Episode Transcript
[00:00:00] Speaker A: Constructive Voices, the podcast for the construction people with news, views and expert interviews.
[00:00:07] Speaker B: Hello, I'm Steve Randall and welcome to Constructive Voices. In this episode, we're talking about two things that are never far from the front of our minds the impacts of climate change and artificial intelligence. We're talking about Wint, a system which detects and stops leaks at the source using AI. It alerts you when water's leaking and automatically shuts it off. Constructive Voices producer Jackie De Burker has been speaking to Yaron Dysian, co founder of Wint and its chief product and strategy officer.
[00:00:40] Speaker C: So Wint is all about managing water in buildings, a topic that's really been overlooked for too many years. Water is kind of a dumb asset. It goes into the building, it gets used by the tenants or equipment in the building and nothing manages it. And things that we don't manage tend to do what they feel like, which means in the case of water spill and cause massive damage or just leak away and waste a lot of water. And the reason we started Wint, it's not just me by the way, it's me and my co founders. Wint is dealing with a problem that I noticed is not being dealt with. In fact, a neighbor of mine had a water leak in his apartment after having renovated it and he had to redo the whole thing at huge costs. And that brought to my attention know significant also. You know, as we see global warming and climate change take place in the world. One of the first impacts that these are having is water scarcity. Places that we would never have guessed, such as London or Paris are now becoming water stressed. It's a topic that needs to be taken care of and so these were some of the reasons.
[00:01:55] Speaker A: Okay. That's really interesting. Aaron. What was your background before you actually found it Wint? What's the paths that led you there?
[00:02:04] Speaker C: I've been doing tech for 30 years, over 30 years now. A lot of startup companies I've dealt with anything from cybersecurity to electro optical systems. And when you look at this world, the world of the built environment, buildings, construction sites, there are some of the slowest to adopt It technologies, there are good reasons for this. But when you think of the upside, when something is lagging behind that many years, you can make a big difference.
It's easier to make a big difference in a place like the built environment by adding it to it than for instance, banks who are extremely advanced in their adoption of it. So it's a great area to be helping and to be making a big dent for the world.
[00:02:56] Speaker A: Obviously, as you've already mentioned yourself Aaron, water scarcity is a press and global issue and unfortunately it's probably not going to get any better as we see climate change happening quicker than most of the scientists imagine it would even happen. How does Wint water intelligence contribute to addressing this challenge, particularly in the context of the built environment that you've just mentioned.
[00:03:20] Speaker C: So here's an amazing statistic. About 25% of the water that gets into a building goes to waste. We don't see it and there's no one overruling overriding reason it's death by 1000 cuts. A leaky toilet here, a tap lift open there. Misconfigured irrigation system, a problem in your HVAC system. All of these add up amazingly fast to waste so much water.
Now, when Wind goes into a building, we place sensors which are essentially water flow meters, the standard type that you would use to meter with your utility company. And we read this data, we read it at very high time frequencies and we use AI to analyze that water flow pattern. And when we do that, we can learn what is normal and what is abnormal. And an anomaly will eventually map out into one of two things either waste, such as a leaky toilet, or damage, such as a pipe breaking in the wall. And so when we go into a building and install our systems, they start alerting the staff that's maintaining the building. And things that have gone unnoticed are now top of mind and you can go and fix them. Usually the fixes are very simple and very easy to do. It's just being aware that that thing is happening. And so our contribution comes from the fact that we can identify that bad things are going on in the building. Alert the staff when needed, shut off water. We can automatically shut water off. And that's very helpful for damage prevention. And by just putting all of this stuff in front of our customers eyes, they can go and take care of it, resolve the issues and save that much water in the building. So typically our customers will save about a quarter of their water, reduce their water bills. And by the way, water has a big carbon footprint. So we will also reduce the building's carbon footprint throughout this process.
[00:05:18] Speaker A: Tell us in a bit more depth though. I mean, obviously these days everybody's familiar with the term AI and plenty of people are fascinated about it, some people are quite worried about it. How exactly does AI work with the pipes and the systems that are already in a building? How is this actually possible? Go into a bit more detail as to how it actually functions.
[00:05:38] Speaker C: So AI is really a very broad set of technologies. The AI that we all know with Chat GPT, which is generative AI, is using the same basis or same technological basis as we are. But it's different, it's not the same. Our branch of AI is what's called anomaly detection. When you do anomaly detection, you look at an environment, you look at a situation, in this case, the water flow in the building, in the third floor or in the irrigation system or wherever. You look at what is happening, there and you just observe it for a while, and you let the algorithms understand that this is normal. There's a variety of ways to do this. It's actually quite complex processing. But think of this and try to imagine this. It's tough on a voice podcast, but think of a chart that shows water flow rate 100 liters per hour, 50 liters per hour, 500 liters per hour. And it kind of creates a chart over time. So you can see, if I open a tap, you will see all of a sudden the flow rate increase. And when I shut it off, you'll see the flow rate drop. You look at these charts and you let the algorithm, the AI algorithm, analyze them. They actually create, if you will, a picture. This picture has normal behaviors, and it has abnormal behaviors. I'll give you an example. If we're in an office building, you do not expect to see that chart demonstrate water is flowing 24 hours, seven days a week, because during the night or during the weekend, the office building is empty. So when you initially go in, you will see that in certain times of the day, the flow reaches zero. But then again, on other times, it is used with certain patterns. So if all of a sudden I've got a weekend and the water continuously flows for 24 hours, that's not normal. So that's an example, kind of a visualization, if you will, of what the algorithm does.
[00:07:37] Speaker A: It's really fascinating. And don't worry about the charts. We can include them on the website for the episode page so people can see them. One of the things that I heard prior to our chat today, Aaron, is you were on another podcast, and you mentioned some quite startling statistics about water and the built environment, stroke, construction sectors. Can you share these with our audience, please?
[00:08:02] Speaker C: There are two areas to the impact of water in the built environment. The first is the waste. We mentioned that approximately 25% of the water going into a building goes to waste and generates fantastic amounts of carbon emissions. So if you think of a single example that we had, and I can mention the customer's name simply because it went public, but the Empire State Building installed us, and we identified water waste in one of their buildings. Electromechanical systems was wasting water at a rate of $100,000 a year that was just going down the drain, very difficult to identify because it was going through some internal mechanism of the system down the drain. Nobody notices an impact except for the water bills. And the second element of this waste is the carbon emissions. That specific example, when we calculated the carbon emissions that it generates was equal every year to flying a full passenger plane with 170 people between London and New York. And that's just out of some hidden water leak that's happening in some electromechanical system. So that's in terms of the waste the amounts of water and their impact on the environment. A second element is the damage aspect. Water damage in buildings is the number one cause of pain in quite a few insurance categories. Insurance companies are now saying water is the new fire. Where fire used to be the big problem for insurance companies in the past, that problem has not been 100% resolved, but has pretty much been diminished to significantly less of a concern. While water has risen up the list, water in certain categories accounts for a full 30% of the damages that insurance companies pay for the full category of insurance policies. So, for example, builders risk the insuring of a construction project, suffer 30% of everything. Every loss that happens in the building goes to water damage. Think of it, construction projects is a very risky location, right? People may, God forbid, fall off and hurt themselves or even die. Fires may. There's a lot of stuff that can happen in this kind of location. Yet 30%, a full 30% of insurance payouts come from a pipe that bursts somewhere in the construction site. So this is a very big deal for insurance companies and they're trying to sort it out and solve it and make it go away like they managed to do with fire. In the UK, for example, there is a group that's coming up with what's called JCOP, joint Code of Practice, which will enforce water leak mitigation systems on construction sites. So insurance companies know the problem is being dealt with and do not need to pay out more than is required on operational buildings, a building that's already inhabited. It's also a very big problem being the number three most common source of issues for insurance or claims. So if you look at it, water has two sides to it. One, it gets wasted with water stress. That's obviously not a desirable thing. And it also has that insurance impact. Finally, if you've ever been in a building or in an apartment or a house that suffered significant water loss, you do not want to be there again. I mean, the damage, it's like a silent killer. You get back into the facility after the leak has happened, the damage is phenomenal. Sometimes walls need to be torn down, carpets need to be pulled out, equipment, it's extremely painful.
[00:11:42] Speaker A: I guess you're talking about the likes of the kind of molds that can be formed after leaks such as that.
[00:11:48] Speaker C: That's exactly right. And they're extremely difficult to get rid of once they're in the wall. It's terrible, especially in wood buildings.
[00:11:55] Speaker A: But in general, there's so many aspects, obviously, from how you're describing yarn, so many aspects to wind. Sustainable construction obviously involves balancing environmental concerns with economic concerns, which is probably one of the biggest challenges at the moment as we try to shift and fight against climate change. How does wint strike this balance yarn and help construction projects actually become more sustainable without compromising their financial viability.
[00:12:26] Speaker C: Yeah, I mean I think that's one of the nice things that we're really proud of. The solution brings a combination of sustainability that essentially pays for itself with even some money left after all the math is done. So on the sustainability side, we reduce water waste, we reduce the carbon emissions. On the money side, that water saving actually translates to real money. So most of our customers will actually see a return on investment coming from just the water savings. There is obviously also savings that are more statistical in nature and a little bit more difficult to quantify or estimate, which is the damage of the prevention of damage. Because how much damage and what would be its cost would have occurred had we not protected the building is a very difficult question. But we've got some comparative analysis showing that we prevent approximately 90% of the damage. So that's also a component of the savings and the financial impact. And I think the combination of these two things ensures that you can both be profitable and sustainable. Many organizations are investing significantly to make their buildings greener. From a lead perspective or from a Brieamp perspective, the cost of a lead or brieamp point could be quite significant if you think of covering the building with special coatings on windows and so on. Lead points and brieamp points coming from Wince Water Management are actually low cost relative to most other lead points that can be gained.
[00:14:04] Speaker A: So they're a quick win essentially quick.
[00:14:07] Speaker C: And low cost win yes.
[00:14:08] Speaker A: Would you like to share costs yarin or is that sort of depending on a project by project basis?
[00:14:14] Speaker C: It does depend on a project by project basis. I will say the model is hardware.
Thousand meters we do not manufacture. We can provide them to the customer or customer can buy them on their own and it'll really vary. Right it depends on the pipe size.
The equipment for a small pipe could be 100, maybe 200 pounds per unit. Large pipes require big equipment that's heavy sometimes even needs mechanical lifting. So these things could go for thousands of pounds. It really depends. When we go into a project we will map it out and recommend what type of systems are needed per that project's plumbing system. In terms of the ongoing fees they will be at the tens of pounds per month per system.
[00:15:02] Speaker A: Okay it seems very economical obviously once you get your infrastructure in place at the beginning which is a little bit like a solar system isn't know your outlays there at the beginning and then the savings can be very worthwhile as you go along.
[00:15:15] Speaker C: Yep, exactly.
[00:15:16] Speaker A: Now next question where is Wind operational at the moment?
[00:15:20] Speaker C: Yarin we directly have offices in the US, the UK, london and Tel Aviv we have partners in other regions of the world. So UAE we have partners in Dubai and Abu Dhabi we have partners in Singapore we are starting to operate in other European countries. So we're actually quite international at this point.
[00:15:44] Speaker A: Excellent. Okay. Now, one thing that popped into my mind, because I am very curious about technology. How does the technology adjust on a construction site? That's something that's much more, at least in my mind. It would be more challenging to find out the norms and then spot because the construction site obviously is changing as construction proceeds.
[00:16:08] Speaker C: Yeah, it's a really good question. Construction sites are more hectic than a standard operating building. Yet, surprisingly, when you dive deep into the patterns, you see that there are things that are very common. And that's really the power of AI being able to decipher behaviors from what we as humans may think of as a mess. I'll give you a really simple example. There is never a flow rate of over 5000 liters per hour in the building. That's a very simple observation, but it's something that comes up and the system can learn it. Water is never used for more than 45 minutes, things like that. They're not all accurate and these are just examples, but that's the kind of thing that the algorithm can come up with. We have done some very deep studies that essentially will tell you what is the probability of being able to stop a leak in certain conditions. For example, a construction site up to a certain volume of water. So I can tell you what the probability of stopping a leak over 1000 liters is in a construction site. And the numbers are good. And so, an important thing to say about this whole technology overall, this is not an absolute, it is statistical in nature. But we are looking of way upwards of 95% of blocking, even on complex construction sites, of blocking these leaks at what is very sustainable levels of water flowing into the building. It will never be zero, by the way, and we will not catch everything. So it's an important observation. We are not 100% guaranteed, but if you look at the statistics, they are extremely good.
[00:17:54] Speaker A: Okay, that's really interesting. And Yarn, one other question comes to mind because I do work a lot on a computer and therefore I'm very curious. With Generative AI, they talk about, obviously all of its benefits, but they also talk about some of the experts that it is prone to lying from time to time. Is that the case with Anatomy AI? And I don't think I've said that very well, but you know what I mean.
[00:18:19] Speaker C: Yeah. I mean, the term they like to use for generative AI is hallucinations.
And I've seen crazy situations where they ask a question about a person. Tell me about that person. All of a sudden the thing makes up a story about him being a criminal and crazy things.
And so it's not relevant in our world. In our world, the world of anomaly detection. This is really not even a possibility. It's a world where you try to identify the anomaly and you do so with a success level. But at the end of the day, if you think of it, there's no room for a hallucination. At the end of the day, the system will tell you I'm seeing a leak or I'm not seeing a leak and that's it. It won't come up with an elaborate story around that.
[00:19:10] Speaker A: So the reason for the hallucinations, because you're a technological person, obviously is overload of too much information to decipher too many different types of information. Why does that happen?
[00:19:22] Speaker C: It really happens because what these large language models are designed to do is build excellent sentences based on a huge pile of examples they've seen. They are not designed to actually provide correct facts. That is not a design goal. Now they're putting that in, but originally it's not a design goal. The design goal is just to speak eloquently. So you ask it to speak eloquently about a topic and it will. It did not check the facts.
So you have to be very careful. If you use chat GPT to check for facts, make sure to double and triple check. I've asked it questions and I asked it to give me a link, a web link to a reference. And the web link always comes broken. And it took me a while to understand that. It's not really giving me a web link. It's creating the language equivalent of what seems to be a good web link. But at the end of the day, there's not a web page behind it. So that's an example, but it really comes from what they were designed to do. Now, there's enormous work all over the world happening now to make this much more factual and they'll get there. But think of it, we're really in generation 1.0 of that technology.
[00:20:41] Speaker A: Sure. Okay. So in comparison to anomaly AI, it's just programmed to check for what the issues are.
[00:20:50] Speaker C: No, correct. I mean, this is a very different type of situation.
Take a look at the water flow and figure out does this look like the typical pattern that you'd see normally or does it look like anomaly?
[00:21:03] Speaker A: Yes.
[00:21:04] Speaker C: No, and of course you need to do it well. What you don't want happening is too many false alerts or misses. Right, too many false alerts is annoying. Too many misses. Wow, I could let the building drown. And the risk is high. So that's the challenge here.
[00:21:21] Speaker A: So the challenge with yours is basically getting the balance right for finding what they need to find.
[00:21:29] Speaker C: Exactly.
[00:21:30] Speaker A: Fantastic. So I don't know what kind of names you're allowed to mention and not mention, but let's go and talk about some real world examples. Success stories about how wind has helped.
[00:21:43] Speaker C: Absolutely. So I mentioned Empire State Building. We helped them in multiple types of situations. One was identify a very large wasteful leak that resulted in around $100,000 a year of water waste and the associated carbon emissions. We also caught at a very early stage a pump that imploded in one of the high floors. Can you think of a pump imploding on the main riser bringing in water into the Empire State Building? Think of that imploding. I mean, the damage could be phenomenal. We caught it when the pump was just starting to leak, dripping. We caught it. We alerted the team there, went and fixed it. So that's an example. We work with companies like Mace. Mace, for example, projects all over London. They have essentially standardized with Wind. They've seen a significant reduction in their pain related to water since we joined. In fact, I'm not aware of any water damage incident on any Mace project for the past few years. Ever since they started using, they've done I think they should be quite happy with what they've done. There places like Canary Wharf, South Bank, right behind the London Eye.
High end apartments. They're using us, saved water, prevented damage quite a few years. At this point, as the list goes on, we have hundreds and hundreds of enterprise customers. We have 10,000 systems deployed, anything from multitenant apartment buildings to commercial buildings to construction sites.
[00:23:26] Speaker A: So, yeah, you've mentioned Mace. Obviously, they have a good reputation of being leaders when it comes to sustainability. So I'm not that surprised that they're one of your clients. Obviously, from the insurance perspective, is there recognition already at this stage by insurance companies about wind?
[00:23:45] Speaker C: Absolutely, there's pretty significant recognition. I can say some insurance companies are mandating Wind for high risk projects. Others are providing incentives in the form of premium reductions or reduced deductibles. And one of the amazing things is, because water has become such a big deal, insurance companies have been raising deductibles to pretty scary levels. It could be a quarter million to a half a million pounds per incident, which means that in many cases, you're simply not insured against water damage.
And so when wind is deployed, quite a few insurance companies will provide a reduction in both premium and reduced deductibles. We also have partnerships, so we have a partnership with HSB Munichry. HSB is a daughter company of Munichry, one of the world's largest insurers. And they have basically come out with a warranty program. So they will warrant that went, operates and detects leaks. And if we miss, they will pay. And the warranty is not too expensive and it covers those half a million or a quarter million, whichever is relevant in your deductible plan. So to us, that was a very big event because first of all, we can provide this benefit to our customers. And also it shows that a very large insurer who is very capable and they've done their technical due diligence, has essentially agreed to put insurance money behind Winst technology and algorithms. So it's kind of a sign of what people think of us.
[00:25:24] Speaker A: Yeah, that's really excellent. Now, obviously, that's been a partnership that you're delighted about? And why wouldn't you be? Are you involved in any other collaborations, particularly when it comes to advancing sustainability, which is one of our kind of pet themes here. On constructive voices.
[00:25:43] Speaker C: Absolutely.
If you look at some of the annual sustainability reports coming from some of our primary customers, water has become a big deal. So if you look at HP Empire State Realty Trust, which is the parent company of Empire State Building, Large pharmaceuticals, many of these organizations are now making water sustainability a top priority. So carbon is always number one, water is typically number two. And you can see that in many of these organizations report. So water sustainability and the benefits and the savings that we provide them are becoming quite a significant thing. I think generally the world has been focused over the past few years and probably will be for the next many years on carbon prevention because that's dealing with the root cause of global warming. But there's also the adaptation aspect. So global warming is happening. We need to reduce it, but it will happen. I think that's kind of common wisdom these days. And so how do we adapt when one of the major impacts of global warming is water stress? That is becoming a major area of focus and many of our customers are working on that.
[00:26:56] Speaker A: Where do you see Winch, let's say five years from now?
[00:26:59] Speaker C: Given the importance of managing water properly in buildings, I believe that it will become a standard. If you think 100 years ago there were not elevators and there were not sprinkler systems in buildings, now you wouldn't even give it a second thought of should I install an elevator or a sprinkler system? And I think water management in buildings is going down that path. We're already seeing that. So while most of our current customers have retrofitted wind into the building, we are already working with quite a few customers who have that forward looking thinking and they are putting water management equipment into the design of the building. And I think in five to ten years it will become an industry standard. And you would be frowned upon if you're not installing water management in the building, because why would you not? I mean, the actual cost of the equipment is not high a valve in a meter per floor. It's nothing relative to the construction project costs. And it makes so much sense to make the building more water sustainable and more waterproof.
[00:28:06] Speaker A: This is constructive voices.
[00:28:08] Speaker B: Another fascinating conversation on constructive voices. Jackie De Burker speaking with Yaron Dysian from Wint. And I'm sure we're all well aware of the damage that leaking water can do. And just great to hear about the possibilities of AI helping to solve a major problem within the construction and building management industry. And of course, you can find out more about Yarin and Wynt by going on to our website constructive Voices.com. Don't forget the dash and check information in the show notes too.
And that's all for this episode of Constructive Voices. Please take a moment to share it with others who may find it interesting. Follow or subscribe to get the latest episodes automatically on your favorite podcast app and rate and review the podcast if you can. You can also listen to the latest episode by saying alexa, play Constructive Voices.
[00:29:02] Speaker A: Podcast here's Constructive Voices here's the latest.
[00:29:05] Speaker B: Episode and on our website where there's lots more information too. That's constructive dashvoices.com. Don't forget the dash. Until next time. Thanks for listening. You're really helping us build something.
