Russ Fordyce 🏆: Welcome back to the winner's circle, everybody. I'm here today with Zach from Rapid SOS. These are the guys that have the network that all of our first responders use to talk about emergencies, get us help and everything else. And they are the newest 2026 Stratus Award winner. So Zach, welcome to the show and congratulations on the award.
Zach LaValley: Thanks, Russ. great to be recognized. really mission driven company. So it's it's nice to to to feel that recognition.
Russ Fordyce 🏆: Yeah, this is one of those companies that you never hear about that's fundamental to pretty much most of America, it seems like, from from the numbers that you guys tout of customers. you guys are that underlying network that all our first responders use, fire police, emergencies, government schools. So that they're off kind of the public internet and they've got their own channels, right? Why don't you give us kind of the elevator pitch for rapid SOS?
Zach LaValley: Sure. So we're the world's largest AI-powered safety network. So we connect 723 million different devices, 250 different enterprises across North America largely, and about 23,000 responder agencies, right? And they take a lot of different different shapes and forms. And the way that I would think about it is traditionally 911 emerged as a way of calling for help. and that was excellent and frankly revolutionary when it first came out, and we all had land lines in our houses, and the landlines never moved, and so you knew exactly where someone was standing when they were talking to you on the phone. The world has evolved tremendously since then, but the underlying technology and infrastructure lagged. and so really we were still very reliant on for a very long time the phone lines being it. And trying to triangulate us just off of where we were. And so we moved to cell towers, and then we're triangulating off cell towers. But now it's so much more than phone calls. So, you know, your car knows when it gets into a crash. your phone knows when it gets to a crash, your smartwatch does. We're surrounded by an ecosystem of technology that is aware and able to help us in our times of need. and so it's more than just the call. It is the call, but it's more. And so we help connect those devices into first 911 and then out to the field as as an emergency emerges. And so your phone provides, you know, your location when you call 911. But if you set up things correctly, it can also provide your health information. And only in the case where your emergency is health related, but would we connect that in? But we're the conduit to make sure that it gets to the people who can help you. and so that's what that's Yeah. Yeah. Yeah.
Russ Fordyce 🏆: Yeah, it's like almost like the connective tissue, right? It seems like. Yeah. So you're you're playing this kind of role in the middle of of kind of the responders, the equipment, the communications. And you talk about yourselves as infrastructure. Walk us through kind of what a deployment normally looks like. Is it software, hardware, a mix of both solutions? What what do you put out in in into
Zach LaValley: some of both.
Russ Fordyce 🏆: a
Zach LaValley: Yeah, some of both. largely the network is internet based. and so we we play in various levels of the internet. So we are available over the commercial internet, because sometimes the the private internets that the nine one one system and responders also connect into, sometimes those go down. So we act as a redundancy backup between those those various levels of the internet for 911. We also provide some hardware in certain cases in order to help digitize and make more available pieces of information. So pieces of hardware that might go into the 911 in order to access and understand what is happening inside the phone call. Radio systems that might live alongside of the radio system between the first responders in 911 So that we can understand what's happening in communication streams between them. So again, so that we can foster better communication and more overall awareness within that community in order to to facilitate again faster, better outcomes for the public.
Russ Fordyce 🏆: Yeah, it it's it's really amazing at how far this system has come to me. I I was in telecom for a lot of years and supported nine one one, you know, through the telecom channels. And as it did, as as I P came into that equation, as we did BOIP, you had to move to kinda E nine one one, which is like kind of self reported, which was horrible. You had to get employees to like sign up to tell you where they were and if they were on you know, we sold a soft phone and you know, that ran on an Android device. How do you do nine one one on that? You constantly, you know, updating an address. It didn't work. And then yeah, and then GPS came in. It seems like this infrastructure, a lot of it is aging. Why do you think it's taken so long for this kind of aging infrastructure and technology to to get upgraded and it become of age?
Zach LaValley: Yeah, there's probably lots of factors. change is hard, right? And so it's also so that's I think the the driving one. there's a real pro and a con to the way that we, at least here in in the US and largely in North America, have designed our nine one one system. it's very decentralized. So there is no national nine one one. It it's county by county. Runs their own 911 centers and they have their own connections out to the field responders that surround their community. And so you kind of have to make this change 6,000 times across the United States. And not everybody's in the same camp, and not everybody's on the same tech stack. And so it is just sort of this evolutionary game that has to be played across this industry.
Russ Fordyce 🏆: So w walk us through you t you were talking about the connected car, everything's connected now. I mean, we've got sensors and watches and bands. They're coming into your earbuds next to your ear, you know, your earphones. I saw that. We have a a customer or a a nominee and winner who's who designs hearing aids and and all that technology is going in there now too. Walk us through kind of what happens when they're not a rapid SOS customer and where kind of the data drops versus kind of what you guys do and where you come in and fill in the parts.
Zach LaValley: Yeah. So for those companies that aren't a part of our network, most of the times that information never makes it into 911 or emergency services. The only way that it would is if a human being read off of the app that maybe you are using and said those words to someone in 911, or and then hope, fingers crossed, that information gets out to the first responders. Because when you call 911, you're not talking to a first responder, like the field responders. You're talking to a 911 telecommunicator or a dispatcher. For some connected devices, you might go to a monitoring center, right? So these are private entities. Your home security system is very is basically designed this way. home security system fires off, it goes to a monitoring center.
Russ Fordyce 🏆: us. So
Zach LaValley: Perfect. Okay. So some of these connected devices are connected to to monitoring centers. and so they are, you know, basically a human-based filter before we get 911. The a lot of these signals are very noisy, right? And you get false alarms all the time.
Russ Fordyce 🏆: Yeah.
Zach LaValley: And so the traditional way, again, before we had AI and before we had a lot of these sort of smarter techniques, was another set of humans would filter those calls. and the way it would work is they might call you back as the homeowner. verify with you if it is real, and and again, 99% of the time it's not real, but if it is real, they would pick up the phone and they would call 911 and then they would verbally relay all the information that they just said to you into 911. now let's talk about if a if a customer or an enterprise is part of our network, all of that digital fingerprinting and is automatically into the network and gets funneled to the right people. So when you call 911 or your smartwatch gets into a crash, it will automatically send its location and any relevant information up through into our network, and we will automatically route it to the 911 center that is responsible for handling that action. And then as they dispatch resources, the relevant information, like location, maybe like a health situation you're having, will get sent out to an app on the responders in their pocket. In the case of monitoring, which still happens, all of that literal game of telephone goes away. So yes, the monitoring agency might still provide you with fantastic customer service, call you back, make sure that you feel okay. Most of the times it's still nothing. But in the cases where it is something, no longer do we have to play that game of phone relay. All of the information that was gathered is now automatically sent into the systems of record. inside of the nine one one agencies and then if again, if appropriate, out to the field responders if it's valuable for them in response.
Russ Fordyce 🏆: That really changes the day the way that I mean, I'm just thinking through like an ambulance, right? You know, if you've fallen and broken something, you know, it really changes the way care is going to be provided, I would imagine, if if the ambulance is arriving knowing, you know, hey, you fell twelve minutes ago and your heart rate is this and your blood ox is this. I i it can really determine kind of life or death things, I would imagine. I don't mean to be too, you know, blunt about it, but I I Yeah it's it seems remarkable we didn't have this layer before.
Zach LaValley: I I think most good things feel obvious in retrospect. and yeah, it just took some work to actually make all these connections. not only is it better responses, it's more appropriate responses, right? So knowing the speed with which you got into a car crash and the number of people inside of the car because you had seatbelt sensors totally determines how people will respond and the level of response needed. If you got into a two mile an hour bumper inside of you know a parking lot that's way different than 40 miles an hour and there's you know four other you know two kids in the back. and so now we can we can get all of that and especially in these cases where man maybe no one no one dialed the phone. I we've got a case out in Cass County, Michigan where we sent in a crash report and it was it showed off the road because we had the GPS. No one called in that person was unconscious. It was on a
Russ Fordyce 🏆: Yeah.
Zach LaValley: county road, but within 13 seconds, the dispatchers, dispatch people, we got there in four minutes and you know, pulled a guy out of a burning car. and so these are the these are the really positive examples of of how having sort of this interconnected safety network really can bring positive positive outcomes to us.
Russ Fordyce 🏆: Yeah, no, you guys chose this interesting approach too, of kind of layering you've you've brought this new layer of data and AI to the existing infrastructure. What what made you kind of decide on that approach versus going in with new equipment and new hardware to kind of do a replacement versus kind of an augmentation?
Zach LaValley: What we have found is that we wanted to resist ourselves. When you go in and you bring your own stuff, you create another silo. And really the problem we're trying to solve is we're trying to overcome all of these data silos. So all of these people who bring in their own hardware and their own networks, they're then incentivized to try to cordon off and isolate and cons a customer that like it is more secure to keep these things completely siloed. and we really think that you don't want to create an inner silo inside 911, but like almost nobody, almost no emergency requires just fire or just police, or like we are surrounded in a connected ecosystem. And so if you put them in silos, all of a sudden fire EMS and law enforcement all show up. Two of them are completely blind if one of them was part of the silo. And so That problem exists everywhere. And it's not just in this public, right? Like if you got into a car accident inside of a commercial parking spot, that retailer might have you know the help to get that you need, right? Or maybe not maybe maybe you're having a heart attack. They might have an AED inside of that building. You don't want to wait for the ambulance to show up if we can turn on that AED and have. The security guide or the manager run out to you and give you life-saving treatment. And so figuring out how to sort of turn on and enable the response that you need to help you across all of these channels is really what we were trying to avoid by saying we've got our solution, you got to do it. We figured we've got to connect everybody.
Russ Fordyce 🏆: Yeah, yeah, and and the infrastructure is there, the internet, the all all of these connected devices, all of these sensors are are built in. And now you guys are layering in this AI for s for intelligence, which I totally get. And it it will definitely make responses faster. How do you how are you using kind of legacy data in any way to to kind of, you know, train the AI and make and make it kind of richer? Have you have you dabbled in that at all?
Zach LaValley: Yeah, we're very keen on on privacy. So because we're tapped into all of these sort of legacy data systems, what we do is we sort of look at those and we create a lot of synthetic data based off of those systems. So again, it's more than anonymized. We're just looking at the shape of that, what what does it need to do? and we take those things in and we evaluate basically a lot of different models off of them, and then we work with the underlying model providers. in many cases to provide them with a synthetic test bed and so that we're constantly in this evaluation mode to say like what models are working best for the use cases of public safety without giving anybody's actual data or information out.
Russ Fordyce 🏆: Well, you are talking about a very vulnerable part of the system, right? It's emergency response. There is a lot of public data there that needs to be protected and and private data. and trust is kind of in your is, you know, the middle name of the company, I imagine, is you've got to have the trust of the agencies and of of all sides, right? Or it doesn't work. how do you how do you go into this these kind of very traditional environments? They're not super tech heavy. they've had that infrastructure for a while and say we're gonna bring this AI in and you just trust me. How do you kind of prove to them that this is this model works and that, you know, that this connective fabric's actually gonna help them and not just, you know, cause more noise.
Zach LaValley: Yeah. first we we come in and we say, like, let's talk about where we want humans to to participate. We can have them be directly in the line, we can have them be on the loop, we can have them monitoring. Let's talk about the problem we're trying to solve with you, agency, and let's make sure that we are crafting a feedback mechanism in a in a structure that makes you feel comfortable. So that's one is let's make sure that we are putting AI where you feel comfortable. Does not need to be making decisions, it does not need to be making actions right off the bat. Great. So that's step one, build trust from that perspective. step two is basically always ensuring so I like to have people think about we never try to replace the old way. What we're trying to do is we're trying to, if we think about the old way as being a highway or a road, we're trying to add a lane and that lane is fast. It's the fast lane. And I want to divert some traffic to that fast lane. And anybody who goes down that fast lane, man, they just see a benefit. But we don't have to put all the traffic down that lane. And so you can basically govern or throttle the amount of traffic you want to put on the fast lane and we just get comfortable with it. And then ideally over time we drift most or all of the traffic that way. And at some point we might we might consider we get rid of the old way. But we've built up the habit and the confidence that the fast lane's going work and then again technology is going to evolve. There'll be another fast lane that we put on somewhere else. But we just want to keep that pattern and we always want to be able to feel like we can go back. The the system is working. Is it working optimally? No, but it works. And so you don't want to just rip it out under a promise that the new one is better. You want to be able to shift people over in a in a consistent, in a, in a feeling confident kind of
Russ Fordyce 🏆: Well, this is this is such a people business. It's grown you know, it started, is that right? You wanted to pick up the phone, dial nine one and talk to a human. Right. That was the idea is you're getting you're getting into that response channel very quickly. And I know a lot of those centers have been kind of, you know, have you know, budget cuts and everything else, you know, forced into, you know, kind of deprecated environments in a lot of instances. What's being done to the kind of the overall system of nine one one that's kind of enhancing that other than this layer you're talking about, are there any kind of fundamental infrastructure changes going on there that that will, you know, benefit us? Hey Zach, if you can just rejoin, we can pick up right where we left off. Period. Not sure what happened, but I'm still recording. Period. The system will just pick right up. And we're back.
Zach LaValley: We'll see.
Russ Fordyce 🏆: no, it's it it it was designed for this. I didn't
Zach LaValley: Right.
Russ Fordyce 🏆: give you that part of the spiel and I I I apologize. If it ever goes down and you just dial back in and it picks right back.
Zach LaValley: Russ, I asked you at your last
Russ Fordyce 🏆: David.
Zach LaValley: question. So
Russ Fordyce 🏆: Yeah, no, that that's fine. We can well let's pick up from Let me pick up. It was kind of all in the trust. Let me see here.
Zach LaValley: I just described the lanes, the fast lanes, and you were about to ask another question.
Russ Fordyce 🏆: Yeah. yeah. You ready? All
Zach LaValley: Yes.
Russ Fordyce 🏆: right. How do you draw the line between kind of what the AI is gonna do and what the human's gonna do? I imagine it's a kind of a customer by customer decision. I guess it actually it might even be a better question. How do your customers decide? And I know it's variable, but in this day and age, where we are at this time, you know. How are they deciding kind of where that edge is between human and AI?
Zach LaValley: Yeah. I I there's different use cases. So a place where we have a human completely in the loop, right? So the human is the gating factor, is AI might auto-translate or auto-transcribe a phone call and it might recommend here is a you know here's an interesting piece of information for you to know. we need this missing piece of information according to your Agencies protocols or standard operating procedures, this is the next question you should ask. And so it's suggesting. But the human is doing all the talking, and the human is is basically responsible for taking all the actions. That's like human in the loop. A human on the loop might be a supervisor is using our AI agent to answer non-emergency calls. So most people don't know this, but because 911 agencies. Are essentially a call taking center. They often handle for a county or a municipality emergencies and non-emergency calls. So things like 311 and other stuff might come into the line. It's not coming in on the 911 line, but it's the same people answering the phones. And so we have an an AI system that can receive those calls and respond to them. And so a supervisor would be monitoring a dashboard of all the transcriptions that are happening there. And being able to step in and say, like, nope, like that's not going the right way, or that person's actually having an emergency, we should do something. and so we design a system that allows for the human to step in. a human over the loop is QA, right? Quality assurance inside of these systems, it was a very human process. And so when you're taking thousands of calls, you know, in a day or a week. You're only going to sample and actually have a QA manager listen to 1% of those calls. But now with AI, that that manager can have an AI listen to 100% of those calls and give them a summarized report to say: here are the types of calls, here's the quality of those recalls, here are two or three that you should go and have a conversation with those telecommunicators. And so You just have to pick your use case and pick the right human AI relationship. And and there's lots of different variants there. I'm just highlighting three of common ones that we see within our our ecosystem.
Russ Fordyce 🏆: Yeah, the the other interesting thing to me is I know there's there's a guy in telecom, Jeff Polver, who's been in the business forever. He was one of the kind of inventors of VOIP and promoters of it in the beginning. I d and I don't know if you know him, but he's working on the metadata for calls and making sure that the metadata around calls is is attached to the call. Is that metadata also being captured by your system and by by the county system so that they're you know they're They're bringing in all the data around that call too.
Zach LaValley: Not uniformly. So it really kind of depends on where and how we can access information. I think it's on our radar to try to increase that because the metadata can be useful in helping guide and understand and create intelligence, right? Like any one of these pieces of data itself is not necessarily valuable. It's in trying to meld them all very quickly to form intelligence that is useful for response. and so the more you can kind of put into that soup, right, AI is really good at being able to ingest an immense amount of information all at the same time and try to come up with some of those those interesting insights and pieces of intelligence that either humans miss because we over pattern match, or it's just literally too much of a fire hose for us to to comprehend.
Russ Fordyce 🏆: Yeah, and and now you guys are this new fabric, this kind of data layer, this new data layer, right? almost adjacent data. what are the systems doing to kind of improve their performance? Not your system, but you know, the overall nine design of the nine one one type of systems, are those being upgraded as well? I imagine, you know, they go through an evolution just like every technology.
Zach LaValley: Yeah. I think we're seeing, especially in offerings that we have, better decision making support. So, you know, I'd say the the last iteration of technology inside of 911 was about record keeping, right? So are we do we have auditability? Are we storing things precisely? that kind of turned humans into robots, right? So humans basically are just entering in forms, they're trying to ask questions and then they they're just entering in forms. AIs can do that piece. And so now the next evolution is being able to provide better decision making tools for the humans, right? Like, do they have more visibility into those things? Conversely, all of the things that we connect in, we can help them make their offering better. Right. Like we we worked with some of the smartwatches, and early days they were spamming 9 1 1. And so we created visibility for that and they improved their product and now they they don't spam nine one one. And so it actually has made their smartwatches safer. and we see that all over the place too, is we can actually kind of show the products that surround us how to make their product better and more useful in these types of situations. and we've seen them advance and listen.
Russ Fordyce 🏆: Yeah, and and that's that pro type of product feedback. Those guys want it, those people want it just as much as as the 911 callers don't want to get those calls, right? They they don't want to send them. the interesting thing here is you're really into this network effect, right? The more customers you get, the more data you have, the more partners that you have on the network, the more data you have. This kind of flywheel keeps going in the next three to five years. How does that change what emergency response looks like?
Zach LaValley: well, I I'm no tea leaf future reader, I guess at a high level, what I hope is that we all just see outcomes response times go down. we see outcomes improve. we see a an optimization. Again, we kind of use that like example of the speed of the car crash. It's hard for us as a society to regulate. How much response is correct? Like over response is a waste of those response and might be a distraction for something else important. Under response has bad outcomes. And so how do we create kind of this feedback loop where we can kind of hone in much more on what is the right level of response, which then should overall allow for more response and like so I like it's that virtuous cycle that I sort of expect for us to be on. And where that cycle takes us, I'm not really sure. but getting on that cycle is the important.
Russ Fordyce 🏆: Zach, very interesting conversation. I you know, it's it's one again, like I said in the beginning, it's one of those things you never knew existed, but you're just happy it does. Right. You you learn about it and you're like, that's very, very cool. I I hope I hope you guys all the best. I wish you guys all the best. I mean, all this stuff is connected in some way and it's the data is there. It's just really putting it all together and getting in the right place. And I'm I'm glad you guys are out there doing it. It's very interesting. So congratulations on the award and thanks for being on the show.
Zach LaValley: Thank you very much. Quick plug. check your phone. Make sure you have your settings correct. you know, again, if if we can help you in your time of need, feel free. But a lot of it's under your control, so just go in and and look in your settings.
Russ Fordyce 🏆: So what are we looking for under our settings? You're looking for SOS or what are we looking for on Android and iPhone?
Zach LaValley: nope, you you just want to go in and say like in the emergency health profile, enable that so that it can be shared during a 911 call. It's the only time it'll be shared is during a 911 call. and also look for and ask for your your technology vendors. When you buy a car, is the car connected? Who is it connected through? Those kind of things. Just starting to create a sense of awareness that you as a consumer can have choice in the types of things that you surround yourself with, and that ideally those choices can come to your aid when you need it.
Russ Fordyce 🏆: Well, Zach, thank you so much. Great advice. Appreciate it. And good luck to the company and all of the partners out there who are connecting to this awesome fabric. So thanks again.
Zach LaValley: Amazing. Thanks, Russ.
Zach LaValley: Thanks, Russ. great to be recognized. really mission driven company. So it's it's nice to to to feel that recognition.
Russ Fordyce 🏆: Yeah, this is one of those companies that you never hear about that's fundamental to pretty much most of America, it seems like, from from the numbers that you guys tout of customers. you guys are that underlying network that all our first responders use, fire police, emergencies, government schools. So that they're off kind of the public internet and they've got their own channels, right? Why don't you give us kind of the elevator pitch for rapid SOS?
Zach LaValley: Sure. So we're the world's largest AI-powered safety network. So we connect 723 million different devices, 250 different enterprises across North America largely, and about 23,000 responder agencies, right? And they take a lot of different different shapes and forms. And the way that I would think about it is traditionally 911 emerged as a way of calling for help. and that was excellent and frankly revolutionary when it first came out, and we all had land lines in our houses, and the landlines never moved, and so you knew exactly where someone was standing when they were talking to you on the phone. The world has evolved tremendously since then, but the underlying technology and infrastructure lagged. and so really we were still very reliant on for a very long time the phone lines being it. And trying to triangulate us just off of where we were. And so we moved to cell towers, and then we're triangulating off cell towers. But now it's so much more than phone calls. So, you know, your car knows when it gets into a crash. your phone knows when it gets to a crash, your smartwatch does. We're surrounded by an ecosystem of technology that is aware and able to help us in our times of need. and so it's more than just the call. It is the call, but it's more. And so we help connect those devices into first 911 and then out to the field as as an emergency emerges. And so your phone provides, you know, your location when you call 911. But if you set up things correctly, it can also provide your health information. And only in the case where your emergency is health related, but would we connect that in? But we're the conduit to make sure that it gets to the people who can help you. and so that's what that's Yeah. Yeah. Yeah.
Russ Fordyce 🏆: Yeah, it's like almost like the connective tissue, right? It seems like. Yeah. So you're you're playing this kind of role in the middle of of kind of the responders, the equipment, the communications. And you talk about yourselves as infrastructure. Walk us through kind of what a deployment normally looks like. Is it software, hardware, a mix of both solutions? What what do you put out in in into
Zach LaValley: some of both.
Russ Fordyce 🏆: a
Zach LaValley: Yeah, some of both. largely the network is internet based. and so we we play in various levels of the internet. So we are available over the commercial internet, because sometimes the the private internets that the nine one one system and responders also connect into, sometimes those go down. So we act as a redundancy backup between those those various levels of the internet for 911. We also provide some hardware in certain cases in order to help digitize and make more available pieces of information. So pieces of hardware that might go into the 911 in order to access and understand what is happening inside the phone call. Radio systems that might live alongside of the radio system between the first responders in 911 So that we can understand what's happening in communication streams between them. So again, so that we can foster better communication and more overall awareness within that community in order to to facilitate again faster, better outcomes for the public.
Russ Fordyce 🏆: Yeah, it it's it's really amazing at how far this system has come to me. I I was in telecom for a lot of years and supported nine one one, you know, through the telecom channels. And as it did, as as I P came into that equation, as we did BOIP, you had to move to kinda E nine one one, which is like kind of self reported, which was horrible. You had to get employees to like sign up to tell you where they were and if they were on you know, we sold a soft phone and you know, that ran on an Android device. How do you do nine one one on that? You constantly, you know, updating an address. It didn't work. And then yeah, and then GPS came in. It seems like this infrastructure, a lot of it is aging. Why do you think it's taken so long for this kind of aging infrastructure and technology to to get upgraded and it become of age?
Zach LaValley: Yeah, there's probably lots of factors. change is hard, right? And so it's also so that's I think the the driving one. there's a real pro and a con to the way that we, at least here in in the US and largely in North America, have designed our nine one one system. it's very decentralized. So there is no national nine one one. It it's county by county. Runs their own 911 centers and they have their own connections out to the field responders that surround their community. And so you kind of have to make this change 6,000 times across the United States. And not everybody's in the same camp, and not everybody's on the same tech stack. And so it is just sort of this evolutionary game that has to be played across this industry.
Russ Fordyce 🏆: So w walk us through you t you were talking about the connected car, everything's connected now. I mean, we've got sensors and watches and bands. They're coming into your earbuds next to your ear, you know, your earphones. I saw that. We have a a customer or a a nominee and winner who's who designs hearing aids and and all that technology is going in there now too. Walk us through kind of what happens when they're not a rapid SOS customer and where kind of the data drops versus kind of what you guys do and where you come in and fill in the parts.
Zach LaValley: Yeah. So for those companies that aren't a part of our network, most of the times that information never makes it into 911 or emergency services. The only way that it would is if a human being read off of the app that maybe you are using and said those words to someone in 911, or and then hope, fingers crossed, that information gets out to the first responders. Because when you call 911, you're not talking to a first responder, like the field responders. You're talking to a 911 telecommunicator or a dispatcher. For some connected devices, you might go to a monitoring center, right? So these are private entities. Your home security system is very is basically designed this way. home security system fires off, it goes to a monitoring center.
Russ Fordyce 🏆: us. So
Zach LaValley: Perfect. Okay. So some of these connected devices are connected to to monitoring centers. and so they are, you know, basically a human-based filter before we get 911. The a lot of these signals are very noisy, right? And you get false alarms all the time.
Russ Fordyce 🏆: Yeah.
Zach LaValley: And so the traditional way, again, before we had AI and before we had a lot of these sort of smarter techniques, was another set of humans would filter those calls. and the way it would work is they might call you back as the homeowner. verify with you if it is real, and and again, 99% of the time it's not real, but if it is real, they would pick up the phone and they would call 911 and then they would verbally relay all the information that they just said to you into 911. now let's talk about if a if a customer or an enterprise is part of our network, all of that digital fingerprinting and is automatically into the network and gets funneled to the right people. So when you call 911 or your smartwatch gets into a crash, it will automatically send its location and any relevant information up through into our network, and we will automatically route it to the 911 center that is responsible for handling that action. And then as they dispatch resources, the relevant information, like location, maybe like a health situation you're having, will get sent out to an app on the responders in their pocket. In the case of monitoring, which still happens, all of that literal game of telephone goes away. So yes, the monitoring agency might still provide you with fantastic customer service, call you back, make sure that you feel okay. Most of the times it's still nothing. But in the cases where it is something, no longer do we have to play that game of phone relay. All of the information that was gathered is now automatically sent into the systems of record. inside of the nine one one agencies and then if again, if appropriate, out to the field responders if it's valuable for them in response.
Russ Fordyce 🏆: That really changes the day the way that I mean, I'm just thinking through like an ambulance, right? You know, if you've fallen and broken something, you know, it really changes the way care is going to be provided, I would imagine, if if the ambulance is arriving knowing, you know, hey, you fell twelve minutes ago and your heart rate is this and your blood ox is this. I i it can really determine kind of life or death things, I would imagine. I don't mean to be too, you know, blunt about it, but I I Yeah it's it seems remarkable we didn't have this layer before.
Zach LaValley: I I think most good things feel obvious in retrospect. and yeah, it just took some work to actually make all these connections. not only is it better responses, it's more appropriate responses, right? So knowing the speed with which you got into a car crash and the number of people inside of the car because you had seatbelt sensors totally determines how people will respond and the level of response needed. If you got into a two mile an hour bumper inside of you know a parking lot that's way different than 40 miles an hour and there's you know four other you know two kids in the back. and so now we can we can get all of that and especially in these cases where man maybe no one no one dialed the phone. I we've got a case out in Cass County, Michigan where we sent in a crash report and it was it showed off the road because we had the GPS. No one called in that person was unconscious. It was on a
Russ Fordyce 🏆: Yeah.
Zach LaValley: county road, but within 13 seconds, the dispatchers, dispatch people, we got there in four minutes and you know, pulled a guy out of a burning car. and so these are the these are the really positive examples of of how having sort of this interconnected safety network really can bring positive positive outcomes to us.
Russ Fordyce 🏆: Yeah, no, you guys chose this interesting approach too, of kind of layering you've you've brought this new layer of data and AI to the existing infrastructure. What what made you kind of decide on that approach versus going in with new equipment and new hardware to kind of do a replacement versus kind of an augmentation?
Zach LaValley: What we have found is that we wanted to resist ourselves. When you go in and you bring your own stuff, you create another silo. And really the problem we're trying to solve is we're trying to overcome all of these data silos. So all of these people who bring in their own hardware and their own networks, they're then incentivized to try to cordon off and isolate and cons a customer that like it is more secure to keep these things completely siloed. and we really think that you don't want to create an inner silo inside 911, but like almost nobody, almost no emergency requires just fire or just police, or like we are surrounded in a connected ecosystem. And so if you put them in silos, all of a sudden fire EMS and law enforcement all show up. Two of them are completely blind if one of them was part of the silo. And so That problem exists everywhere. And it's not just in this public, right? Like if you got into a car accident inside of a commercial parking spot, that retailer might have you know the help to get that you need, right? Or maybe not maybe maybe you're having a heart attack. They might have an AED inside of that building. You don't want to wait for the ambulance to show up if we can turn on that AED and have. The security guide or the manager run out to you and give you life-saving treatment. And so figuring out how to sort of turn on and enable the response that you need to help you across all of these channels is really what we were trying to avoid by saying we've got our solution, you got to do it. We figured we've got to connect everybody.
Russ Fordyce 🏆: Yeah, yeah, and and the infrastructure is there, the internet, the all all of these connected devices, all of these sensors are are built in. And now you guys are layering in this AI for s for intelligence, which I totally get. And it it will definitely make responses faster. How do you how are you using kind of legacy data in any way to to kind of, you know, train the AI and make and make it kind of richer? Have you have you dabbled in that at all?
Zach LaValley: Yeah, we're very keen on on privacy. So because we're tapped into all of these sort of legacy data systems, what we do is we sort of look at those and we create a lot of synthetic data based off of those systems. So again, it's more than anonymized. We're just looking at the shape of that, what what does it need to do? and we take those things in and we evaluate basically a lot of different models off of them, and then we work with the underlying model providers. in many cases to provide them with a synthetic test bed and so that we're constantly in this evaluation mode to say like what models are working best for the use cases of public safety without giving anybody's actual data or information out.
Russ Fordyce 🏆: Well, you are talking about a very vulnerable part of the system, right? It's emergency response. There is a lot of public data there that needs to be protected and and private data. and trust is kind of in your is, you know, the middle name of the company, I imagine, is you've got to have the trust of the agencies and of of all sides, right? Or it doesn't work. how do you how do you go into this these kind of very traditional environments? They're not super tech heavy. they've had that infrastructure for a while and say we're gonna bring this AI in and you just trust me. How do you kind of prove to them that this is this model works and that, you know, that this connective fabric's actually gonna help them and not just, you know, cause more noise.
Zach LaValley: Yeah. first we we come in and we say, like, let's talk about where we want humans to to participate. We can have them be directly in the line, we can have them be on the loop, we can have them monitoring. Let's talk about the problem we're trying to solve with you, agency, and let's make sure that we are crafting a feedback mechanism in a in a structure that makes you feel comfortable. So that's one is let's make sure that we are putting AI where you feel comfortable. Does not need to be making decisions, it does not need to be making actions right off the bat. Great. So that's step one, build trust from that perspective. step two is basically always ensuring so I like to have people think about we never try to replace the old way. What we're trying to do is we're trying to, if we think about the old way as being a highway or a road, we're trying to add a lane and that lane is fast. It's the fast lane. And I want to divert some traffic to that fast lane. And anybody who goes down that fast lane, man, they just see a benefit. But we don't have to put all the traffic down that lane. And so you can basically govern or throttle the amount of traffic you want to put on the fast lane and we just get comfortable with it. And then ideally over time we drift most or all of the traffic that way. And at some point we might we might consider we get rid of the old way. But we've built up the habit and the confidence that the fast lane's going work and then again technology is going to evolve. There'll be another fast lane that we put on somewhere else. But we just want to keep that pattern and we always want to be able to feel like we can go back. The the system is working. Is it working optimally? No, but it works. And so you don't want to just rip it out under a promise that the new one is better. You want to be able to shift people over in a in a consistent, in a, in a feeling confident kind of
Russ Fordyce 🏆: Well, this is this is such a people business. It's grown you know, it started, is that right? You wanted to pick up the phone, dial nine one and talk to a human. Right. That was the idea is you're getting you're getting into that response channel very quickly. And I know a lot of those centers have been kind of, you know, have you know, budget cuts and everything else, you know, forced into, you know, kind of deprecated environments in a lot of instances. What's being done to the kind of the overall system of nine one one that's kind of enhancing that other than this layer you're talking about, are there any kind of fundamental infrastructure changes going on there that that will, you know, benefit us? Hey Zach, if you can just rejoin, we can pick up right where we left off. Period. Not sure what happened, but I'm still recording. Period. The system will just pick right up. And we're back.
Zach LaValley: We'll see.
Russ Fordyce 🏆: no, it's it it it was designed for this. I didn't
Zach LaValley: Right.
Russ Fordyce 🏆: give you that part of the spiel and I I I apologize. If it ever goes down and you just dial back in and it picks right back.
Zach LaValley: Russ, I asked you at your last
Russ Fordyce 🏆: David.
Zach LaValley: question. So
Russ Fordyce 🏆: Yeah, no, that that's fine. We can well let's pick up from Let me pick up. It was kind of all in the trust. Let me see here.
Zach LaValley: I just described the lanes, the fast lanes, and you were about to ask another question.
Russ Fordyce 🏆: Yeah. yeah. You ready? All
Zach LaValley: Yes.
Russ Fordyce 🏆: right. How do you draw the line between kind of what the AI is gonna do and what the human's gonna do? I imagine it's a kind of a customer by customer decision. I guess it actually it might even be a better question. How do your customers decide? And I know it's variable, but in this day and age, where we are at this time, you know. How are they deciding kind of where that edge is between human and AI?
Zach LaValley: Yeah. I I there's different use cases. So a place where we have a human completely in the loop, right? So the human is the gating factor, is AI might auto-translate or auto-transcribe a phone call and it might recommend here is a you know here's an interesting piece of information for you to know. we need this missing piece of information according to your Agencies protocols or standard operating procedures, this is the next question you should ask. And so it's suggesting. But the human is doing all the talking, and the human is is basically responsible for taking all the actions. That's like human in the loop. A human on the loop might be a supervisor is using our AI agent to answer non-emergency calls. So most people don't know this, but because 911 agencies. Are essentially a call taking center. They often handle for a county or a municipality emergencies and non-emergency calls. So things like 311 and other stuff might come into the line. It's not coming in on the 911 line, but it's the same people answering the phones. And so we have an an AI system that can receive those calls and respond to them. And so a supervisor would be monitoring a dashboard of all the transcriptions that are happening there. And being able to step in and say, like, nope, like that's not going the right way, or that person's actually having an emergency, we should do something. and so we design a system that allows for the human to step in. a human over the loop is QA, right? Quality assurance inside of these systems, it was a very human process. And so when you're taking thousands of calls, you know, in a day or a week. You're only going to sample and actually have a QA manager listen to 1% of those calls. But now with AI, that that manager can have an AI listen to 100% of those calls and give them a summarized report to say: here are the types of calls, here's the quality of those recalls, here are two or three that you should go and have a conversation with those telecommunicators. And so You just have to pick your use case and pick the right human AI relationship. And and there's lots of different variants there. I'm just highlighting three of common ones that we see within our our ecosystem.
Russ Fordyce 🏆: Yeah, the the other interesting thing to me is I know there's there's a guy in telecom, Jeff Polver, who's been in the business forever. He was one of the kind of inventors of VOIP and promoters of it in the beginning. I d and I don't know if you know him, but he's working on the metadata for calls and making sure that the metadata around calls is is attached to the call. Is that metadata also being captured by your system and by by the county system so that they're you know they're They're bringing in all the data around that call too.
Zach LaValley: Not uniformly. So it really kind of depends on where and how we can access information. I think it's on our radar to try to increase that because the metadata can be useful in helping guide and understand and create intelligence, right? Like any one of these pieces of data itself is not necessarily valuable. It's in trying to meld them all very quickly to form intelligence that is useful for response. and so the more you can kind of put into that soup, right, AI is really good at being able to ingest an immense amount of information all at the same time and try to come up with some of those those interesting insights and pieces of intelligence that either humans miss because we over pattern match, or it's just literally too much of a fire hose for us to to comprehend.
Russ Fordyce 🏆: Yeah, and and now you guys are this new fabric, this kind of data layer, this new data layer, right? almost adjacent data. what are the systems doing to kind of improve their performance? Not your system, but you know, the overall nine design of the nine one one type of systems, are those being upgraded as well? I imagine, you know, they go through an evolution just like every technology.
Zach LaValley: Yeah. I think we're seeing, especially in offerings that we have, better decision making support. So, you know, I'd say the the last iteration of technology inside of 911 was about record keeping, right? So are we do we have auditability? Are we storing things precisely? that kind of turned humans into robots, right? So humans basically are just entering in forms, they're trying to ask questions and then they they're just entering in forms. AIs can do that piece. And so now the next evolution is being able to provide better decision making tools for the humans, right? Like, do they have more visibility into those things? Conversely, all of the things that we connect in, we can help them make their offering better. Right. Like we we worked with some of the smartwatches, and early days they were spamming 9 1 1. And so we created visibility for that and they improved their product and now they they don't spam nine one one. And so it actually has made their smartwatches safer. and we see that all over the place too, is we can actually kind of show the products that surround us how to make their product better and more useful in these types of situations. and we've seen them advance and listen.
Russ Fordyce 🏆: Yeah, and and that's that pro type of product feedback. Those guys want it, those people want it just as much as as the 911 callers don't want to get those calls, right? They they don't want to send them. the interesting thing here is you're really into this network effect, right? The more customers you get, the more data you have, the more partners that you have on the network, the more data you have. This kind of flywheel keeps going in the next three to five years. How does that change what emergency response looks like?
Zach LaValley: well, I I'm no tea leaf future reader, I guess at a high level, what I hope is that we all just see outcomes response times go down. we see outcomes improve. we see a an optimization. Again, we kind of use that like example of the speed of the car crash. It's hard for us as a society to regulate. How much response is correct? Like over response is a waste of those response and might be a distraction for something else important. Under response has bad outcomes. And so how do we create kind of this feedback loop where we can kind of hone in much more on what is the right level of response, which then should overall allow for more response and like so I like it's that virtuous cycle that I sort of expect for us to be on. And where that cycle takes us, I'm not really sure. but getting on that cycle is the important.
Russ Fordyce 🏆: Zach, very interesting conversation. I you know, it's it's one again, like I said in the beginning, it's one of those things you never knew existed, but you're just happy it does. Right. You you learn about it and you're like, that's very, very cool. I I hope I hope you guys all the best. I wish you guys all the best. I mean, all this stuff is connected in some way and it's the data is there. It's just really putting it all together and getting in the right place. And I'm I'm glad you guys are out there doing it. It's very interesting. So congratulations on the award and thanks for being on the show.
Zach LaValley: Thank you very much. Quick plug. check your phone. Make sure you have your settings correct. you know, again, if if we can help you in your time of need, feel free. But a lot of it's under your control, so just go in and and look in your settings.
Russ Fordyce 🏆: So what are we looking for under our settings? You're looking for SOS or what are we looking for on Android and iPhone?
Zach LaValley: nope, you you just want to go in and say like in the emergency health profile, enable that so that it can be shared during a 911 call. It's the only time it'll be shared is during a 911 call. and also look for and ask for your your technology vendors. When you buy a car, is the car connected? Who is it connected through? Those kind of things. Just starting to create a sense of awareness that you as a consumer can have choice in the types of things that you surround yourself with, and that ideally those choices can come to your aid when you need it.
Russ Fordyce 🏆: Well, Zach, thank you so much. Great advice. Appreciate it. And good luck to the company and all of the partners out there who are connecting to this awesome fabric. So thanks again.
Zach LaValley: Amazing. Thanks, Russ.
