes unfold.

[00:06:13] So that was a great example of early, hey, predictive maintenance is a big deal because there's all these ripple effects. And so that was an early indication AI before Gen AI will unpack a bit more. When people say AI today, they mean Gen AI and LLMs, but there's still the broad AI spectrum, which remains very relevant. And so then we started to think about beyond individual systems.

[00:06:38] When we plunged into buildings and cars, you had these complex environments where we realized, hey, we have to model the physical world aspects of that first. What does that all look like? And you had the 3D models, et cetera. And then how do you attach that signal from the physical world into that? And what are the systems that already exist, like the building management system or building automation systems and other components? What can I already harvest that I can bring into that model?

[00:07:06] And then where is there opportunity to look at new censoring capabilities that we can add into the mix and unlock additional scenarios? So that's also where the semantic web came in. That wasn't a new concept. Academia had been plowing into that for decades. But people tried to build what are known as knowledge graphs into SQL constructs, et cetera. Graph databases started to emerge. And we realized the importance of ontologies.

[00:07:36] That's, if you listen to Ballantyre and other big players now, that's a big, important concept to unlock the full power of AI into what we do, right? I often talk about bits and atoms. The traditional digital transformation was all about that. All the digital systems that people have been using, right? Software, language, and data. All the AI assistants and co-pilots today often still mostly live there, too.

[00:08:03] And web pages, social media, emails, spreadsheets, and documents, et cetera. So now we're talking about the next frontier with operational AI where, you know, you enter the world of atoms where the physical world comes into the mix. And that's where intelligence meets reality, right? Buildings, infrastructure, grids, cities. And operational AI is not just another digital layer. It's the unifying intelligence connecting the physical, operational, and human layers of the build world.

[00:08:33] So that's how it really morphed and emerged and advanced pretty quickly to all these different layers that I touched on. So it started with bridging the world of the building information modeling, the 3D models, the traditional on-premise building management systems that always control HVAC, sometimes lighting, sometimes fire and security.

[00:08:53] But then you got the explosion of all the in a built environment, your elevator, your smart meters, your waste management, smart parking, battery storage now, solar PV systems and microgrids, EV charging, indoor air quality monitoring, occupancy monitoring. Now we end up in research lab with fume hoods and this explosion of things that all somehow can intersect, interconnect, influence each other.

[00:09:20] But that all have been managed thus far as individual systems, data islands, system islands. And so, yeah, this is where started with the knowledge graph, how you build your data estate for your real estate and everything that sits in there. And then how do you bring that all together? And then how do you bring that all together? The spatial, the cap, the BIM, the static, the live data.

[00:09:42] When I say static data, that means things like warranty records, your work order tickets, your single line diagrams, riser diagrams, you name it. Anything of relevance that has relevant information, you bring that in. And then there's the IT systems that can give you data about whether your utility bills, how you manage your leases, all your maintenance, management and work order tickets, as I mentioned, et cetera.

[00:10:09] Any LOB type software you may have running that also feeds that purpose that I mentioned of the activity that builds that the building is serving. So, yeah, that's one aspect of it. And then, as I already mentioned, when people say AI these days, they often mean Gen AI and LLMs, but there's still the whole broader spectrum.

[00:10:31] Gardner did an interesting paper on this a while back where they talked about composite AI and say, hey, there's still the traditional expert systems, rules, engines, heuristics. There's optimizers, there's optimizers, machine learning, deep reinforcement learning, came back very much into the mix with DeepSeq, the knowledge graphs itself, simulation techniques, and then obviously Gen AI and LLMs and all the agentic aspects. Right. So it's a whole spectrum.

[00:10:58] Now, what was interesting about the paper at the time was that they said, hey, you know, not all of these techniques are optimized for all the use cases. Right. Some of these AI techniques remain highly relevant. And there's obviously cost performance aspects, energy consumption aspects, all of those things into the mix. And so at the time they mapped a bunch of use cases and were each of these different approaches and models were applicable or strong or had weaknesses.

[00:11:27] And then say, hey, oftentimes you want to combine them as well. And so we realized we had been on this multi-year journey where that's effectively what we had been doing. And so we mapped all of our use cases in similar fashion to help people explain that wanted to go deeper in the technical weeds on how it all works to express the unique nature of what we were doing there. So ultimately that led us to this whole concept of operational AI. You get all the data, spatial, static, live data.

[00:11:57] You combine that with the AI skills where different domain experts can come into the mix. That drives insights that are not just the traditional anomaly detection, but rich insights with quantified impact scoring. Impact can be cost. Obviously, it can be energy, it can be carbon, it can be wear and tear, it can be productivity impact. And all of these things can get modeled in and it's ultimately about driving action. That can be the traditional way.

[00:12:24] You create a work order ticket, somebody does a manual intervention. Increasingly, there's more and more autonomous approach to it, right? Where the system itself or an agent can take action themselves. You always want to start with human in the loop on those aspects. But that ultimately drives specific value outcomes, right? And as you go over all of these systems and start to understand how they all intersect and interact with each other and influence each other, that's where you get very novel outcomes in a way.

[00:12:54] And then there's a co-pilot on the side that's always there that understands all of these components and where you can go on reasoning streaks and ask follow-up questions where you can get strong advice. If you're a technician or an operator, you get an insight, but you also get diagnostics. We can do root cause analysis because we can traverse all these relationships and root cause more proficiently. The co-pilot can then help you prepare if you need to go on site to fix something, to say, these are the tools you should bring.

[00:13:24] These are the parts you should bring for this specific issue. When you're on site, you can get assistance and instructions on how to approach the issue. Now you run into an error. Traditionally, we know of these scenarios where a technician has to call into headquarters to find that manual. It takes a few hours to find that. And then on page 145, you find that error code. But now with the co-pilot, that's instantaneously, right? Because all those things live in the knowledge graph, live in Willow.

[00:13:51] And so that's the richness of it all and how it's evolved. But yeah, that was from these early sparks, how machine learning came into the mix. So we're going on that journey and then as new AI techniques emerged, and certainly when Gen AI popped onto the scene, we were very much an early adopter to bring that into the mix and build our co-pilot as well. And now it's obviously all the agents coming into the mix as well as we look at some of these autonomous behaviors.

[00:14:21] There's some interesting examples like things that people always go like, oh, I assumed we would have been able to do this for a long time. But occupancy-based HVAC control, right? A lot of buildings run on fixed schedules. They ramp up in the morning, the sunset when the sun goes down. But lots of these spaces, lots of these environments within the buildings are not constantly used, as we know. Certainly with hybrid work, et cetera, that's worse, especially in office buildings. But that's true in general.

[00:14:50] It's a very popular topic in higher education campuses as well. The labs aren't always used, et cetera. So if you can detect that, if people invest in occupancy detection for different reasons, right? Are we effectively using the space? Do we need all the space? Well, when you have that system in Willow, then we can now correlate that with what we call active control. And so if nobody's there, why would we heat and cool it the same way as if people were there?

[00:15:18] And that can be a number of signals, certainly sensors real time, which can be from very basic presence detection to advanced people count solutions with AI-based solutions, but also all the event scheduling, right? And in an education setting, you have classroom scheduling software. So it can be anticipatory. It's something scheduled to take place in that space. And you can combine that with sensors if they are available or with sensing data on occupancy signal, right?

[00:15:47] So it can be both anticipatory, but also reactive. I did prepare the lab because we anticipated people to be there for whatever reasons that got canceled or not moved. We don't detect anybody. Within a period of time, we can throttle it back down or do the setbacks. And that same is true in an office building on how you get your meeting room schedules, whether it's through office or Google Calendar.

[00:16:12] In a hospital, we can tie into systems like Epic, the EMR, EHR systems. Is there a patient in that bed or not? So those are all signals that we can take into account to bring smarts into that. And it turns out you can save tons of energy, tons of carbon by doing this at scale. And so these are like popular features. We now do similar things with indoor air quality.

[00:16:36] When you can sense those, we have customers on that rodeo where you can, if your CO2 levels get too high, you can pump in more fresh air. And again, these are typically independent systems that were acquired for different reasons that never talk to each other. So now with Willow, you can bring all of that together. So yeah, we can go on this journey, but it gives you a bit of a storytelling on the question and how AI came into the mix and where it's gone.

[00:17:07] Yeah, such a cool story. So many magical moments there. And before you came on today, I was doing a little research and I was reading how, Willow, you describe the platform as an active management system rather than just a passive dashboard. So what does this shift mean in practical terms for those building owners and operators that might be listening? No, another excellent question.

[00:17:30] Historically, most building software has been observational and you get dashboards that in essence tell you or show you what happened yesterday. But operators today don't need more dashboards, right? They need concrete guidance and action. And they need a co-pilot that can help reason over complex systems and problems, as I already gave some examples. That can help with root cause analysis versus treating symptoms.

[00:17:54] Then increasingly we'll need and want agents that can take autonomous action on things that are repeat actions that have high impact or once validated and trusted. Always human in the loop starting point. But what just means delays or no action, right? Or lower or no impact. Executives need help with capital planning. We have some cool blogs out there on examples of people use co-pilot now in those activities. How to address deferred maintenance.

[00:18:25] They need a life facility condition index or assessment. Another once every three year static inventory assessment. That's what's mostly the case today. Property managers for commercial real estate can get a view on all the components related to tenant satisfaction. So those are all those components. So the shift from an active system or to an active system means that the platform is continuously analyzing operations and recommending or triggering actions.

[00:18:53] There's a bit of a cool background story here where we talk about that with both great humility and pride. Most people or many people will know the Gardner hype cycles. And they obviously go through whenever there's innovation triggers. There's typically the peak of inflated expectations. Huge expectations at first and then the trough of disillusionment, etc. So we build our own version for Willow.

[00:19:19] There is obviously one on digital twins in and of itself, but I would say innovation trigger in 2018. People were all giddy about the core concept with a knowledge graph and bringing all this data together. And we were all going to solve world hunger and world peace and some other gnarly problems through digital twins. And it was all initially about that data aggregation, right?

[00:19:42] Captured the digital signal across the organization, the structure, model the data in the Willow knowledge graph in a secured cloud environment. People saw like that must be great, right? And saw early demos and it demoed sexy and all of that. But then they went through that through a trough of disillusionment, right? Like then it was like, well, what's the value, right? What's the ROI?

[00:20:05] And so Willow then spent two years doing all the traditional inside overlays and traditional business analytics on top of that and dashboards. And I would say even today, that's all necessary, but not sufficient. So then that started us on the slope of enlightenment, right? That's Gartner would call it in the hype cycles. Insights connect and synthesize that spatial static and live data. And then we created what we call our activate technology, right?

[00:20:34] Our original tagline was Willow know your world. That evolved into activate your world, right? What are you going to do about it? What's the value coming out of that? So an AI ready knowledge graph, AI skills that started to deliver measurable impact, saving, sustainability, security. We can dig into some of that more. But then the operational AI that we're now talking about was really the game changer ultimately, right?

[00:21:02] We now will activate, co-pilot the agents, deliver efficiencies into business operations, automating, optimizing, enhancing processes, right? So that's been that journey, right? So I went from know your world, activate your world, went from connect and monitor to analyze and improve to now transform and expand, right? So that's kind of the evolution. And we feel pretty strong and good about where we are today. And it really resonates with customers.

[00:21:31] It's a bit of a flywheel. And there's more and more evidence and customer stories and reference customers that make people very excited. So the other important thing I would say, we've now proven it at scale. We do large campuses and the Walmart campus with HPE. We have all the corporate real estate around the world. So these are massive deployments basically. So, and we're proving that out.

[00:21:58] The more you got, the more you have, the more you say, the more impact you can have. And we will get into some of the categories of ROI as well, I'm sure. But yeah, it's come a long way. It's really matured since that 2018 period to today. It's been a long journey. And as I said, we tend to talk about that with great humility on all the things we had to learn, but also great pride of where we are today and the real impact we're delivering.

[00:22:25] And so much of what we're talking about here has come a long way. And when we, I think when many people listening think about digital twins, they were once seen as visualization tools. So I'm curious, when we talk about coming a long way, how has AI moved from that static models into systems that drive measurable operational outcomes across different types of buildings? Because again, it's a relatively short amount of time, but we've come a long way here.

[00:22:52] Yeah, no, it's such a great and essential question. Like for a while, we actually stopped using the term digital twin altogether when I joined the company, because I felt it meant so many different things to different people. BIM providers, building information modeling, providers, culture solutions, digital twins, some merit to that. All right. All of the interpretations had some merit, but it became hard for potential customers and buyers to distill clear signal.

[00:23:22] From the noise. And then the earliest, more expansive and advanced solutions like Willow that brought in that life and spatial and static data were initially perceived as highly complex and expensive endeavors, only accessible to the richest organizations. Right. We would show SoFi Stadium and one Manhattan West from Brookfield at 2.1 million square feet high rise in the middle of Manhattan and cool demos.

[00:23:49] But people were like, yeah, it's just for the rich and famous. Right. Right. And I'd say a lot has happened since, right? The AI revolution has helped spotlight the relevance and significance of digital twins. Even Jensen at a video head talk said there's events about digital twins routinely, the concept and knowledge around knowledge graphs, ontologies, which it touched upon. And the relevance of those in enabling this operational and physical AI as well, my bits and atoms comment.

[00:24:19] And at the same time, we've been hard at work to democratize the concept by not only making it affordable to own and operate one, but also fast deployment using also their cutting edge AI techniques on how we ingest all of this data quickly from all of these systems and documents, which takes out the cost and complexity. And it really accelerates the ROI.

[00:24:44] It's such an important component that's often not talked about enough, but we do believe we're kind of leading edge there when it comes to that. Like we, if all these, these document types that I mentioned that help us build a knowledge graph, we used to have an army of architects and high end mechanical engineers that would read through that. And yes, we would have scripting tools, et cetera. But now things that used to take days to just ingest now take minutes.

[00:25:11] So that's, that's a game changer in and of itself. And so that allows people to, to stair step their way in to what we call our activate packs, right? You can think of those as thematic value bundles or ROI bundles aimed at clear outcomes. And there's no set of costs, which is mostly unheard of in the industry. So it's, it's truly now easy to get going. These activate packs can be building energy and operations. We have one around sustainability.

[00:25:41] There's one around occupancy. It even goes to kitchen equipment, for instance. The nice example of going well beyond HVAC. That's the whole industry is still myopically focused on, on HVAC and energy. Nothing wrong with that. That's very impactful area. But we already talked about the intersection between the HVAC and things like occupancy sensing or air quality monitoring and, and how you then may affect or control the HVAC system. But it goes well and beyond, right?

[00:26:09] That there's refrigeration systems, which can be thermal units, et cetera. But big impact in, in both grocery stores and how you manage, manage those things and cost avoidance and the cost of downtime. When any of those units can't hold their, can't hold their set point or go down altogether. Now you have all the rapid ripple effects on FDA regulations. You need to throw the content because that becomes unsafe.

[00:26:35] It can be 10 to $60,000 per rack, depending on the content of it, whether it's yogurt or wagyu beef, right? Different effects. But then you go into the medical field and warehouses where you distribute vaccines and things like that. Right. So, so it goes, it goes pretty far, but yeah. So the early generation of visual twins were primarily visualization tools, essentially a rich 3d or data model of a building.

[00:27:00] We then added that static and live data context from IOT, OT and IT systems. And what AI has added is, is context and reasoning, right? It's a built on top of a knowledge graph. AI can now understand and reason over how equipment spaces, sensors, and operational processes all relate to each other. And an interesting concept as well. Any of the listeners that use chat GPT or Claude or Gemini, whatever, there's the concept of memory, right?

[00:27:29] So these systems now learn about you, who you are, what you do, what kind of questions you've asked. And so they get more and more proficient as they get to know you. Well, Willow does that for your physical assets, right? So you build up a history. So if you now have a couple of assets that may yield exactly the same insight on any given day, they may be the same brand maker model and they were installed on the same day. They may have a different history, right?

[00:27:58] So we have all kinds of, we have the warranty aspects of it. We have the work order tickets to happen. One of those units may never have gone down since you installed it. The other one may have already had four incidents. And so we can give you different guidance and recommendation based on that history, on that memory. So just as people experience this at the personal level, this is now true for all your physical assets as well within the Willow environment.

[00:28:25] So yeah, our mission is, as I said, to create that world where every building on the planet responds to people, purpose, environments they serve. When it comes to driving outcomes, we focus on three main buckets. Direct savings. Those are the clear and measurable financial improvements that show up immediately on the P&L. It's kind of hard map. It's energy, energy efficiency, shifting time of use, energy avoidance, like the occupancy-based H5 control, etc.

[00:28:53] There's maintenance and assets in that, the reduce of the wear and tear, commissioning, ongoing commissioning and validation of the assets. Proactive and predictive maintenance that feeds into that. Labor and work efficiency, workforce efficiency, right? So first time fix because of, as I said, I know which tools to bring, which parts to bring. I don't have multiple truck rolls for the same issue. I had a poor work order ticket.

[00:29:19] When the technician got there, it turned out to be quite different than what it seemed in a poorly formulated natural language query versus something that Willow generates, which has lots of precision in it, etc. So that creates maintenance efficiency. Again, all go straight to the bottom line. Very measurable. Then there's cost avoidance, right? Avoided revenue loss. Something goes down.

[00:29:42] The avoided downtime in general, like what we find in hospitals is they care a lot about energy, but the cost of downtime is everything, right? Given a trade-off, that's an easy decision. Avoided remedial work, etc. And then you have a third bucket, which we call the insurance bucket, right? The protection and risk group prevention. Why do people buy cyber software? Well, not because it contributes to the bottom line. You want to avoid the black swan event that can wipe you out, right?

[00:30:11] The same is true for your physical assets, right? We monitor all of your environment basically on an ongoing basis. And so there's safety and compliance protection around that cascading impact prevention. Water leaks, like very common in all environments really, but like in schools in particular. If you can do preventative detection of that or predictive better, that's one aspect. But even if it happens, if you have early detection, you can fix it.

[00:30:41] It's a relatively inexpensive fix. You mop up the puddle as opposed to it's been going all night. You come in, your classrooms are flooded. You have to remove the furniture. You have to bring in the massive dryers and dry it out for days. Now I have to do mold prevention. So your $300 problem became a $30,000 problem overnight. These are way more common than you think at airports, in hospitals, and in schools in particular. So that's a whole bucket that we guard against as well.

[00:31:11] And we put numbers around that when we do have those detections. And they're very non-controversial, even though they're not like precise math. But most people have their own history coming back to that memory. You look at the big school district and they will tell us, yeah, we had 62 of those incidents last year. It amounted to $200 million of cascading and ripple effects. So these are pretty common things. So I think that's kind of the holistic attempt to answer your question.

[00:31:40] Listening to you there, one of the things I love about you is you're not afraid to think bigger. And looking back at your career, you've already worked on smart cities and energy grid modernization. But I'm curious, again, if you look back, how do the lessons learned from large-scale public infrastructure, how do they translate into commercial real estate, healthcare facilities, airports, industrial campuses, etc.? Any crossovers there? Oh, yeah. Love all your questions. These worlds heavily intersect, right?

[00:32:10] A corporate or a higher education campus are small cities in and of themselves, right? Many of them are composed of hundreds of buildings. And they're really small cities. The Microsoft campus, where I spend so much time, is a city in and of itself in many ways. It's the Walmart campus that we operate as well in Bentonville, similarly. DFW Airport, one of our key lighthouse accounts and customers. And we have great customer stories around that on our website as well.

[00:32:39] But it's the size of Manhattan. People are always slabbergasted about the size of the thing. It has its own zip code, has its own police force, has its own health infrastructure. So, yeah, the comparison is very apt, right? So, in that context, Willow literally started with the traditional entry point, which is HVAC optimization in a single terminal, then went across terminals. But now it goes across multiple systems, across all terminals and beyond.

[00:33:08] There's digital glass, there's lighting, there's HVAC, there's occupancy. We do passenger boarding bridges with ground power units where the planes plug in and get their fresh air while docked. And that gets fed from the central utility plant. So, I should talk about grids, right? A lot of these big settings have their own central utility plants. They have their own microgrid. Some of them have their own solar and wind farms. There's conveyance systems, including escalators, elevators, luggage handling, water systems.

[00:33:36] I already talked about leak detection. In 2015, they had an undetected leak that ultimately created a sinkhole near the runway. And so, what a great example of this cascade effect, right? So, flight skip canceled, boom, boom, boom, $4 million impact. So, those are all great examples. Parking systems, EV charging of ground support equipment, et cetera, right?

[00:33:59] So, it started in a single spot and then it keeps fanning out, which is the value of something like Willow, right? And where these activate packs come in, you can start in a logical place. And again, democratizing this whole spot or this whole play. And you get ROI quickly. And then you can say, what other layer of the onion can we peel here? We want to be the gift that can keep on giving. And so, that's so key.

[00:34:25] The industry, as I mentioned, is still mostly myopically focused on HVAC and energy. Lots of play there. As I mentioned before, nothing wrong with it, but not sufficient, right? So, one of my biggest lessons from dealing with smart cities, also there are lots of PowerPoints. But if you then look at the actual success implementations versus the multimillion dollar projects that the curtain was pulled over.

[00:34:53] But that one, grid modernization, which we'll come back to in a second, is that scale forces that interoperability, right? Cities and energy grids involve thousands of systems from different vendors that were never designed to work together. And the same is true within buildings and certainly larger settings like the ones we operate in. And optimization tends to only happen at the system level, right?

[00:35:19] But you can't optimize energy maintenance and occupant comfort independently, right? AI needs to understand the whole operational ecosystem. So, I've been personally very active in and around the grid over many years. I've been a long-term board member of the Alliance to Save Energy. I've shared for five years the Active Efficiency Collaborative. And building pioneering demand flexibility solutions in Scandinavia, originally, which we won some awards for.

[00:35:49] So, everyone is now aware that there are plenty of big issues to be solved around the energy grid today. Explosive growth of data centers, unless you've been hiding under a rock, you've heard or read about these stories almost on a day-to-day basis. There's the reshoring of manufacturing, right? So, this is where the intersection of the build world and the grid is becoming crucial. The DOE has been very active, and we've been very active on the concept of grid interactive efficient buildings, or GEPs as they're called.

[00:36:19] Buildings that are aware in real time of the grid's needs, wants, and struggles. Buildings that are also available in the concept of virtual power plants, right? So, that's going to become a key aspect. There's been tons of interest in there. I've always claimed their aim, and we've done many settings as part of these collaboratives. Where for non-residential buildings, you absolutely will need something like Willow, if it's not Willow. So, AI that can operate across fragmented systems, that's where it's going to be at.

[00:36:49] That's exactly what we're now bringing into individual buildings, campuses, and large infrastructure settings, as some of the ones I described. So, yeah, it's very exciting. Part of the reason why I joined Willow was that I really got bitten by the bug on the big need. Like, I feel this is like a transformation, like the Uberization or the Airbnb transformation, where there has to be a fundamentally new and different approach to this,

[00:37:18] and which we believe we're offering a real solution and a real alternative to the status quo. So, yeah. And it feels like the right time as well, because energy efficiency is now tied directly to cost control, sustainability goals, ESG scores, etc. It could go on and on there. So, how are you seeing AI helping building operators move from that reactive maintenance to predictive performance and performance-driven decision-making as well?

[00:37:48] Again, big value add here. Yeah. Yeah. As you said, historically, building operations have been largely reactive, right? Something breaks, an alarm fire, somebody dispatches the technician, they get a work order ticket. AI changes all of that by identifying patterns before failures occur. For example, subtle shifts in temperature curves, vibration patterns, energy signatures often indicate that equipment is drifting away from optimal performance.

[00:38:15] Lots of additional things that goes across system now and understand how different systems influence each other. I mentioned like the digital glass. When it tints, it changes. You need to change the lighting. Your floor plate heat changes. UV lights get blocked differently, etc. So, how do you automatically adjust all of those? Today, all of these things exist. They're all physical assets with some SaaS solution attached to it, but they all work in isolation, right?

[00:38:44] That's where Willow comes in. We also often hear and get huge head nuts where people would say, hey, my technician comes in in the morning. They open their 16 tabs in the browser. And we always say, we don't want to be tab number 17, right? We want to be the one that, you know, you can live on Willow and do most of this. It doesn't mean we're completely displacing those other tabs remain available to you if you want them. But in most typical cases, you won't need them. So, yeah.

[00:39:12] So, instead of fixing things after they fail, operators can address issues weeks or months earlier with our solution. We've managed to do that within big settings where people had amassed the telemetry data of assets but felt they were still being very reactive. I mentioned refrigeration in grocery settings. Often, lots of millions spent on getting that signal in, but then still finding that we're too late, right? We can analyze after the fact, but we still had a food throw, right?

[00:39:43] Because once it goes bad, then the clock is ticking, right? You have to get to it within a certain amount of time. Otherwise, you still have the food throw. So, those are kind of essential capabilities where we could also do the simulation backwards and say, hey, we can replay the last year by ingesting all your data.

[00:40:04] And if you would have had Willow, we can prove to you that we would have detected three days, seven days, months before an issue that you have hard evidence that led to a food throw in your environment, right? So, you can also simulate and prove that Willow would have prevented those. And with some of these settings that can really run into the billions of dollars, it's all part of the shrinkage scenario.

[00:40:29] And as I mentioned, once you talk about vaccine, refrigeration, et cetera, any loss there is a different order of magnitude, right? But that applies to many, many other systems. I mentioned fume hoods, which has been a relatively recent addition as we get into hospitals and research labs. There's compliance aspects, right? You need to maintain negative pressure. How do you prove those things if the Joint Commission visits you?

[00:40:53] The intersection between fume hoods and a HVAC system, right, which all deal with air and air circulation and ventilation and how did they play off of each other? All those type of things become really interesting. So, yeah, instead of fixing things after they fail, operators can address issues as described weeks or months earlier. Reducing or avoiding downtime, big deal in certain settings more than others, but almost everywhere.

[00:41:19] Extending equipment life, we'll come back to that one as well, and cutting energy waste. So, that's why many organizations are literally seeing double-digit operational savings. And we discussed those three different buckets as well as ROI comes in in different ways. And then we have things like people do a facility condition assessment, and that leads to a facility condition index.

[00:41:44] And typically done every three or so years where people come in and they run around with clipboards and iPads, and they go assess all the assets in the environment and what state they're in. And that's where concepts like deferred maintenance come in as well. Certain assets have a predicted life cycle or lifespan.

[00:42:06] And then people mark those, and you want to do preventative maintenance because if this thing is end of life based on normal averages and documentation, then you want to, if it's a critical asset, you want to preventatively swap it out. Well, what if there's nothing wrong with that unit? And you could squeeze another two, three, four, five years out of it, and you have predictive maintenance running on it. Those things go straight to the bottom line, right?

[00:42:32] So you could start to then look at, well, with something like Willow, you get more of a life facility condition index as opposed to these manual activities that ultimately lead with a PDF produced. And that's out of date within at least months. So yeah, the high cost of preventative versus predictive maintenance, Willow helps reduce wear and tear as well. We do do continuous commissioning as well.

[00:42:58] One of our impact scores goes on wear and tear specifically. Predictive maintenance can go straight to the bottom line in that way. So yeah, just a whole setup and spectrum of things that went into that question and the answer. Yeah. So if we look further ahead into the future when AI becomes embedded into the fabric of physical infrastructure,

[00:43:23] what capabilities should building owners start investing in right now to ensure that their assets remain competitive and resilient over the next decade? Any advice that you would pass on to those building owners listening there? Yeah. I mean, I mentioned the catchy phrase on build your data state for your real estate, right? So that's a foundational thing. Create that unified operational data foundation. We know AI is only as good as that data foundation, right?

[00:43:51] So most organizations have fragmented systems. Bringing those into a common data layer is essential. So Willow can help you get there quickly, as mentioned, without huge setup costs, etc. Second, really consider digital twins and the relevance of knowledge graph. They provide the structure that allows AI to reason about physical environments as we unpacked quite extensively.

[00:44:16] And then third, really start taking advantage of these operational AI co-pilots and agents, right? The workforce running our infrastructure is aging as well. That's another one that's very widely recognized, right? There's some different interesting studies, but 68% of technicians and operators are above age 45. 22% are above retirement age. So that's a huge problem that people are grappling with, right?

[00:44:44] So many organizations struggle recruiting experienced operators. So that's why they're hanging on to people at retirement age. AI can help capture domain expertise from the seasoned operators, dramatically augment those teams and make the profession more attractive to young people, quite frankly, as well. It's a very different approach, right? That's hence my transformative thought around Uberization and Airbnb fundamentally changing the equation.

[00:45:11] So we feel that there's a similar motion happening that our type of solutions will compel to young people. It's a very different type of interface, very different type of interaction. And so organizations that do all the three things will be in a very strong position as AI really becomes embedded into the infrastructure of cities and buildings. No question. Exciting times ahead.

[00:45:37] And for anybody listening wanting to dig even deeper and unlock some more insights from you, I know this is an area you're incredibly passionate about. Anyone wanting to find out more information about Willow or connect with you or your team, where would you like me to point everyone listening? Yeah, of course. Go visit our website, right? There's a great unpack of what we do, how we do it. It's willowinc.com. So easy to go to.

[00:46:01] There's a very extensive blog section there as well, where we tend to want to explain a lot of our most differentiated capabilities, right? Where we think we're highly unique. We also have a bunch of customer stories and videos up there. They provide relatable testimonials and organizations' journeys with Willow and the impact it has had under operation. So definitely go to there. You can find me on LinkedIn. Search for Bert van Hoof if I should show up there pretty quickly.

[00:46:30] I share insights on digital twins AI, the future of operational technology. And of course, if anyone listening or is working on complex infrastructure and wants to explore how AI can help, we'd love to continue the conversation for sure. And I would encourage anyone listening to do just that. I will include links to everything you mentioned there in the show notes.

[00:46:53] And I cannot thank you enough for coming on here, not just talking about how AI is transforming outcomes for customers of various types of buildings there, but also I think most importantly is sharing the story behind it, the origin story, spending decades at the forefront of advancing technology, smart buildings, energy grids, energy modernization, and to present day, talking about how Willow is using AI to help the physical world run in so many different settings there. Absolutely fascinating.

[00:47:23] I'd be interested in any takeaways from people listening today. But more than anything, thank you for starting the conversation. Well, thanks so much for having me, Neil. I loved, enjoyed the questions. They were very on point and pertinent. So hopefully, yeah, listeners learned something new today. One of the things that stood out to me in this conversation was not just how much the role of AI is changing in a built environment. It's no longer about monitoring systems or collecting vast amounts of data.

[00:47:53] It's about creating a foundation where buildings, infrastructure, energy, and operations can all be understood as one connected environment. And then using intelligence to improve performance, cut wastage, and support better outcomes for the people inside these spaces every single day. So if you want to learn more about Willow and the work that Bert and his team are doing, please head over to their website.

[00:48:20] And as always, let me know your thoughts, techtalksnetwork.com. And let me know your thoughts on whether we are getting closer to a future where every building becomes responsive and intelligent. But while you marinate on that, I'm going to leave you now. But I will return again tomorrow. Thank you for listening as always. Bye for now.