Advanced Modeling Prevents Flash Flood Surprises

Pete Miller [00:36]: Welcome to the Predict and Prevent podcast. Flooding is a major problem worldwide, damaging homes and vehicles, claiming lies, and disrupting businesses and communities. While river and coastal flooding risks are mapped and often covered by government warning systems, surface flooding or flash floods caused by storms and rain often catch us by surprised and unprepared. But what if we could better predict when and where flash flooding is imminent? In this episode of Predict and Prevent, we explore how technology is transforming flood forecasting and prevention. We learn how a cutting-edge flood prediction system uses sensors and predictive modeling to provide earlier, more targeted alerts. Giving advance warning can help businesses and communities avoid the worst impacts of surface water and flash flooding.

Our guest is Avi Baruch, the cofounder of Previsico, a leading flood forecasting solution provider that started in the UK and has expanded to the US. Avi holds a PhD from Loughborough University, where he conducted research on advanced flood warning systems. Along with his academic supervisor, Avi spun out the technology he developed into Previsico with the mission of minimizing the global impact of flooding. Well, Avi, thank you for joining us. I’m really very grateful for your time and certainly for your expertise. Tell us a little bit about your background and how you came to co found Previsico.

Avi Baruch [2:10] It’s a story that actually began well before I even got into research. So I joined a university in the UK called Loughborough University as a PhD researcher back in 2015. And I joined a project on flood forecasting that had been running for about 20 years. So Loughborough University has an academic team that really focused on flood forecasting and flood modeling,

And I did my PhD at the university working on an adjacent project to a flood forecasting solution that was really focused on capturing pluvial flash flooding and complementing existing government warning systems. And it was a very successful research project that was piloted with the UK government. And it was so successful that there was a lot of kind of calls for it to be commercialized and spun out. And the research project was being led by my PhD supervisor at the time, Professor Yu.

And so having finished my PhD, I had done some further research on the project. I felt that it was a really important next step was actually to try and commercialize it. A lot of research these days stays on the shelf and doesn’t get used. And there’s so much coming out of our universities across the globe that could really make a huge difference to people’s lives. And so I’ve been really keen to make sure that actually the output of the technology gets used. and so convinced him to form a company together and given that both of us were academics we also then brought a CEO in, a person called Jonathan Jackson who’s a serial entrepreneur, and together we built the company up really around a mission to minimize the impact of flooding globally and we do that by providing an early warning to help people take preventative action.

Pete Miller [04:06]: So just as a level set, there’s different types of flooding, right? And you’re involved in one of them. Can you just describe those different types just so that everybody’s kind of on the same page?

Avi Baruch [04:17]: Yes, of course. So there’s three main different types of flooding. There’s rivers. So I think we all are familiar with when river level rises, over tops. Likewise, coastal flooding, particularly across the US, a huge problem, storm surges. But actually the largest flood risk, and in most countries it is estimated to be far more than rivers and the sea combined, is surface water or pluvial flooding or stormwater. It’s called different things in different places, but ultimately when heavy rainfall overwhelms local drainage.

And the challenge, if the Mississippi is going to overtop, you’ll know about it. There’ll be really effective government warnings. There’ll be advice. There’s systems in place. But when heavy rainfall is overwhelming your local drain outside your home, that’s when actually that can catch you off guard and can cause all sorts of problems. And so that’s the type of flooding, which is the largest flood risk, and it’s also the flood risk that catches people off guard most of the time. And that’s where we’ve really made sure always complementing the wider government warning systems that are already in place by them capturing those localized floods.

Pete Miller [05:29]: So it’s that flood I’m sent out in my raincoat with my shovel to clear the drain. I’m very thrilled to do that. Exactly that kind of stuff, right? So can you explain your solution, and you mentioned earlier, you have a flood warning system. How does that differ from other flood warning systems?

Avi Baruch [05:49]: So the flood warning systems most of us will become familiar with is from the government. So the National Weather Service provides weather alerts and also flood warnings if, again, if large river systems or coastal storm surges are occurring, there’ll be a warning out that these areas are at elevated risk. And so those are really based on gauges on the ground and engineering teams who might be out there in the field who know what’s going on.

And those are run by governments and are really effective and have proven typically the return investment of those sorts of systems are about four to five hundred to one based on the dollar in dollar out. So they’re really effective. Now what we do is those are also available to the government, available to businesses, but what we do then is we complement those by again capturing those flash floods. The floods that do really kind of catch people off garden that pluvial and small streams and surface water.

And so what the way we do that is using a combination of different technologies, the main ones being live hydrodynamic modeling and also onsite sensors. So we’ve got our own IoT devices where we’re monitoring what’s happening on the ground with our own devices, as well as other existing devices. And then we’re running continuous simulations to be able to forecast exactly where in the current conditions where we expect flooding to occur.

Pete Miller [07:17]: So I noticed on your website, one of your testimonials is from a small town, I guess, Whalley, So obviously they’re a customer. What is your typical customer? Who is your target customer?

Avi Baruch [07:31]: So Whalley is quite a unique one where it’s again, it’s a small community. Our average customer tends to be kind of larger infrastructure companies. So we work with the likes of National Grid, United Utilities, Network Rail, large infrastructure companies, BV, Camp Water Utilities, Power Energy. But the thing we noticed straight from day one, everyone floods and everyone wants to have better warning for floods. And so while our kind of our sweet spot has always been kind of the big national infrastructure organizations, actually we also work a lot with real estate and housing. So our largest customer segment is actually social housing and then we also work with local governments. So that’s also a huge sector for us. But then…other sectors would also be retail and real estate and highways. So wherever flooding occurs people often we get a lot of inbound interest from all sorts of different parts.

Pete Miller [08:37]: Great. So when we think about early warning, and our podcast here is about predict and prevent. So how early are you able to give warnings of potential flooding? What kind of timeframe?

Avi Baruch [08:52]: So our standard solution is really to provide up to 48 hours, so two days flood forecast ahead of time, that’s using rainfall and modeling. But what we are also able to do now is also provide a lot of really useful insights based on our on-site instrumentation, so our sensors, to be able to indicate actually is this location now actually at an increased risk than it normally would be. But our standard solution provides up to 48 hours warning ahead of time, which is, I think, what most people really need to start preparing.

Pete Miller [09:26]: So I imagine it’s a series of overlays, right? It’s like this sort of piece of real estate, if you will, this location is more prone to flooding and that’s one sort of view on it. And then you have other views on top of it, like the weather and stuff like that. Is that kind of the way that works?

Avi Baruch [09:45]: So there’s a little bit about the history, but the challenge with all history is, again, we are living in a very changeable world and what used to flood actually might not always be what’s now going to flood in the future. So just because your property might not have been flooded in the past, actually, it still means it might be at quite high risk. So we’re running continuous models ahead of time using event-specific modeling. So you can have situations where actually less rainfall might cause more flooding, purely based on where it’s falling.

So all of our models run event specific ahead of time. Now, what that means is actually, I mentioned that the two-day lead time is kind of what the forecasting is provided up to that level, but actually a lot of the time, too, you will just have really short, sharp cloud bursts. And these can cause flooding that actually can only be predicted three or so hours ahead of time. So it’s really very much weather dependent, particularly on how much lead time we’re able to provide.

Pete Miller [10:44] The recent flooding in Dubai, was that surface flooding of the sort that you would have been able to predict? Or is that it… I don’t know what the nature of that flooding was.

Avi Baruch [10:56]: So yeah, that was intense rainfall, completely overwhelming the local drainage systems and causing streams and rivers. So that is our sweet spot, those sorts of floods. So that’s where the government may have known that there’s heavy rainfall forecast, but they might not have known exactly where it would flood. Using modelling, you can then predict exactly which specific locations are at immediate risk. And then using sensors, too, you can then understand exactly what’s happening, where and at what time.

Pete Miller [11:22]: So that, so that there would bet there would have, if you were to have contracted with the Dubai government, you would have installed a series of sensors and, and done some real time mapping, right?

Avi Baruch [11:35]: Yes, we were not there at the time. So we’re not currently live in Dubai, but I’m sure we’ll be speaking to them at some point.

Pete Miller [11:45]: Yeah, yeah, it was just it’s just been in the news. So I thought I wonder if that’s if that’s that would have been a good application. Yeah. So this is very fascinating to me because obviously flood as you said is a huge risk and the damage from flood monetarily is very significant. As you look out a few years, how do you see your company expanding and what kind of enhancements do you see to the solutions that you have today?

Avi Baruch [12:16]: So we started off in the UK and that’s still where a large number of our customers are based, but last year we launched in the US and have grown quite quickly. So we’re now kind of live across 11 states and we’ll be expanding across the rest of the US. And…

I mentioned earlier a few of our target customers, some of those core segments. We work very closely in partnership with a large number of different insurers. And the reason we do that is to support their customers. So especially if, so insurers, we work with the likes of Zurich and Liberty. So when their customers are immediately at risk, we work with them to support them ahead of time with early warning and practical insight on where the impact might occur.

And that then causes a win-win because the customer prevents loss and sure has the lower overall risk. So everyone’s a winner there.

Where we are growing very much is on both geographically, but then also working, especially in bed income, in the wider insurance proposition to customers to ensure that they’ve got all the tools they need to prevent the loss before it ever occurs. And in terms of the peril itself, there’s lots of different drivers of flooding and what is already there. And there’s certainly quite a few other areas which still catch people off guard and those are areas that we’re growing into.

Pete Miller [13:46]: Let me just dig into how you work with insurers. Um, like in terms of integration, are there different tools or approaches that you use, um, to, to integrate? Can you just describe how that works? Um, especially like your warning systems, or your warning notifications within their processes?

Avi Baruch [14:08]: Of course, and it really does vary customer by customer. So the easiest way for a lot of insurers and then different insurers might have different levels of tech stacks available to support customers. So what we often do to get started is an insurer will see our service working for one of their customers and say, well, actually, you know,

One of my other customers might actually also really benefit from this. So let’s work on a project to support that other customer. And the great thing there is we work directly with the customer. The customer signs up. The customer uses the service and does the operation. So the insurer itself can actually keep quite… can be as hands on, as hands off as they want. They don’t need to be involved at all. But what we do, for example, with Liberty is actually they use our API. So all of our data, all of our insights are available via API.

So they then embed, they’ve got their platform called Risk Reduce, so Liberty Specialty markets, and they embed all of our insights into their risk reduce platform so that they can then deliver those to their customers. So that’s then kind of part of the wider risk management solution that the insurer provides. And what we’re seeing a lot of…right there are quite a few insurers really expanding their kind of risk management proposition that you know, their prevention proposition every insurer of a certain size has some form of you know, prevention proposition that they give to their customers and what we’re seeing is just huge demand to support those because it’s really what is starting to differentiate insurers the last few years, but it’s very much an accelerating trend.

Pete Miller [15:46]: Yeah, I mean, we very much believe that, right? It’s like, let’s try to prevent things from happening rather than trying to indemnify after the fact, right? Everybody’s happier. So what are some of the challenges? Most of the people that listen to this podcast will be in risk management and insurance. So as an organization working with folks in risk management and insurance, what are some of the challenges you see for insurers and their customers as we make the move that you mentioned to a more preventative market.

Avi Baruch [16:20]: So I think it’s a number of stakeholders that are involved, firstly in actually taking preventative action. So sometimes if you talk to a chief risk officer, they might not know exactly across their portfolio.

about too much about the day-to-day flood incidents. So you might have one factory go down for a day due to flooding that actually might not ever go into insurance claim and might never come be escalated. Likewise, kind of incident response team will be very localized. So some chief risk officers will be very close to those day-to-day operations, some of them won’t be. And so…

You’ve got maintenance teams, health and safety teams, ops teams really on the ground taking those preventive actions. Then you’ve got the risk management team and insurance buyers who then have their relationship with the insurer. Then you’ve got the insurer who then normally is being used to dealing with large claims. And ultimately.

What you then have is a huge opportunity from the insurer, first of all, to say, well, actually, right now, you have your total cost of risk. You have the incidents that are occurring on the ground that never go to the insurance side. You’ve got the premium, you’ve got all of these other, you’ve got the deductible, you’ve got all these other costs. We want to grow our proposition to help you well beyond just your insured risk. And that’s what a lot of insurers are assigned to do. And…right now, especially where they’ve been at in the last few years, is actually not getting too involved with those kind of more day-to-day incidents.

And as you do so, you’re involving a wider range of stakeholders, especially kind of the onsite ground team. And this is where we’ve had the most success is actually getting the solution to the hands of people who actually then using it operationally. And sometimes that works really straightforward and really, uh, simply. Other times, you know, you do need to then talk to the organization, kind of learn a little bit more about what they’re, who’s responding unto and where.

The challenge is really around getting the solution from the high level kind of insurance risk management and then right down to the end user, the people who are then responding to floods on the ground. And maintenance teams are often the ones who are involved in that, but there’s also health and safety, there’s ESG and risk management and operations. And often those teams…do not normally engage that closely with the insurance buyer. And they should, because how well a risk is managed is all part of how well the insurance buyer then tells their story, presents their risk to the broker, and shows how insurable they are. And so the better you are at risk management, the lower your risk in theory should be, that doesn’t always happen because it’s all about how you tell your story and how you sell your own risk to your insurer.

And that’s certainly one thing we’ve noticed is such a huge opportunity for the insurer to get closer and for organization to those on site operations teams and for those teams to actually better align what they’re doing to what the insurer wants to see across organization. And to really support the risk the insurance buyer it because that’s certainly there’s just so much that you can do to showcase and to prove that you’ve actually reduced and better managing that risk.

And so it’s, as you said, it’s certainly a challenge, but it’s also a major opportunity because it allows interests to get way more embedded with their customers.

Pete Miller [20:03]: Yeah, that’s a great answer. And that’s, as you say, I think what I hear you saying, and I agree with, is it allows the insurer to kind of create more stickiness, if you will, with the insurance buyer, because you’re bringing more expertise, more people in a better relationship, right, across a broader spectrum of sort of risk management.

You know, a lot of the good people that work at insurance companies that maybe don’t always have customer facing roles, they can in this kind of thing, which should make that a more stable and long lasting relationship. That’s kind of what I’m hearing you say.

Avi Baruch [20:40]: Absolutely, and certainly on the insurance side, I think a lot of organizations look to the likes of FM Global as real leaders and really getting out there on site with the customer really integrating the risk management to the wider insurance proposition. And I think a lot of other insurers look to FM Global as an example of how you can really connect those in a far more effective way. And certainly, what we found is our technology is one of those examples of how you can really get closer to that on-site team to prevent the worst.

Pete Miller [21:11]: Tell us what your experience has been in terms of communication or storytelling both internally to an insurance company and externally with their customers.

Avi Baruch [21:25]: Yeah, sure. So as I mentioned, a lot of the end users will be people on the ground, site managers, operations teams, and what the work that they do on responding to incidents, be they floods, wind, hail, anything, a lot of what they do doesn’t always get them escalated back up.

And a lot of the time it’s because they don’t know how valuable their actions are actually across the rest of the organization. And what we see day in, day out is really effective risk management going on. And insurers really wanting to hear about that, but actually that message not always being passed up and down the chain, kind of what exactly insurers want to see and hear from their customers. And the customers realizing what they need to present back to the insurer. Certainly, as you said, risk and prevention is a huge part of the insurer proposition moving forward, but to make it effective, you need to see how well it’s being used and to do so, you need to get closer to the people who will be then using it. And so I think that there’s a lot of opportunity for better communication across the risk management chain on actually…

How is risk being managed? How could the risk be better managed? How does that make a difference to the wide insurance proposition? And certainly…

That is another win-win for both the insurer and the insured because the better they’re communicating their risk, the more likely that they’re able to build a better relationship with that insurer. So it’s a huge opportunity moving forward and I think any insurer that’s starting to work on prevention in their agenda  will be seeing that right now.

Pete Miller [23:17]: You know, it’s funny you say that because in the course of this podcast, we talked to several organizations that are doing this type of predict and prevent, and that’s, that has been something of a theme. And you know, another way to look at that is, you know, does the leadership of the insurer, they’re very busy doing great work, making a lot of people’s lives better and safer. But if they look at this, I think as, you know, it’s a mindset change, that’s actually an organizational culture change, that needs to be communicated and sort of, you know, led throughout the organization. Does that make sense to you? Does that resonate?

Avi Baruch [23:59]: Yeah, absolutely. And I also think, you know, it’s a hugely exciting and challenging role to be in the risk management team. Because the risk manager, especially chief risk officer, has so many things that they need to focus on, so many risks that they need to address. And some of them will get more stuck in than others, but certainly they need help. That’s certainly a common theme and they need help both from the insurer and they need help from their internal team.

Pete Miller [24:29] Avi, you mentioned that you recently made the move from the UK to the US. Can you just give us a little more detail? How did that look? Who are you working with and how extensive is that?

Avi Baruch [25:22]: As soon as we started the company back in 2019, we grew up in the UK, but we started processing data in the US. So as part of our solution, there’s a lot of data preparation and processing to make sure that our forecasting is as accurate as possible. And last year, we finally made the leap and launched in the US in partnership with a number of different insurers. So we’re working with both insurers and customers now in the US to reduce their overall risk.

And we’ve had some really good case studies from both New York and other states in the area, use it with both the sensors and the forecasting. And so we’re starting to see a lot of growth in that space, particularly across the East Coast.

Pete Miller [25:22]: I think your vision of how insurers and their clients will work going forward is exactly correct in my opinion. And certainly the technology that you’ve built is sort of a key enabler, I think, for organizations that utilize it in the flood area to kind of realize that vision for the insurer and for their customers. Yeah. So thank you very much.

Um, certainly appreciate your time. Um, this is fascinating to me. I have to say I went through your whole website and, uh, looked at the testimonials and they’re pretty compelling. I would definitely encourage folks to look at your website. Um, and certainly, you know, best of luck. I think you’ll have great success, certainly in the U S because it’s a big market and has a lot of need for this. So, and thank you for your time. I really appreciate it.

Avi Baruch [26:18]: Thank you.