Retail Crime Prevention | Using AI to Respond to a Growing Threat

Welcome to the Security Cocktail Hour. I'm Joe Patti. I'm Adam Roth. Adam, we're going to get right into it today. We have not one, but two distinguished guests on.

Wait, wait. Redefine what you mean by getting into it. Are we going to fight on the podcast?

No, we're not going to fight. We're doing a podcast. We're doing a show. We're going to do what we're doing. No fluff or anything.

I heard a ding ding and I was ready to get out there.

No, I'm not calling you out or anything, okay? Anyway, so we have Dean Takacs. Dean, welcome.

Hi, everyone. Very much appreciate it.

And Doug Horsting. Doug, how are you?

Good. Good. Good evening, everyone. Thanks for having me. Okay.

Now, these two gentlemen are doing something really interesting and actually very important, too. And they're making a real difference. And it's not exactly cyber security, but it's very related to cyber security. And surprisingly so, it was a little unexpected for me. So it's gonna be, we're gonna have an interesting talk. But first, please, if you're watching on YouTube, like and subscribe. If you're on Spotify or another platform, please follow us, it helps. And because this is the security cocktail hour, and it is actually the evening, we are actually doing this at cocktail hour today. So we can feel good about it.

Well, it's Friday. It's the end of the week, right? So we got to celebrate.

That's right. And it's Friday. So OK, well, have a good weekend, everyone. All right. Who wants to start and tell us what you're all up to? It is very, very cool.

So I guess I'll begin if that's cool with Dean. Is that good?

Yeah.

All right, cool. So about approximately a little over three years ago, I started working with Rite Aid as organized retail prime. law enforcement liaison, basically covering the New York area. And one of the things I was tasked with was trying to combat external theft, retail theft that was taking place. And as many of you can see in the news, if you turn on any news channel, you always see those videos of somebody running in a store and bagging up and running out. And unfortunately, it's become way too common and almost a learned behavior on some levels. What I did was I reached out to NYPD, and it started out in the 4-4 precinct in the South Bronx, New York, for those who don't know, by Yankee Stadium.

So you started out with something really easy, I guess.

Yeah, yeah, yeah. What we did was we just met up with the NYPD's NCO unit, Neighborhood Coordination Unit. And those are guys who are, basically they address conditions, guys and girls in the NYPD who address conditions in an area, stuff like larcenies. And we discussed the steps that we're taking on the inside and asked them for their help and what they can do to address this. Because the problem we have, and most people, most retailers have is, These suspects are in and out of a store in an average of about one to three minutes. Your average 911 call is about a minute to a minute and a half. So they're long gone before, if you include dispatch and then response, and they're long gone before they get there. So the question is, how do we catch them live and in the act? And how do we really throw the book at them, charge them for all the crimes that they've been committing? So one of the things that we started to do is we ran these special operations. also known as blitzes at locations. And before we would do that, we would do our homework and try to research who our top offenders were at the location, the times and days that they're most active or going to hit. And then we would use that data compared to the 911 calls for service. We'd ask the law enforcement to run that and pick the most peak time. And then set up an operation at that location, looking for that individual the flip side and where and then Dean can touch on this a little bit. I'm sure and go into detail is we would case build on these individuals. For example, we don't know their names. So they would come in and whether they're stealing, let's say Duracell batteries. This guy or suspect comes in and he steals Duracell batteries. All the time, he likes Duracell batteries. So we give him a name, Duracell Battery Guy.

Let me jump in here. This is where Joe and I are going to laugh, because part of what we do in cybersecurity is something called APTs or Advanced Persistent Threats. And these APT threat actors, they get these names like Cozy Bear. So now you have Cozy Duracell or something. So on the retail theft side, you're assigning names to these threat actors, right, Joe?

Yeah, we do the same thing. These threat actors are the nation states, Russia, China, North Korea. And they don't know who they are, but they give them a name like that, some kind of clever name. So when we heard you talking about this, we're like, yeah, same thing. But you have some funnier ones. I think.

We've had one guy, you know, they steal city bikes, and he painted it red, so it was red city bike guy. And he would drive his red stolen city bike like as if he owned it, you know. And you would know when he's coming in, this red city bike guy, plaid shorts, crazy hair, you know, I mean, it's just, it's endless. And it's even funnier when the cops are calling it, you know, they're called, Hey, hey, you have any reports on crazy hair, female, you know, chip lady, crazy hair, chip lady, you know, like, yeah, actually have 20 incidents on her. And okay, great.

And oh, that's wild. I was gonna say so. So you're tracking down all these people, there are multiple offenders that do other shoplifting or whatever. And we should probably say you work for a, I guess, we're going to say a chain of retail stores that has a lot of locations.

You actually mentioned the name, Joe.

Cover all this stuff. Oh, you did?

Yeah, right. Right.

Okay, that's right. All right. So you're doing that. And then, so you have the issue of trying to take all these things and make sense of them together and identify these people. And Dean, that's where kind of you come in and where you've been doing a lot with that, right?

So just real quick, the reason for that is because these individuals that come in, if we just get them for the one time, the one, anything under a thousand dollars, it's just a petty larceny slap on the wrist. They're in and out and they're back in the store. So that's the need and really the importance of the case building.

Doug kind of touched on basically the core value prop of detective analytics. It's just connecting the dots at scale and doing it without requiring manual work of investigators.

So, you have detective analytics. It's a system. It's a piece of technology. It's software essentially that the way you described takes a lot of, takes all these case reports, all these incidents and kind of correlates them together and matches them and gives you some intelligence over it.

Yeah. Yeah. And I could go into as much detail or the background story at its core, right? What our core technology is and the core value property is we can connect up to millions of incidences. In fact, we have many retailers that are part of the crime linking network. So we have quite a bit of technology around the space of regardless of how you report, whether it's through our system, whether you're reporting to another system, or if you're capturing data that is more transactional in nature, such as e-commerce fraud, all the way to shoplifting, all of this data, if humans can read over it and make sense of it, if I gave you guys 10 incident reports and I had you read over it, you might be able to see, aha, two of them are a blonde guy with a tattoo on his neck, very likely the same person, right? And three of them are a guy with a ponytail with brown hair. So very likely the same person. The other five are unrelated. But if I said, instead of 10, how about I give you 100? And you say, well, Dean, this is going to take us an hour and a half, two hours to read over. OK. Now what happens if you go to 1,000? A lot of retailers, some of the bigger ones can have tens of thousands, if not hundreds of thousands. And so no human can actually read through this. And so our software does it instantly and is able to connect everything. And we have some clients that have millions and up to tens of millions of transactional records that are tying their investigations back together using the same technology.

I have two questions then. One, Doug, are you able to use his software to build up attribution to one threat actor, a core threat actor, and does their multiple incidents allow you to link a larger theft. So in other words, if it's $500 maybe because you saw them steal 10 things of detergent and then another time they stole 400 packs of toilet paper, does that equal above the limit that allows them to be charged at a higher level? Is that how it works?

So absolutely. So just to dig a little deeper, like for example, these suspects don't just stay at one store. They go to different locations, different stores. And detective analytics actually shows you the stores that it's their most probably going to hit next in the area for our stores, right?

Oh, for your stores. Does it go out to other? So we- Stuff too?

Detective analytics doesn't, because it's a relationship between us and the vendor. Dean can speak to that, but we work together. Organized retail crime managers, I work closely with CBS, Walgreens, we all work together. Ulta Beauty, Stop and Shop, I mean, you name it. Because the individuals that are hitting us are also hitting them. So what we do is we take, if I have, for example, tide guy, which will have a store number or division number attached to it, right? We then create a bolo for tide guy and we'll put, we'll send it out to our stores in the area of where he's going to hit again. So when that individual walks into their store, they can then put that individual name as, oh, that's Todd guy. He came in to our store. And then it's all put together for us.

Let me just jump in there for those who might not know what Bolo is, is be on the lookout. Because I know a lot of our people focus more around cybersecurity and maybe they do watch cop shows and, you know, things like that. But I want them to know what a Bolo is.

Right. And that's where this gets interesting because, you know, what you're doing, and this is why this resonated with me when I first heard about it, what you're doing is actually really similar to what we do in cybersecurity. One of the big things in security is, you know, trying to detect incidents, trying to detect signs of breaking and everything. And so we have, you know, people who are familiar with it know we have SIM, it's called Security Information and Event Management. where we take the logs from everything. And, you know, they can be thousands of events a minute, put them into this big repository. And that's the easy part. The hard part is making sense of it. And for everyone, when you see a lot of things that, you know, the analysts and then the security operations center, the place that looks like mission control with all the screens, getting these alerts in where the machines have figured out, it's like, okay, of these million alerts, these three look like they're related to each other and they're significant. That's a lot of what they're doing. So what you're doing sounded a lot like that to me, that kind of correlation and of unknown actors, because you don't know who they are.

And the parallel there, Joe, is IOCs or the incidents of compromise. So we also do that. And then the other part of it is like you were talking about, Doug, and then you, Dean, we talk about attribution, right? Who is the individual, in this case, we're not talking about cyber, who can we attribute these incidents to? We see the same person maybe in five different stores. If I understand correctly, tell me if I'm wrong, Dean, we do a little bit of a predictive analysis, right? We're going to say, we think this person's going to do this at this time, because when we aggregated all these incidents, it usually happens on a Thursday at 3.59 p.m. This person comes in for their Tide, you know, the Tide repository and steals five Tide things, whatever. Is that right?

Yeah, yeah, that's very correct. We've really made the system become as free-thinking as possible so that it can really mimic how a human investigator would approach that. So, yeah, if you have patterns both in time and in areas and how they hit, it just increases the likelihood that it's the same people, right? And so what winds up happening is, you know, whether or not you actually truly identify the person, or you get some piece of information, or, you know, like Doug mentioned, you just create a nickname for the person. All of a sudden, instead of saying, hey, this is definitely the Tide guy, all of a sudden the system automatically links back to 38 other incidences and says, yep, this is what the Tide guy did. And then you kind of go from there and whatever you need to do. And by the way, the investigators, what we want is we don't want investigations to be kind of the way they were in the old days where it was, you know, someone has to read the report, someone's got to send the Bolo, and everyone's got to read every Bolo. And you just got to get lucky that everyone's memory overlaps on a Bolo and everything else that, you know, you've connected it. And so that's kind of where we've kind of tried to step the industry away from. and say, no, we're going to facilitate the whole information sharing thing from the point of view of an AI that's reading over millions of these reports continuously.

I have a question for Doug though. So Doug, when you're looking at the footage, these video camera footage, can you always show attribution to the same person? Like you say, oh, this is absolutely the Tide guy. Is there a time when you say, wait, I can't really tell this time, maybe we're going to make it as tied guide number two or tied guide woman number two, whatever it is, or you can pretty much do attribution to all of them.

Yeah. So, I mean, with this, if there's any doubt or uncertainty, I don't put it in the case. If there's any uncertainty that it's not one and the same individual. You know, it has to meet meet a bunch of character set certain set of characteristics that it that it is 1 in the same that we're going to be able to positively ID that individual for law enforcement and district attorney purposes and legal proceedings, you know, because. Push comes to shove, we have to go in front of a jury and there are plenty of times you have to go through each 1 of those incidents for video. And if he has a mask in 1 and you can't tell, but. He doesn't have a mask and another, but you can tell, you know, it all depends. But what we do also is we will turn it over. to answer your question, we'll turn it all over to law enforcement and let them decide as well. You know what I mean?

Yeah. You know, that's interesting, because that's so much like what we do in cyber. You know, one of the toughest things is attribution, you know, knowing who it is. And actually tying it to the individual or to the group is extremely difficult. And really, a lot of times, only the government or the really, really high level guys can do it. saying, you know, within your own network or your own sphere, like, you know, these two events are being done by the same actor, you know, to say, like, oh, we got someone trying to break in over here, and we think he was knocking on the door scanning us last week, and that's the same guy. That's really tough to do, but really valuable if you can.

That gets me to my next question for Dean, right? Dean, do you have different customers participating in information sharing across your platform, like let's say, I don't wanna use one specific organization, but like ABC and DEF, can they see each other? Is that allowed information sharing, showing attribution? Like, hey, yes, that tied guy came into this store and then went into the other store somewhere else that wasn't part of the same chain.

Yeah, yeah. So the answer is yes. In fact, we actually have patents filed in that space for an invention that does that. And so there's an interesting thing that the industry's tried to solve here, right? And it goes at odds with what legal teams want. On one side, right, you can always just take all of your data and say, hey, here's a shared repository. Everyone from all, you know, 300 retail companies can have access to it. All it takes is one bad actor and all of your data is compromised. All it takes is one mistake and all your data is compromised, right? So legal teams do not want to give direct access of their data to people outside the company. And it's not just the compromised data. If someone makes a bad report or says something obscene on a report, they cannot have that risk of vacuum, you know, public knowledge.

Yeah. So within your system, each client's data is segregated.

Yes. And this is the key thing. So with the Detective Analytics Prime Linking Network, we have retailers reporting through our system, through different case management systems across the board. And what we've invented is a way to encrypt that information. and just share the connected incident IDs. So this means if I'm in an investigation at, you know, a company ABC, I see all of the data. Let's say there's 12 incidences linked to two suspects that are working together. I can see those 12, and if it so happens that that investigation is linked to another investigation at company XYZ for, let's say, five incidences, I see the five incident numbers, and I see the dates of when they occurred, and I have a contact information from that company. So at any time on a per-investigation basis, we can collaborate, but you never risk actually sharing your sensitive data. I don't even know the names of the people or why they're connecting, but the system knows.

So at some point, Doug, you might get a call from Dean's people and say, hey, we think that someone who's hitting this other company is hitting you too. Is that kind of the way it works?

Yeah, absolutely. And then, I mean, I let Dean still take I don't want to steal the thunder on this 1, but for the, for the New York City mayor's retail theft task force and for NYPD, he's, he's running a program where he's inviting retailers in the New York City area. if they want to be a part of it, to sign a release and share that information to NYPD, strictly to NYPD, for investigative purposes. And so just picture it this way. So as a detective or a lieutenant, a sergeant, however NYPD decides to set it up, on one screen they'll have all their information on the suspects, and on another screen they'll be able to see all the retailers who are participating, the information on the suspects, and say, wow, this guy has He's got five open cases with this company, six with this company, seven with that company, and now it paints the whole big picture. And now you're putting together an entire case to actually get some jail time for these individuals. Do you want at all?

Well, okay. I'm going to ask this and hope you can answer. This is when we started thinking about, that's a lot of information about people. And as Dean, like you say, someone gets in, I'm hoping you got really good cybersecurity because it's a problem if someone, especially an insider, gets into it. But one of the things that you guys have to deal with that we don't, is stuff like civil rights. Is this basically what detectives are doing anyway and what you'd normally share with them? You're just being able to do it on a larger, I guess, with higher quality rather than less guessing or have to do less investigation?

Yeah, it's much quicker as well. It does a lot of... It basically puts it all together for you. and for them. So, in other words, it'll put all the stuff together for TideGuide, all their reports, all their video, all everything, and I just, all I would have to do is just send the link, and it's all right there available at the link. So, if we're doing an, and the other flip side is when you ask for the, it does heat maps of when the TideGuide's most likely to hit again. We set up an operation during that time, we locate the individual, catch them live, committing the crime live, and then we're able to go back And pull up everything that we've had in the last X amount of time period on Todd guy with video and Incident and police report numbers and so forth and then it makes it I mean, it's it's done in real time So yeah, it makes it a lot easier and quicker and faster.

So yes, how often is that happening? Like how often? Does it make a prediction? And I'm not trying to like, you know, I don't know how to say this, but I might like saying Dean's software is not doing the work. But does it happen that often where it's so good that you're able to say, oh, we expect him on Thursday, almost like in a calendar item. Thursday, 2 p.m., we'll see him then. or see her then, and they literally walk in the door, boom, they're arrested.

Yeah. Yeah, I mean, I don't... Yeah, does it seem magical, you know? Yeah, pretty much. I mean, some of them, some suspects do operate with a rhyme and a reason and some just don't, you know, it's just kind of random. But, I mean, also to make sure that we're utilizing our resources correctly, we do, when we do set up these operations, we do ask the law enforcement agencies that we're working with to run their calls for service to see how they match. For example, if If we're getting hit between the hours of 4 to 12, 100 times versus the hours of 8 to 4, 30 times, let's do an operation from 4 to 12 because we're going to get more suspects and send the message out there as well.

So Dean, and I know some of this may be a little bit proprietary, but to the extent you can talk about it, you know, What kind of attributes and things are you looking for and how are you correlating them? I mean, is there a basic kind of thing? You mentioned an AI. Is there an AI that's looking at people's behavior, patterns, other AI stuff we probably don't understand? How do you put it together?

We want you to put the actual algorithm on our podcast so everybody can see.

Yeah, please do.

Yeah, I'll open source it. I'll open source it. And then on top of that, what I'll do is I'll just sign away all rights to you guys and that's it. We'll make it clear.

You didn't read that part on the release form? No, there's no release form.

I've received the worst forms as a mistake to sign off on. So, you guys look at a laugh, I could talk about that for a while. Basically, what we've done, so think about this, this is a problem that DA has been focused on for something like seven years now. And so, what you can't do, and a lot of these case management systems do it, is they're just like, you have to create these specific fields and they'll build everything around these couple of fields. What's the color of the hair? What's the color of the eyes? What's the color of the car?

Your customer tells you that?

No, so we don't do that. That's actually what we don't do. We don't actually build to pre-selected features. In fact, what we've designed is a system that actually just looks at everything in its entirety. So if you abstract away from the creation of hey, what's the height of the person stated here versus the height of someone else that match that together? We just say, no, just read the damn incident report, literally AI, read the incident report, extract all the information that is required from everything. And then when you do that, it turns out you can have some really cool results. And if you think about it, It's, we don't operate as Chats UPT. It's a very isolated type of connecting. But if you think about kind of the results you get, once you abstract away from manually coding stuff up and having to look at features, you wind up having really cool results. So some stuff we'll get is, we'll get matches when it's a woman and a man, and it's actually brother and sister working together. We'll get some interesting stuff where, Yeah, and it's designed to do that because the system has seen millions upon millions of crimes across, you know, mostly retail, but other industries as well. And it knows how to recognize similarities between them.

So if I can ask then, you know, talking about the patterns and all, you said it's not JAT-GBT, so it's not generative AI, which kind of doesn't surprise me, just because you haven't had time to incorporate it. Is it basically the more traditional AI, like machine learning, those sorts of techniques?

Yeah, so we're using deep learning. We're using deep learning algorithms applied to this, and there's actually a lot of proprietary approaches to how to efficiently do this, right? How do you efficiently manage a whole set of investigations for an organization without your investigators having to build that? So one thing I'll say is, like, when retailers go from using one of our competitors, which is traditionally, like, a case management system, where you're sending BOLOs and you have to manually build everything, they just connect up to Detective Analytics and all of a sudden they have 9,000 investigations for an average size of 30-something apiece. And they're like, what just happened? And they say, well, that would have took us two years to build. What we've really built is a system that's managing investigations in real time by looking at incoming incidences. That's kind of the core of it.

So I want to jump in on Doug now. So Doug, how do you utilize Detective Analytics? Is it in a computer, on a PC, on the desk? Are you putting it in stores? Is it on tablets or is it more your management tool?

So Detective Analytics for us is something just that I, it's a program that I use on a computer. And right, it has a special operations center that they're using and monitoring it as well. Nice.

So, Dean, I'm guessing it's in the cloud, right? It must be.

Oh, yeah. Yeah, it's in the cloud. I mean, you know, it's funny. Our platform has expanded significantly, and I'll tell you, when we initially created it, the thought was that every single provider of case management solutions would play nice and would love something like that, as you could have, as in no, because it's, you know, we've actually gotten pushback from the same organization, same solution providers, who market the idea of collaboration. And so, you know, the core of what I'm talking about, about DA, and we've really since expanded into case managing and reporting, but what Doug's talking about, the pure intelligence, we can reduce down to just a handful of people at a major retailer can manage all of your investigations, which traditionally would take the mobilization of hundreds of your in-store folks. and really just a whole army going to Bolo Alert, Bolo Alert. So it's like, instead of giving everyone a better shovel, we went ahead and invented a construction equipment that's a digger, right? So that was kind of the thought process. Yeah.

So I wanna know this, Doug. Have you been able to use detective analytics to show attribution to an individual in more than one state, maybe even multiple states, like a crew?

Yeah, so there are organized retail crime crews that travel multi states and same thing. We give them case names. It's just the same thing on a different on a bigger on a bigger scale. And it tracks them and it also includes the more information again, just like any case management system. It's only as good as the information provided and put in, right? So the more information that's put in, the better chance we have tracking these individuals and these suspects. Once you start talking mobile suspects that are traveling different states, you start talking about rental cars, license plates, all that information. And once that stuff's in a report, yeah, it picks it up and it connects it as well.

I'm going to ask this question. I don't know if you can answer it, but Is there an attribution that our audience will say from a value-wise, oh my God, like 100,000, a million, 10,000, is it a crazy number?

I mean, so I can say this, for New York alone, I'll give you one example. For New York alone, it was hundreds of thousands of dollars for all three pharmacy retailers, hundreds of thousands of dollars. And it was by one crew that we created a name for, and the other two retailers, pharmacy retailers as well, used the same name. And we were tracking them only through New York, but they hit all five boroughs. They were hit all five boroughs, hundreds of thousands of dollars. And the way we caught that, one of those individuals, unfortunately, is through setting up a blitz operation. We set up a blitz operation in the Bronx and they came and they hit we would have gotten all 3 of them that day. But the 2 of them, they ran, they dropped the stuff they got in the car. There was product all over Mars Park Avenue in the Bronx. We're picking it up. The guys, the guys are flying all over the place, but we got the 1 guy. And so, and we're able to, you know, then it comes into. Once attorneys get involved in just attorneys get involved in New York, especially, as you know, each borough is its own thing and its own entity. So it's got its own and its own what they're going to take what they're not going to take and so forth. But, yeah, just even within New York City, the 5 boroughs tracking that 5 boroughs, 1 crew, all 3 pharmacy retailers. It was hundreds of thousands of dollars.

Wow. That's so awesome because, you know, being in cyber, I mean, for us to actually catch someone and have them, you know, prosecuted or picked up, God, it's rare. It's really rare. And a lot of the really nasty stuff like this ransomware is all happening from overseas. You can't touch these people.

I mean, it's just not realistic. Yeah. Ranter wears literally sometimes kids, right, Joe? And they're making millions of dollars.

Well, it can be kids, but also the really bad ones are the operations. They are corporations. It's not just organized crime. It's really crime corporations. It's unbelievable the way they run things, but it's all overseas. But it is so amazing to me how what you're doing from the data collection and the intelligence and the attribution is so much like what we're doing in cyber. Because like, Dean, you were talking about, having, you know, doing the work of thousands of people or thousands of hours or whatever quickly. You know, in cyber also, when you have that SOC and you have all these alerts and all these things, the goal is to be able to deal with them effectively with as few people as possible.

And that's my next question for Dean. So, Dean, we know what Doug is telling us in hundreds of thousands. That would mean that your software, your application, Detective Analytics is probably and have been able to solve millions and millions of dollars in losses and have attribution to multiple threat actors. Do you know what that number might be today or no?

Yeah, it's over.

It's in the hundreds of millions, the total attributions to the system, yeah.

I mean, you have some of these crews, so like certain areas, right? Like pharmacy in a city is going to be isolated typically, right? But when you get some of these higher end retailers, and some of these traveling crews, I mean, we have some investigations where it's 20 plus people that are connected together. And so like we literally at any moment are tracking these massive crews, they've hit 200 and something times. The best part is no investigators are sitting there building it together. I mean, managing an investigation that size is very difficult. So our system is designed that you can just literally monitor the top groups as the data's flowing in, The AI is actually recognizing what investigation they're connected to and updating the investigations in real time. And then you just monitor them and you say, hey, okay, I want to get this person. Like Doug had mentioned, right? He'll, you know, they'll be tracking the Todd guy or whoever they may be. All of a sudden, boom, you can just say, now I want to go get him. He got too good. You know, he hit us 38 times. That's it. It's time. So, yeah, I mean, I'll say one thing too, guys. One thing that was really fundamentally changing to me was very early on, one of our customers, They said, by the way, these folks that were hitting up in Chicago, your system connected them back to Florida. We would have never known about it. No one would have ever communicated. And they said that actually bumped their thresholds up to a significant felony. So they're actually going to go do time. And that was really kind of fascinating. I know you gentlemen are involved with law enforcement and whatnot. Now, from a software engineer's point of view slash a math guy, it was very weird to say, wow, we're actually having an impact on society in such a way that Some of these really prolific criminals are now going away and doing time. And so I just want to say, like, I've tasted the slightest bit of what law enforcement feels like from, you know, doing a job and solving crimes. It's kind of been a little bit, you know, life-altering in a sense.

Well, isn't that, you're kind of attacking a big problem, because as I understand it, and Doug, I think you had mentioned it too, you know, a lot of these crimes, individually, they're tiny, they're not prosecutable, or they're not very serious or whatever. But when you're able to bundle them together, then it seems like you're finding people who are, it's almost like they're approaches. If we do so many small things, even if we get caught on one of them, they'll never put it together. And I'll never forget NAMP for the whole enterprise, but it sounds like now you're able to see what they're doing on a wider scale.

Yeah, and to be completely honest, and to give credit where credit really is due, we couldn't do any of this without law enforcement. And a huge thanks to them, especially to NYPD. I can go down a long list. Worcester County Police, Yonkers Police, Osley Police. So when I started up until July of this year, in New York City alone, five boroughs, just NYPD, we hit 7,000 arrests. 7,000 arrests. And that includes blitz operations and something that we call off-site arrests as well. So not all of our arrests take place during a blitz operation. We'll case build on the individual, like we said, crazy short guy. And crazy short guy then gets assigned a detective in NYPD. And based on what the tools and resources they have, they would be able to do ID procedures and ID crazy short guy. And then they can go out and issue I-cards and a warrant for their arrest. And then they go out and get them as well. So we had 7 over a little over 7000. And it's a huge thanks to law enforcement across the board throughout New York state. They're doing it. Uh, chief Cecil of Syracuse police on board. We have stores up there that, um. Just on top of it. Rochester police, Westchester County, Ardley police. And the other thing too, I'd like to say is, once you move out of New York or the borough area, right, you have different agencies in different counties and that all technically fall under New York State, but a lot of them don't really talk to each other in real time as well, where we'll have the information and say, hey, listen, even if somebody gets arrested in an Ardley location, We have information of them hitting our stores in the Bronx with NYPD. So we got, they're being held at Ardley police. And then we make a phone call to the NYPD and say, hey, listen, that suspect that we have 20 open cases on from the Bronx, they have them in custody in Ardley for stealing. And then they'll be able to pick them up and take them and vice versa. And then those charges will be met in different jurisdictions. Very nice.

Now, I got to think that I know law enforcement deserves a lot of credit, but I suspect that one of the reasons that they're doing this and cooperating with you is that you must have a very good relationship and a very good track record with them, because that's something that we talked about in an earlier episode with Jennifer Gold, who she was talking about in Cyber, It was so important, the collaboration between the government and the private industry. It was so important to build those relationships so that when something happens, they know how to talk and there's a level of trust. It sounds like you're way down that road.

Thanks. Thanks so much. I mean, I appreciate that. The thing is that retail speaks one language, law enforcement speaks another language, and you got to kind of I kind of bridge those and bring those two those two languages together to be one. The other reality, the situation that we look at is, I'm sure you gentlemen are familiar with the broken window theory, right, in the criminal justice world. It's so true. I mean, out of the 7,000 arrests, a large percentage, if not all, have a tremendous amount of criminal history with some serious, heavy crimes in the background. So we're taking guys off – bad guys off the streets who have open warrants for robbery. We have – we picked up – we did a week-long operation in the Bronx once, and we were able to pick up a guy. Who have who they were looking for for a bank robbery pattern out of Manhattan. And we were able to get him for stealing from from several of our stores in the Bronx. So, you know, it goes back to law enforcement appreciates the fact that we're taking the bad guys off the street, the common. Misconception is they're just shoplifters. They're not just shoplifters That's the shoplifting is when you're just getting for yourself for personal use or these people are doing it for resale And to try to make a profit and make some money off of it, but that sounds amazing I mean did do you mean that you have guys who are like?

They're like, you know bank robbers they're doing serious stuff and they do that they go and hide they're good but then I don't know, and they need a candy bar there. They go and, you know, steal, you know, grab a grab a beer out of out of the case. Yeah. In a rioting. You're getting them because of that. They must feel ridiculous on that.

It's taking bags, Joe, like plastic bags, garbage bags. And they're literally walking down the aisle and taking 200 shampoos. Look, I get it.

I mean, these are the pros even in that.

Some people would even say that organized retail crime is the new drug dealing because the laws, the change in the laws, there's less consequence for your action if you do this and there's a lot of money to be made out there. It's a very large industry, where they sell it on social platforms or in brick-and-mortar stores, some mom-and-pop businesses, or they'll go down the street and pop up a table and sell it without a license, without a resale license. So yeah, it's the new – some say it's the new – It's near drug dealing.

And I think it's safe to say the mother that goes into a store that literally does not have money and is taking a gallon of milk is not going to show up on Detective Analytics because they're doing it so often. Whereas the people that are literally walking in there taking 200 things of shampoo that's expensive or cosmetics. A lot of it's cosmetics, it's makeup, it's things that people can sell very easily and they'll sell them literally on the streets and try to sell that same product somewhere else. So it's very organized, but I don't think Doug or Dean is chasing down the person that's literally, again, I don't condone it, but the person that's trying to get milk or cookies for somebody that literally is hungry, they'll handle that, but you're focusing on the 20% or less that's creating 80% of the robberies or the thefts.

Correct.

That's correct. Yeah. I mean, the truth of the matter is, me personally, If someone can't feed themselves, I don't want to say I condone it, but I certainly feel it falls in a different category of the other stuff. I mean, the goal of DA, it's really, we're playing a high-level game of cops and robbers with professionals. I mean, some of these guys, you hit the numbers of hundreds of thousands or even some of them millions of dollars, you have really complex schemes going on. I mean, society just can't handle that, right? And so you've chosen that profession. And we have good guys as well, and we're empowering the good guys to kind of prioritize you. And so, yeah, if someone comes in and takes a bag of potato chips, it's not helping your ROI to go after and investigate the guy. And I don't, listen, I don't speak for any LP group, but the truth is the guy taking a bag of potato chips to feed his kids or a sandwich or something is nowhere near the guy going in and taking a garbage bag, like you mentioned, 20 shampoos or everything of lipstick and going reselling them, and who's doing it 40 times. Or worse yet, if I'm someone who's paying less than people that aren't well off and are living on the street, and I say, here's a list of product, here's 10 people, I give them a list of product, I say, every week you meet me here, I'll pay you exactly what's on that list. you wanna uncover that person, right? Because that person is perpetuating a lot of problems in society. And so that's really what DA is focused on. It's focusing on the worst offenders who for most purposes, they know what game they're playing. They're playing a very high risk game anyway. Eventually they're gonna get caught.

And it's not a victimless crime, because I'm paying for that. Joe's paying for it. You're paying for it, Dean. And so are you, Doug. We're all paying for this. When an organization has to raise its prices because, you know, that external theft is happening so often and so flagrant, Well, I mean literally you see videos all the time on the internet guys walking out woman's walking out with a bag of sneakers From uh from a retail store Like they they took 20 pairs of sneakers or even worse they're going I'm not saying it's not worse that it happens to like small stores or chains of uh of uh drug stores, but when they're walking in taking 50 Bags worth $5,000 each it is absurd It is not right and it's causing all of us to pay You know higher prices to everything that we purchase Right and the other thing you want to talk about dollar value, right?

I Safety and security of our staff and employees and our customers is number one. And these individuals doing these things are not good people. They're dangerous people. And our job is to get them off the streets as quickly as possible with law enforcement. With that being said, you look at a child online with his mother, right? And he's standing there and she's standing there with her one bottle of Tide and maybe her thing of milk, and they're online together. They're watching the 2 or 3 other people fill up garbage bags full of tide and walk out. And now that that child says his mother, mommy, what's going on? And she's standing there paying and, you know. It's it's it's become too. the norm. It's become to the norm in metro areas. And that's why a lot of law enforcement agencies have stepped up and have really are ready to go above and beyond to address the situation.

So we spoke about how there's a loss, there's loss in value. You know, hundreds of thousands of dollars of items are stolen out of stores. Do we track, I want to say we, do you track the actual incidents where maybe somebody came behind the counter and hit the store clerk or the person or the cashier to grab stuff, maybe even money. Do you track the physical assaults against the store employees?

So, yes, I mean, first of all, that, like I said, the safety of our employees and of our customers is number 1. If anybody gets even threatened, verbally threatened, let alone if there's any action that takes priority, it's all hands on deck until that person's in custody. And I can actually give you a most recent example of where that happened in Manhattan. in the confines of the 3-3 precinct. And I have to give a kudos to them, Special Operations Lieutenant Caprino, Officer O'Shea, the commanding officer of the 33rd precinct. There was a man that came in and said some threatening stuff. And that man was in custody almost immediately. He was found and in custody almost immediately. And with that being said, was able, and Char, was able to go back in He was in also for shoplifting. He was a shoplifter past crimes, a shoplifter, and I was able to pull up the case management of what we had on him in the past and send it over to law enforcement as well.

So God bless your employees. God bless you, Doug, and you, Dean, and law enforcement and NYPD for acting so swiftly to ensure the safety of everybody. I mean, that's what we're really talking about. Look, yeah, it really sucks to lose a monetary value, but that does not compare to the safety of anybody. Even law enforcement, everybody can get injured or hurt by these people that are stealing items.

So Dean, you sound like you have a huge amount of data. And I know that when cyber people see it, and even when, you know, IT people see it, they're going to say, that's valuable. And we can learn a lot from that. So are you also doing, um, things like, uh, you know, trending that, that kind of, that kind of analysis, macro sort of, uh, stuff.

So, you know, I mean, the, the key thing, right, is that the trust of your, of your customers is, uh, absolutely critical, right? And so, we never ever share any piece of information, a single piece out of our customers' data sets. Within there, customers, yes, we do a lot of tracking for them, a lot of high-level stuff. Do we ever do any kind of thing together? No. We have a very restricted approach to collaboration, which is the Crime Looking Network, actually. That's a very rigid implementation of of facilitating information sharing in a very encrypted, very precise and a very secure manner. But now, do we do that? Yes. Is there a trend analysis happening? Yeah, and you do see it. I mean, a lot of what we do is we'll find anomaly detections, and that's actually a very critical way that TA is able to pick up when there's new suspects. When you see certain locations that just go out of control, starting to get hit, You can either say that, well, randomly, every criminal happens to be coming home from, you know, vacation, or it's likely one crew doing it, right? And when you see 20, you know, not 20, but you see three, four, five different people repetitively just all of a sudden show up in an area, what does that tell you? It tells you they're probably working together. And what we've seen in the past is you wind up connecting the dots on these crews and you get one of them, right? And as soon as you get one of them, they share. They share information. Why is that? Because they don't have to go to jail. and they're happy to get all their friends to go to jail in their place. And that's one of the funny parts about this ORC thing is that there's not a lot of sacred codes of loyalty amongst them, at least from what I've seen. I'm sure Doug could probably speak more of that from the bigger crews.

I guess what I think Joe was asking, and I think you answered it anyway, was we would be interested if you were able to anonymize the data and say, guess what? In 2023, our product was in place and we found out there was 55,336 retail crimes that happened in the city of New York, something like that. But I obviously understand that even the anonymization does not allow you to probably give that out.

But one of the things we actually could do, and it's funny you bring up New York, because New York we're working closely with, is retailers that decide as a subset that they want to collaborate together, we facilitate that, right? So the idea of working with the NYPD to allow them direct access, one of the things we're doing in that space is allowing retailers to sign up and say, hey, we want to allow closer cooperation with the NYPD. And by doing that, the NYPD therefore will have access to the aggregate data of that information, which is very powerful. You know, just having them be able to see crimes, let's say, immediately within one retailer is huge. But what we're talking about is actually transcending the different retailers across their different case management systems. And that's really where we come in and say, we don't care what you're using. We're not trying to step on the toes of any other software providers. We'll actually work right on top of that together. And that's actually a ton of our success comes in just with that. And by virtue of what you had asked, NYPD will be able to see aggregate statistics across any retailers that wanna participate. Now, the hard part is getting retailers that participate en masse, but it's something that over time, they've been kind of coming together on.

Well, that's great. And I'll tell you, it is, you know, we've done a couple episodes where you talk about, you know, surveillance capitalism and all the, you know, data data collection practices and everything. And I mean, to have, you know, someone who's collecting a massive amount of data on people and especially something, you know, sensitive, you know, law enforcement related stuff. and actually thinking about protecting it and worried about it and taking some steps. Dean, you're a superstar as far as that goes, really. We don't see that too often in our business. And that is good to see, especially since it's in such an important area, too.

I know we're coming to the last call, John. I just have one other question for Dean. Is there any other hints you can give us at some of the features that might be coming out that we should look forward to?

Oh, yeah. I mean, I'll tell you guys. So we're coming out with our mobile web app. We're teasing the idea of launching it across the country. Now, one of the really cool aspects is that DA is actually being used now on digital crime as well. So we can talk more about that. And one of the funny things is that we kind of had our hand forced is the whole case managing and reporting. So initially, as a founder, I said, they're all going to want to work nice with us. They all want to collaborate. It's the mission of all these solutions providers. And every day, it seems they get less likely to collaborate unless they're all making money. So I said, you know what? Let's just create our own case management system. So we're actually coming out with a really, really top state-of-the-art case management system we just launched. And that's really going to change the game, as well as our new mobile app for sharing of retail intelligence crime. And so looking forward to that. And that's, you know, what I can share with you guys. So thank you.

Thank you. Hey, Doug, any other words here?

No, just thank you. Thank you very much for the time and for having us on the on your show, we really appreciate it. And it's a huge thank you to all my AP teams out there. Really, everything we do is a team effort. Our store teams, our AP teams, and our law enforcement partners. I'm just a small piece in a big picture, so thank you.

Yeah, it's safe to say that people don't realize how much LP, the loss prevention people do behind the scenes. You don't see them, you're not supposed to see them, but they're working very hard out there.

Yeah. And it is great to see that you're, I know it's a big problem, but you're, but you're making progress. You know, one of the things we always try to do at the end of the show is sometimes, and we're talking about, you know, information security and cyber, and we talk about some stuff where the situation is so bad, it's really discouraging sometimes. And we got to say, okay, let's find the silver lining in this. But it sounds like you've got the silver lining is very clear. You're really seeing results. And that is, and it does help everyone. So, you know, well done. That's absolutely great to see. Thank you.

I want to say one thing about, you know, Doug, for what Doug has done in the city is very atypical. Like, you do not see this across the country, and it's a model that really, really should be followed. Doug's ability to combine, you know, use of technology and build these amazing relationships to the point of Doug's doing stuff that, you know, you don't see, I don't want to say it too loud, You don't see many vice presidents doing the stuff that Doug is doing. I think Doug's work, alongside of what he's done, the hard work of the police, the hard work of his associates as well, but you talk of a number like thousands, 7,000 people, these are people that would have committed more crimes. How many assaults did you prevent, Doug? How many potentially homicides did you prevent because of this? I mean, you think about the impact that some of this stuff's happening and Doug's just a phenomenal example of this. So I think we might not realize just how much of a talent we have on the show with us. I just wanted to point that out and it really should be followed by folks in the industry.

Thanks, Dean. I appreciate that. Again, I'm just a small picture, small part of the big picture, but I appreciate your kind words. Thank you so much. And again, a big, big credit goes to law enforcement partners, especially partners over at NYPD. Too often we pick up a newspaper or turn on the local news, and we always have something bad to say, but they're people too, and they're out there doing the best that they can with what they have, and they're doing a great job from what I see on a daily basis.

Absolutely. God bless them, white people. Yes, and that's great to hear, because we count on them. So, gentlemen, this has been absolutely fascinating. You know, sometimes we do get a little bit outside the realm of cyber, This is very, very interesting. It is amazing to see what kind of similarities they are. And it sounds like you're making a lot of progress. So thank you. And Adam, you want the last word?

The last word is, guys, thank you. I know we kind of met. you know, through social media, and this was a really good find, and I really appreciate it. I felt like, you know, a lot of what I've done in my life, this makes a lot of sense and adds to everything I've done. Thank you.

And I'll say one more thing. I really appreciate it, guys. Thank you so much. I absolutely love NYC, NYPD, and I also want to say I think New Jersey is just a little bit better than NYC.

Make sure you got it back on that.

Well, you know, I have been known to say that Adams in Staten Island because he just can't deal with Jersey. I always go to see him. He never comes out of here.

So I, you know, look, you know what? I shouldn't say anything bad about New Jersey because the natural progression is you move from Brooklyn to Staten Island to New Jersey. So maybe I might end up there one day.

I'll keep an eye out for a house for you. Okay.

All right.

Thanks, everyone. Thanks, everyone, for listening.

Thank you. Thanks, guys.

Okay. Take care, everyone.