transcript

Speaker 1:
[00:00] This episode is sponsored by BLTN, powered by Multitude Insights. See how BLTN connects law enforcement agencies nationwide with real-time bulletin sharing at multitudeinsights.com. Hey, welcome back to Policing Matters. I'm your host, Jim Dudley. Well, if you talk to almost any cop today, they'll tell you the same thing. We have more information than ever before. Alerts, emails, bolos, bulletins, intel reports, fusion center updates. But here's the real question. Are we actually getting the intelligence we need when we need it? Or are officers drowning in information while the critical pieces slip through the cracks? Well, today, we're talking about the future of intelligence sharing and policing. And my guest is Matt White. He is the CEO and co-founder of Multitude Insights, a public safety technology company focused on building modern intelligence infrastructure for law enforcement. Matt founded the company with his MIT classmate, Izuku Izu. And the idea came from Matt's own background, conducting more than 1,000 hours of combat reconnaissance and intelligence missions where getting the right information at the right time could literally mean life or death. Their flagship platform, BLTN, the Bulletin Platform, is designed specifically to help law enforcement agencies share real-time intelligence and collaborate across units and jurisdictions. Matt White, it's a pleasure. Welcome to the show.

Speaker 2:
[01:53] Good to see you, Jim. Happy to be here and I'm looking forward to this conversation.

Speaker 1:
[01:58] So let's talk about the problem with information. And Matt, let's start with the big picture. When you say intelligence infrastructure, what does that really mean for the police department and for the officers on the front lines in radio cars getting inundated with all of this information?

Speaker 2:
[02:17] You kind of said it is intelligence can mean a lot of different things to a lot of different people. But when I say it, I think a lot about the information that a police officer, whether it's a patrol officer in a cruiser, somebody on a beat cop or a detective or crime analyst, the information they need to do their job. And at the end of the day, that's protecting public safety and driving solvability on crimes. There's a lot of the big players that people know about. There's Flock, there's other ALPR cameras, e-mail based bulletins and crime intel updates from fusion centers. There's radio calls coming in live while you're cruising down the street. There's a chorus of information that comes in to any police officer on any given day. Compound that across large agencies in major metropolitan areas, and you got yourself a complex information environment. And that's the type of environment I came from in the Navy side of things. And so I tried to, when building bulletins, I tried to take some of those lessons learned and provide a tool and a service for police officers.

Speaker 1:
[03:17] Yeah, I mean, some of that info, we've streamlined ourselves, right? I remember as the Deputy Chief of Operations, we took a look at over 300 bulletins that we issued in a year. And we said, how is it possible for our officers to maintain knowledge about 350 bulletins, some more information bulletins, some more training bulletins? And then we broke it down into ABC priority, where A, you had to know, B, you had to know something about it, and C, well, you could research it if you needed to. Are you doing something like that to funnel this information?

Speaker 2:
[03:58] Yeah. So I'll start with kind of how we think about it. If you zoom out, before I answer your actual question, you zoom out and you think about what a bulletin is inside of law enforcement. It's something, one, that every single one of the 18,000-ish agencies in the United States has a process for. They're either creating them or consuming them. And when do you release a bulletin? Well, sometimes you do it for some CYA purposes, but generally you do it because you have a crime nexus that you're trying to solve. So whether it's a new pattern, a series, an explicit event, those sort of things. It is a police department raising its hand and saying, we need help with something. Missing person, identifying a suspect, these type of things. And so if you zoom out and you say, wow, if we could connect a lot of those, we could start to do something pretty impressive. And if we could start to filter those, as your question kind of asks, we could do something very powerful by really driving the most important thing to the right officer at the right time. And that's really what we try to do. So you ask about filtering A, B, C, and D, that's a great way to do it, pretty manual. And so what we do is we have adopted a basically a content recommendation system model. Now that sounds pretty fancy, so I'll put it in plain English, but it's more or less what social media does, where it takes all this content that's out there and it says, Jim Dudley is really interested in barbecue. I don't know. I'm just firing from the hip here, Jim, but we're going to show you lots of that. Good. I thought you might be. Yeah. We'll see. You can tell it's like lunchtime when we're recording this because I'm getting hungry over here. But all to say that I can show the robbery detective all the appropriate robbery bulletins that are good for his or her beat. I can show the narcotics detective all the fentanyl information going on in town, and I can show the school resource officer the gang information around their local high school. They may not need to see that 7-Eleven break in bulletin. And so by doing that, I can take the limited attention they have, drive it and focus it towards the right intel at the right time, drive situational awareness, therefore, delivering better safety and better crime solvability to the PD. That's the goal.

Speaker 1:
[06:18] Yeah, no, and it totally feeds into the research and the science that says our brains can only accept so much, right? It's not about a lack of information. Sometimes it's too much. And in my university where I teach, we teach small chunks, important information. Just as you said, we don't talk about juvenile crime when we talk about a parole section or an adult-related crime, because it doesn't really have the relevance. I totally understand what you're saying, and it makes sense. Well, you've spent a lot of time in military intelligence environments. What lessons from military intel operations could policing take and use it in policing situations?

Speaker 2:
[07:06] Yeah. So a couple of different things. One is I flew a signals intelligence and reconnaissance aircraft called the EP-3 for the aviation people out there. They can look it up. She's since been retired, so it gives you a sense of how long I've been out of the Navy at this point. But what we did was pretty special, which is take in all sorts of different sources of information, whether it be communications intelligence, electronic intelligence, all these different sources, and try to make sense of them and distribute them to the right recipient. So whether that was a three-letter agency, another component inside the Navy, or another service altogether, Air Force, Army, etc. That was our job, is triaging and doing that. So we did it relatively manually, but we also had some software to help us out as well, and decide where things need to go. Now, this is a little meta, but going back to post-911, that was the mandate. We had this intelligence failure. We were not good at knowing what needed to go where. And across the Global War on Terror, those were the systems that grew up and matured. Notably, Palantir really helped with that type of thing. And so I took the lessons learned in the water. I was swimming inside the Navy and said, hmm, law enforcement really has nothing like this. And I found this, I was kind of, an interesting story is, I was a graduate student at MIT at the time after active duty service. And the Boston Police Department came in and gave a lecture on the state of public safety technology. It was really cool. They had some cool tech. They had like the Mark 43 record management system that was really neat. But that was really about it as far as like, you know, advanced technology. There was a mention of the word fax machines at some point, right? There were things like this, and this was 2020. And so, you know, I listened to that briefing, and I ended up speaking with the captain that gave the briefing afterwards. And I said, hey, sir, you know, I'm Matt, former Navy guy. Like, I'd love to come see what you do. And he invited me down and took me in for a ride along and showed me the tech stack of a modern police department. And from there, I said, wow, like, I've got all this experience doing this thing kind of in the Navy side of stuff. Wouldn't it be cool to apply those lessons learned to a problem here at home now? And so while the outcome is much different, we're looking for community safety, not warheads on foreheads, so to speak. It's very different outcomes, but this technology can be applied in similar ways.

Speaker 1:
[09:40] Yeah, so you mentioned breaking it down. So you talk about the relevance of information. You say, okay, here's the gang investigator, the narcotics detective, the school resource officer, and everybody's getting their own flow of information, but what about the overlap? So I love the fact you brought up 9-11, and we're coming up on the 25th anniversary. And so I teach a component of terrorism where we want those ground officers to be aware in any response of what it looks like if somebody's making IEDs, right? And the guy is not a plumber, but he's got all these pipe fittings all over, or he's got jars of nails, and he's got liquids. Yeah, and all that stuff. So how do you account for that overlap where you want everybody to know their own domain information, but you also want that cover spread too?

Speaker 2:
[10:35] That's right. Yeah, and this is where some of our special sauce technology kind of comes in, which is to say, I want to deliver that 7-Eleven break-in to the robbery guy, but I also want to make sure I hit him with things that he may never have seen in the past because we think they're relevant for him. So we really tune the algorithm to deliver to that, let's use this robbery detective who does commercial break-ins, like that's his thing, that's what he does. But we start noticing a trend of these commercial break-ins with a very unique MO that's a lot more sophisticated than normal. Our system will pick up on these things, link them together and then deliver that to the officer as well. And so it is designed to kind of think like the way your best crime analyst did, right? The way your very, very best crime analyst or detective is used to like kind of stringing these things together, coming up with a theory and presenting it to the officer, that's what we try to do too. So whether it's in our delivery mechanism, which is the feed, or whether it's in our smart linking technology that links together this content, we really try to, we're not trying to silo them. We want to just drive up engagement by showing the most important things to them. And then when there are those, the guy with a few too many extra pipe fittings and things like that, well, we're going to detect that in your post. And then I might also detect it two states away, right? Because this is a nationwide system for us. I might be able to see it two states away, and that actually gives me better reach and allows San Francisco PD to immediately get in touch with Oklahoma City and link together and say, hey, we had a pipe fitting thing happen too. We arrested three guys with that in the last year. You have two. Maybe we should compare deeper notes. Maybe there's something bigger going on here, and maybe this should get referred to Homeland Security.

Speaker 1:
[12:25] Yeah, that is so great to hear because I come from, when I was an inspector investigating patterns of burglaries in San Francisco, I would literally have to recognize the trend, pick up the phone, call Oakland, and ask, are you guys seeing the same thing? So, we were lucky if we were meshing and going regional, maybe within a county or two, but wow, to go across states or across the nation.

Speaker 2:
[12:54] Yeah.

Speaker 1:
[12:55] How great.

Speaker 2:
[12:56] Yeah, and Jim, that's literally the point is, it is so hard to rely on an individual officer's ability to network, connect and pick up the phone, because they're so busy. They've been under manned for a decade, right? They are, you are relying on Jim Dudley's personal network, which is good, but ad hoc at best, right? Yeah. I know you're a well-connected guy, but that is even still, you don't know Oklahoma City PD probably, right? And so, we wanted to systematize that knowledge, network that knowledge and allow the detective who's really getting after it in San Francisco Bay to get in touch with somebody they need to get in touch with and let the system prompt it as well.

Speaker 1:
[13:45] Yeah. Well, shout out to my former classmate, Rodney Burris from Oklahoma City. So, there.

Speaker 2:
[13:51] Hey, there we go. Okay, good.

Speaker 1:
[13:53] Glad to hear it. So, yeah, from the FBI Academy, he was awesome. Cool. So, you know, you and I talked a little bit off camera about AI and technology and how, what a great, important time for all this emerging technology to come in and help law enforcement agencies at a time when our recruiting strategies are really kind of stumped. Across the nation, we've done surveys, IACP and PERF and all the others that say 90% of agencies are down people, down and some are down significantly numbers. So we're filling the gaps with technology. And now everybody's talking about AI and managing intelligence flows. And I need to know from you, you're the guy, do we still need, I'm sure I think I know the answer to this one, but human analysts have to be somewhere in there, don't they?

Speaker 2:
[14:58] Yeah, absolutely. One, I think anybody who knows AI well and lives and breathes it the way I do, knows that it's still prone to quite a bit of mistakes. The way that police departments should be thinking about leveraging AI is to take care of these routine things that have to happen, that keep the gears of the machine running, that normally would take up to three, four hours of an officer's week, day, whatever it may be, and automating that stuff, submitting it to human review and freeing them up to be out in the community or doing more or solving more or any of the million other things a police officer or command staff needs to do. So I think about it as, yes, most police departments in the United States are down, 200, 300, 400 personnel and larger PDs. And AI technology can certainly help bridge the gap in that space. Now, Multitude's approach to AI is what we call human-in-the-loop technology. I borrowed the term from the Navy. It's not something I invented, but it's relatively self-explanatory. For example, when we issue a smart link connecting four or five bulletins across the Pacific Northwest, we don't just issue, Therefore, it was Jim Dudley in the billiards room with the candlestick. Like there is no statement of crime. It is, hey, the AI tool connected these things, and then an explainable, plain language narrative of like, here is how I made this connection. The way a crime analyst would explain their work to you. So we believe strongly in explainability for two reasons. One is, if you go on the stand with an AI tool, it said, well, the AI said it was Jim Dudley in the billiards room. Like that's not going to fly today. Maybe 100 years from now it will, but not today. And two is we also just think that there's a lot of other contexts that a given tool, whether it's ours or another company, it just doesn't have and so you need a human to take a look at it. But that does allow a detective, a crime analyst, a police officer to run 10 or 15 of these things at once instead of one. And so we can move through that backlog of cases better. We can be more proactive. We can deliver better policing with less resources.

Speaker 1:
[17:26] Yeah, 100 percent. So listen, we want to talk about some real world examples of how intelligence platforms are helping agencies solve crimes and connect the dots across the jurisdictions. You and I touched on it briefly about the regional coverage and then the state to state coverage, but I want to touch on those, but first I need to thank our sponsor. Better Bulletin's Solve Crimes. BLTN is the Nationwide Intelligence Sharing Platform built by law enforcement for law enforcement. One centralized system to create, distribute and analyze bulletins connecting agencies in real time, so critical intel reaches the right people when it matters most. No more inbox sprawl, no more missed leads, just faster communication and better outcomes. Visit multitudensights.com to see how agencies are closing more cases faster. That's multitudensights.com. All right, and I'm back speaking with Matt White, CEO and co-founder of Multitude Insights. And Matt, we're talking about real world impact now. And I'm going to ask you for a real world example. Walk us through a scenario where better intelligence sharing actually helped investigators connect the dots or help them to prevent a crime or a crime series.

Speaker 2:
[18:57] Yeah. There's a million to choose from. I'll kind of choose to one high and one low, and meaning one regional nationwide and one kind of PD to PD down the interstate type thing. So at the more regional or national level, like I said, our SmartLink tool is able to link together safety bulletins and detect things. We had a great example with a large city police department in the Midwest, posts a bulletin with a suspect and the association with a online kind of virtual militia type thing. Hey, officer safety bulletin, we've heard about these guys. If you see them, like might be something to ask them about. Just an awareness bulletin. They fired out. Well, unbeknownst to them, two or three weeks earlier, a state level agency in the deep south had published a awareness bulletin themselves about that same online militia and a known target, a weather radar, and saying this online militia really wanted to go after this weather radar. Well, our system was able to link those things together and send an alert out to both detectives and each of the two agencies. Those police officers then got on the phone and shared more notes and more detail than any crime bulletin would ever have, but it connected the dots. They got on the phone, and I know for a fact that arrests were made and the FBI was brought in as a domestic terrorism incident. So this system, again, on the backs of the humble bulletin, was able to make the connection and allow these guys to detect a major crime before it was about to occur. Now, importantly, it was just a connection, right? Saying, hey, there's a intelligence nexus here. It is not a probable cause. It is not a reason to arrest, but it did spawn the investigation, which then, of course, can lead to some sort of enforcement action if needed. So that was fantastic to see. And that was an early success of our platform. And, you know, importantly happened while one of the customers was a pilot customer. So we were happy they signed the contract right after that. And then on a more like case to case basis, we have a whole Slack channel at Multitude of like customer wins as they flow in. And it's amazing because I can get a new one in there every day. And a recent one was with some agencies in Southern California, two county level agencies. One sheriff's deputy posts a crime bulletin of a gentleman robbing a Target store and taking about $1,000 appliance out of Target. Next day, same thing out of a different agency, one block south basically. Photo very similar, MO is the same, everything. AI picks it up, sends out push notifications to those detectives. They get on the phone, they set surveillance, they go make an arrest. And thankfully, with the combined information across jurisdictions, gets way into felony territory, right? It's no longer just misdemeanor, right? So you're able to link those things together. And this has knock on benefits for the retail loss prevention teams out there, has knock on benefits for, you can imagine applying this across a vehicle theft task force at the state level. Like all these sorts of things, we have a lot of success with catalytic converter gangs and groups, and as they move across jurisdictions. So we just kind of keep stacking success stories like that, and that's really what's driving the growth of our network.

Speaker 1:
[22:28] Yeah, that's awesome. Those are incredible stories. And now we're getting into the season of all of our big events across America. We've got the FIFA World Cup coming up. Like I said, we've got the anniversary of 9-11, and I'm sorry, but the bad guys love anniversaries. So we're on high alert, and yeah, we can only expect Multitude and others to give us the good situational awareness picture to help us.

Speaker 2:
[23:01] That's right, Jim. And I'm kind of listening to you talk about some of the education work that you do at the university level. Like I'm working on an article right now, hopefully it'll be out soon, that talks about how the police officer on the corner is actually the more likely guy or gal to encounter nation state level bad actors and pre-operational surveillance and these sort of things, more so than any intelligence professional out in the world. And how do we prepare them? Well, you need to give them tools so that they understand what the heck is going on around them at all times, right? So this is, I sort of see what we're building at Multitude as like a continuation of this, like it's still in the national security context for me, especially as we approach these anniversaries.

Speaker 1:
[23:49] Yeah, well, I mean, it's hard to predict or to articulate prevention methods, but I don't care. I'm just glad you're preventing stuff. All right, so one challenge we hear about is information getting stuck inside the agency. How does BLTN help intelligence move not just across the department, but push out, if you don't have somebody at the button at a local agency saying, yeah, we should really send this out.

Speaker 2:
[24:20] Yeah.

Speaker 1:
[24:20] Do you do that?

Speaker 2:
[24:22] Yeah, great question. And like one of the core problems we set out to solve when we build, which is to say that let's say Officer Dudley, you're in charge of sending out the Crime Information Bulletin. You'll probably send it to your San Francisco All Sworn Distro, and you might send it to a couple of fusion centers in the area, and then maybe your job's done. Well, that relies on your personal knowledge of where you think it should go, and the email distribution list you have access to. Well, what we've tried to do is basically turn over that decision-making to something that has a lot more context than you. So while you can use our product to share things very narrowly with a specialized unit or subgroup of units, if you want to, most of our users, about 90%, hit the Share with Bulletin Network button. And when they hit that button, the Bulletin is open for distribution across our law enforcement-only system, so it's only police officers, right? But it is then shared where the algorithm thinks it needs to go, which is, like, of course, it goes to all your guys in your agency as appropriate, but it only shows up on some of their feeds, right? And then it goes out to, maybe you didn't realize that Walnut Creek in the East Bay needs to see that bulletin that you sent out, right? Maybe you would have only sent it to Oakland. Well, we actually know what's going on in Walnut Creek, and so we're going to send it to them. That type of thing. And so, yeah, that's a long way of answering a simple question, which is to say that, like, yeah, we try to take that on and then give the officer the ability to add some and remove some as required.

Speaker 1:
[25:57] No, but, I mean, we're talking about intelligence, right? And we're talking about how much we get and what's pertinent. And I think all of that is pertinent and relevant. And I think we're going to be hearing some chiefs and people in command that are saying, wow, this is the kind of system we need, not now, but right now. What does implementation actually look like? When does the cop on the street get to benefit from this kind of a system?

Speaker 2:
[26:24] Yeah, the nice thing is that this system, and crime bulletins in general is our seed data, right? It's relatively lightweight to implement. So this is not a RMS style overhaul, which I'm sure you've gone through, Jim, before where you purchase something and it's like at the end of a 10-year contract, it's going to take 12 months or 24 months to spin it up. And we are in the weeks territory of turning on. So from contract to go live can be very, very short depending on the size of the agency. So we measure in weeks, we'll put it that way. It's pretty quick to get online. Onboarding and training, we white glove it, we bring every agency and we train them in person, a bunch of electronic resources as well to the folks that couldn't make it that day or whatever. And it moves really quickly and time to value is as short as possible. So again, that's something else like AI can generate or create for us is like the time to value from when a chief says, hey, I want this, I need this, to it helping out Officer Jones on the corner. It has to be quick.

Speaker 1:
[27:33] Yeah, and shame on us if we're dragging our feet, hoping for the latest next iteration of a product or waiting till we get closer to the event. And I mean, yeah, we're great at procrastination. And I can think of a couple of events that went sideways because equipment or personnel or the other supplements were weighted on way too late. So we've talked about regional, we talked about the jurisdictional boundaries and we're talking about how AI can help. But on a platform like this, we want to share with the region, the state, federal, maybe even transnational. I mean, we've had such great success when we've pushed out to the transnational information sharing, especially when it comes to the drug epidemic, right? And we've seen that. I think we're seeing some tangible results. And when I say, you know, it's hard to measure prevention, well, when we go from 10,000 overdose deaths in America to 100,000 in this span of time, and then we're starting to see that that reduction of those overdose deaths, I can only attribute it to, I mean, going outside of our own borders and really being proactive. If we're going to share intelligence on things like organized crime trafficking, transnational crime, terrorism investigations, how do you do it?

Speaker 2:
[29:12] Yeah. Yeah. Well, they're more or less the same way is the short answer. The slightly longer answer is we obviously have to account for our federal rules and codes around what we're allowed to share with who, when, those things. Thankfully, our business speaks caveats and security classifications natively, so we understand how that works. The system today is really designed to be for the state level entities and below. So state, county, city, university, those levels today, because we found that's one level of nexus. Now, you're peeking into my roadmap of where BLTN will go in 2027 and beyond, is being able to do some of those other types of sharing. Again, cordoned off the way that it needs to be cordoned for compliance reasons and all other reasons. But there's some real value in being able to push out to those folks. And man, I talk to, we'll put it like, DOW clients today, and they're very hungry for this information. So they can support that broader mission of overdose reduction and drug enforcement, that type of stuff.

Speaker 1:
[30:22] Yeah, and I mean, even if we think about things like human trafficking, right? Or the sex trade, if you will. We've got a triangle here in Northern California, and Washington state, and Nevada state, where you get arrested a couple of times, you move up. It's this triangle, it's right, it's this carousel. And the same can be said with the drug trafficking.

Speaker 2:
[30:46] Yeah.

Speaker 1:
[30:48] Get enough money, make enough money, you get arrested a couple of times, then bet back to your native country. And then let it cool down a bit, come up with a different name. And we can't keep tracking people who operate like this, right? Because they're going to different states and different cities, back to their home country and back up. So incredible to get this kind of information, hold on to it, and then bring it back to law enforcement agencies when they go active again.

Speaker 2:
[31:19] That's right. Yeah, absolutely. I mean, that type of load of tracking that is something that only federal agencies have the resources to do. And even then, it's constrained because there's a million different ways to do this. And so my hope is that with more tools becoming more available at these local and state level and below levels, we can reduce a lot of that load and help them prosecute where they need to be able to prosecute and drive a lot better community outcome.

Speaker 1:
[31:46] Yeah. Well, Matt, you've shared a lot of time and information. I appreciate it. It's been a fascinating look at how technology is changing the law enforcement landscape and how we share and use intelligence. Thanks so much for taking time with us and for the work that you're doing at Multitude Insights.

Speaker 2:
[32:06] Yeah, thanks, Jim. This was fun. Always happy to jump back on and talk more.

Speaker 1:
[32:10] Awesome.

Speaker 2:
[32:11] Yeah.

Speaker 1:
[32:12] All right. And to our listeners, thanks for listening. Thanks for watching. I hope you enjoyed Matt White, the CEO of Multitude Insights and what they're doing there. Check it out. As he said, I'm going to put the link in our show notes for you to check, click and learn more for yourself. And if you want information on what we talked about today or how to get in touch with Matt, drop me a line at policingmattersatpolice1.com and I will answer every email we get. All right. Well, thanks again and take good care. And I hope to talk to you again real soon.