Using AI to Transform Emergency Management

TGIF and welcome to November. Thanks for joining us on this week's episode of In Case You Missed It, where we speak with thought leaders around GovTech and identify innovations, market trends, and issues that we believe you need to know about. Joining me today is Leo Zeman, the CEO of Ladrus, a company using data to improve all aspects of emergency management from preparedness through response. Such a timely topic that we have the opportunity to talk through today, given the events that we've seen throughout the country recently. Well, many, you know, many of us are still recovering from those events, we thought it'd be an important discussion to welcome Leo to, to kind of share what he's seeing in and across the market. So Leo, thanks so much for taking time to join me today.

Thanks, Joe. It's great to be here. We appreciate it.

So before we get too far, maybe you could spend a little bit of time just letting our audience know a little bit about yourself, your background, and how you got to be in the role that you're in today. Absolutely.

Well, we're very fortunate. Let me start by saying very fortunate to be able to work on this problem. You know, myself and our founding team, management, the whole company thinks it's super important and one of the problems of our time, right? These disasters that are facing us. So my own story started in a little town, Nevada City, which is actually in California. Nevada City, Nevada County, State of California, Eastern California in the Sierras. And actually I grew up with a person who had become my co-founder when we went to Berkeley. So we went to school together at Berkeley and, you know, AI was, one of those things that a lot of people were looking at in 2017, 2019, around that time. And also around that same time, you had the fires in 2017 and '19 in, you know, Northern California. And it kind of put that issue, I think, on the national map for, you know, for one of the first times that it had been there. And it just sort of intersected. So, you know, we had very much a technological background but fell into this problem that was so timely and also impacted the community that We were from, and when we moved back after starting the company to our hometown, we were fortunate enough to be able to work with the local OES emergency services team in Nevada County, and that was the start of it. And turned out, you know, it made a lot of sense for a lot of places, and we took it from there.

That's great. You and I have the opportunity to both be from Northern California. We were in person together not too long ago, but your points are right. I remember thinking back to those fires, that timeframe, and, you know, coming out into my yard. And, you know, having originally grown up in Canada, you're used to snow, but it was very similar at that time except it was ash that was falling from the skies here in my part of Northern California. It was very, very eerie. Maybe you could give us a little bit of overview of that company. You kind of referenced Lattice, but maybe there's people in our audience who are not yet familiar. Maybe you can kind of give us that breakdown of what are the primary challenges that you're working with government today to solve.

Absolutely. So, you know, kind of as we mentioned, Ladrus, our company, works with emergency managers specifically to help them quantify and handle the risk that they face. So evacuations, wildfires, flooding, all hazard events. And really, you know, those are the folks on the front lines of the problem in most cities and counties. It's going to be the, you know, the fire department, the sheriff's office, PD and emergency management. And so when we entered into this space, you know, we kind of looked at what was available and said, you know, are there ways that we can better assess how we're performing today, right? What actions can we take to increase the feasibility of a mass evacuation or ensure that we're creating our fuel breaks appropriately for reducing the spread of wildfire? And for us, that started with quantifying the problem. which traditionally had been a challenge. You know, historically these disasters hadn't happened very often.

Mm-hmm.

So there wasn't necessarily as much of a need or an impetus to quantify just how bad will the evacuation be. Will it take 5 hours or 7 hours? Will the fire spread at this rate or that rate? And I think all of that changed again in kind of that 2017, 2019 range where these disasters are now happening often enough that we want to be able to quantify our performance very quickly. and then understand what actions can we take that are the most impactful that can reduce that risk. And so that's the nice thing about artificial intelligence is it does that, it can do that for you very quickly. And then delivered through software allows you to embed that into your plans, into your response tactics, into your recovery, and ultimately your mitigation and the full cycle of emergency management. So that's basically what we do. We provide software that gives those answers and guidance to emergency managers in a decision support capacity. for all the hazards that they face.

You know, data's always been there. It's always been something that these organizations have relied on. But I think you're spot on that the introduction of AI has rapidly and drastically changed the landscape. And I think you're right in the center of 2 issues that are, you know, gonna dominate, I think, our livelihoods, right? AI and climate. But let's get into some real-world examples here. You kind of gave us the problems that you're solving, but maybe you can share with us as much as you can, you know, some on-the-boots, you know, specific examples of where your technology is being used today and what impact is it having? Not just reactionary, but sometimes before there's ever an event.

Absolutely. So imagine an evacuation scenario, if you will. So this is a scenario that, you know, every county, every city prepares for, for different kinds of disasters that they have. And in the preparation phase, plans are a big part of that process. So for example, your hazard mitigation plan generally, which is a document by the way that is useful during and after a disaster for getting reimbursement or funding from FEMA to assess and address the problems that you have, that relies on understanding how your evacuation might progress. And that requires a plan to be written for that. Your community wildfire protection plan. Which basically defines the strategic priorities in response to wildfire and protecting communities that are at risk for wildfire, that requires an evacuation plan. And so in both of those cases, the fundamental element of those plans ought to be how long does it take to evacuate? If we called the order today and everyone got in their cars and left, do we know how long it would take and sort of what are the big variables that would make it take You know, a time that's feasible or a time that isn't because the fires move very quickly and we got to make sure we evacuate everybody before the fire reaches those communities. And historically, again, those didn't happen very often. These events were very spread out and now we see this happening every year, unfortunately. And so it's important that we can really precisely quantify how effective our plans are going to be. And so emergency managers will take their hazard mitigation plan or their wildfire plan. or their evacuation plan, and they will use our artificial intelligence models and our software to essentially inform the numbers that drive those plans and the strategies that they write into those plans, and then practice with those plans to deploy them during an event.

So let's continue with this a bit. So they do all of this, they walk through the plans, it's data-driven, data-informed. Then we get to the unfortunate scenario where we're in the middle of a disaster and the government is, is responding, they're dealing with the chaos of that moment. How also can your solution and the insights that you're bringing to table help them deal with the chaos of that moment?

So for example, one of the things that our software provides is this kind of bird's-eye view into the evolving incident. So for example, let's say that you have a community up here near where I'm from that's, you know, evacuating in the mountains, and they can see that You know, the evacuation has started, the what we call the emergency operations center has been activated. This is going to be a multi-day event, and we see the fires, you know, over here in one area, and we've evacuated some people, but we see it spreading. And the question on the plate of the emergency manager now is, when do I call the next evacuation order, and who do I evacuate, and where do I send them to? I mean, these are big questions, right? Really really impactful questions that impact thousands of people and tens of millions of dollars in property value. And with LADRS, they can see, all right, we have a congestion, a choke point forming with the traffic on this road, and the simulations are showing us that if we run an evac— if we call an evacuation right now, it's going to take us 5 hours and 30 minutes. But if we phase that evacuation out, it could take as low as 2 hours. So we're going to make the call earlier to start phasing that evacuation out earlier and not run into a really long prolonged evacuation. Our models are showing by the time we evacuate, the fire may be burning over those roads, over those houses. And so it's quite literally can be the difference between successful and unsuccessful responses, just because we get that data coming in in near real time to seeing how the situation is evolving.

It's been roughly a month or so since we've seen Hurricane Helene and Milton, make an impact of the communities across the country. How are you seeing technology, technology like yours or yours being used after those events as well? Is it being used to speed up the process of recovery? How are we also leveraging those same insights to, you know, to maybe get people back into their homes or get life back to normal as much as possible?

Absolutely. So the way we're seeing that happen right now is it's a function of understanding, again, to your point about data, What does the data look like before an event so that we know what we're trying to get back to, right? And being able to track that data, which is already a key part of the workflows, you know, related to the recovery process and related to FEMA, is being able to understand, you know, what assets are likely to be impacted by a risk event or by a disaster so that we know what our target for restoration is and then tracking our progress toward that target. So again, the theme of quantification, putting numbers to it, immediately comes right back, you know, into center stage. If we know here's how many people were in this area, you know, here's how we had traffic flow, here's how we had the assets arranged before the disaster hit, and now we know, okay, here's what the disaster took out, and we can look at our data from before and say, how close are we to getting it, you know, back to baseline or whatever we determine the new baseline ought to be. And that's, you know, only really possible if you have a system that can look at all that data at once. And again, that's what machine learning is so good for, is it can look at— I kind of compare it to, you know, a map, right? So if you're an emergency manager, you look at the map and you see, hey, I have a flood here, I have a power outage here, they're really close to each other, there might be a correlation there. And the nice thing is that now we have systems like the ones that Ladrus provides that can look at, you know, millions of these layers at a time and tell you Yeah, here's the correlation between those 2 points of data that are coming in, and here's what we suggest you might do to remedy the underlying causes as quickly as possible. And here's the people and the property that would be impacted if you did that. So it's really a full-cycle thing. I mean, we wanna keep communities safe and keep them operating well, and part of that is understanding where we go back to during the recovery phase.

You're talking about resiliency, you're talking about safety, you've kind of hit on throughout the conversation so far the criticality of both the data but also the events that we're dealing with. And obviously, like this whole discussion on AI and data-driven decision-making can't, you know, we'd be foolish to also not talk about cybersecurity and data privacy. What steps are you taking to ensure that the insights that you're providing are, you know, are relevant, that can be depended on, that are secure for these government agencies that you're working with?

It's a great question. And I think it hits on perhaps the single most important aspect of what we do, which is that Ladrus, you know, has this goal of being the models that emergency managers trust. Trust is key to everything that we do, especially in risk management, right? you've gotta be able to trust the data and the models that you're looking at. So there's a couple ways that we do that. So first off, there's the, you know, there's the standard sort of, I'd say industry standard best practices, you know, secure GovCloud hosting, you know, ensuring that you have compliance with security certifications like SOC 2 and NIST and others and making sure those are on your roadmap. And certainly those are things that, you know, either we're in process of doing or on our roadmap. And in some cases we do actively. But then there's another level to it, especially when it comes to predictive modeling, that we take a close look at. So on the privacy end, we don't use any data that isn't fully anonymized, right? That's fundamental. You know, that's, it's just, that's not our business. Everything's anonymized, and there's actually a lot you can do with data that's really anonymized because we just want to make sure that everyone gets out safely, that everyone's protected. When it comes to actually making sure that the data is validated, and in particular that the modeling is validated, we maintain an advisory board of luminaries in the space that we work with to inform our modeling, to inform our data. And we also do a lot of research. So we have a team at Ladrus under our CTO that is focused on, you know, alliances and partnerships with research labs and universities that that do work in this space independently, and we are able to then cross-check our data and our models against their results and make sure that things are in the same ballpark and generally going in the right direction. And then the third thing that we do, in addition to having independent domain expertise applied, is we work with our customers because ultimately nobody knows the area better than the emergency manager at the place where they live and work. And in many cases, you know, they're local, they've been there for years, which is great. And they understand these issues, you know, on a level that even the data doesn't always capture. And so when we deploy for an emergency management agency, one of the first things that we do is we go through both a training process and a data validation process. And we check to make sure, you know, do we have the data that we expect? Are we seeing the kind of results that we might expect to see under baseline conditions? And what did the surprises tell us? And from there, we're able to work with them to you know, address any gaps in that data or calibrate the models differently to their particular region. So there's no one-size-fits-all solution, but, you know, generally you want to put trust first, and that's honestly just a massive part of what we do at every level.

Well, given those connections, the engagement points with your customers and prospects and talking to those luminaries, I mean, I imagine you've got a tremendous pulse of what's happening today in the emergency So I'm gonna ask you, if you had a crystal ball and you could kind of, you know, look at the future of where is it going, where do you envision emergency management and technology going? And what areas do you see that are evolving right now that might be either have the biggest gaps for unmet needs or things that are capturing your interest right now as we look at the future?

Great question. So let me start by saying we're very fortunate to work with our customers that we have today, and a number of them are also innovation partners, I think, even beyond the agreements with them and the work that we do. And I will draw, I think, on our experience in working with them in answering this question here. So one of the great things about this space is that we really do see those agencies leading the charge. And in particular, one of the things that we see is There's a great demand and a great need, I think, from the emergency management community to be able to prioritize and respond to the right types of incidents and events. So, for example, perhaps we get notice that there's, you know, potential flooding that we think is going to occur, and we say, all right, we have to evacuate this area. But it turns out that was a false alarm, and now when the flooding actually starts, 5 hours, 6 hours later, we don't have the resources we need to do that 'cause we already deployed them, you know, elsewhere. And so I think one of the biggest and most impactful changes that I, you know, and personally we see happening at the company and in the space in general is the ability to prioritize, sort of separate the, you know, false positives and the false negatives and go after the incidents and the events that will be the most impactful. So the signal intelligence, situational awareness, and we see next, Gen situational intelligence tools being built. You know, we're working on some of our own, I have to say, and we think that that's going to really drive the efficient allocation of resources for planning and ultimately for response, right? Because if you're able to plan more efficiently and get more advanced notice earlier, then you can make those decisions in a more informed way in the moment, and ultimately you have more You have more leverage, right? Because the earlier you know about a problem, the more lead time you have on responding to it. So really pushing the boundaries there and delivering technologies that will do that for emergency managers, that I think is perhaps the space that is primed for the most innovation that we see today.

So I'm sure it comes as no surprise, but, you know, we love individuals like yourselves, organizations like yours that, you know, are drawn to solving problems in the public space that are are driven by a mission, right? And I, I'd ask you as a question now, as you look at yourself in your role, what lessons have you learned from entering the GovTech space, uh, and what advice would you have to other startups or to the next Leo that's thinking about entering the GovTech space? What words of wisdom or advice would you have for, for that, uh, that individual?

I think you should focus on what's possible and not what's been done before. There's— we're really in a new age right now. I mean, these are 2 of the biggest problems and opportunities of our time. 2 things are happening, and you pointed this out at the beginning. One, you have information being generated, you know, at an exponentially faster and cheaper pace than it's ever been generated before, right? So AI is, as some fine folks have said, has lowered the marginal cost of creation of that information. The other thing that you have happening is these disasters are happening more frequently and more often than they've ever had before. And so this changes the rules, right? Things that we may have done before and approaches we may have taken to everything from how these systems are built to how they're procured to how they're deployed, all of that is really ripe for change because the problems that we face require that these systems evolve for us to competently solve them. And I think it's not really a nice-to-have or something that's optional. It's It's very much like this affects our life, it affects our property, and it will affect the generations, you know, to come. And it's our responsibility, and, you know, quite frankly, my generation's responsibility to solve that problem. And that's what we're doing. And we should all be ready to go keep doing that, I think, is the mission that we go after every day.

Well, I love that. So let's say that someone in our audience is like, heck yeah, I want to be, you know, I want to do just that. Where do they go to connect with you, and where can the audience go to learn more about your company?

Well, you can visit our website at www.ladrus.com. You can find me on LinkedIn. It'll be my first and last name on LinkedIn, which— and as well as, you know, our email and our phone information is on the website and our LinkedIn. And, you know, I encourage anyone who's interested in learning more or working with us in any capacity to reach out.

And—

or if it's just even to share the other work that you're doing, we're always interested in learning more It's going to take a lot of collaboration to solve these problems. So yeah, our website or LinkedIn would be a great place to start.

Awesome. Leo, thanks so much for joining me today and taking time to, to share your insights with the In Case You Missed It in Government Technology audience.

Thanks so much, Joe. I appreciate your time. Thank you.

Have a wonderful weekend, everyone.