Guardian Alarm Systems’ new central station employs 20 operators who monitor just under 8,000 subscribers. Through its integration with CHeKT, all video alarms are processed through the MicroKey automation software in the central station.
By Laura Stepanek, SDM Contributing Writer
One of the most dynamic trends in security today is the use of surveillance cameras as sensors that generate intrusion alarms, which are then transmitted to a central monitoring station to be acted upon in real time. Video analytics — especially analytics powered by artificial intelligence (AI) — are driving this trend, because they are capable of precise detection with many fewer false alarms than the video motion detection technology of earlier camera generations. Because AI-powered video analytics are adept at ignoring unwanted alarms, it makes them an ideal type of alert to be processed by a central station. Professional monitoring ensures that someone always sees video alarms, instead of only being stored as data to be reviewed later.
With monitored video alarms, a whole new realm of monitoring now is before the industry: one that focuses on detection, deterrence and prevention of crimes. What’s more, where central station monitoring is involved, so are security integrators and dealers with the potential for them to earn recurring revenue from video alarms.
The concept is different than video verification, in which cameras commonly are used in tandem with alarm sensors (passive infrared detectors, or PIRs, photoelectric beams, magnetic contacts, and glass-breakage sensors) to corroborate an alarm. Video alarm verification, when first introduced a dozen or more years ago, is still well-regarded by law enforcement as a method for reducing false alarms and prioritizing real alarms.
This new category, called video alarm monitoring or proactive video monitoring, relies on the intervention of monitoring center operators to act on an alert as it unfolds. Actions could include locking or unlocking doors and gates, turning on lights, and/or activating audio talk-down — the latter being most common. When the analytics, which could be edge-based on the camera, server-based, or in the cloud, detect a zone violation — meaning a person or a vehicle has crossed a virtual boundary within a scene — the operator receiving the alert views the video feed and decides whether further action should be taken. By executing an alarm response such as audio talk-down, the central station has a good chance of deterring the perpetrator from going any further with their criminal intent.
“Proactive video monitoring is very different from traditional video monitoring with video verification,” explains Justin Wilmas, president of Netwatch North America, Lake Forest, Calif., a company in the Netwatch Group which also includes NMC, a wholesale monitoring center. “Video verification is when an alarm comes into a central station, [operators] pull up the video footage and can verify that it’s actually an alarm that they need to have law enforcement respond to. And that becomes a verified alarm, which should get priority when it’s sent to law enforcement for dispatch.” Netwatch is a pioneer in the use of proactive video monitoring; its platform provides customers with intelligent remote video monitoring to proactively detect and prevent incidents before they begin.
“When we’re voicing down and we’re able to deter the incident 98 percent of the time from happening, that’s proactive,” Wilmas says. “[With] video verification, the alarm has come in, they see a bad guy doing bad things. That bad guy is still going to do the bad things up until the point that law enforcement shows up. And that timeframe can [vary], depending on how busy the jurisdiction is, if they even show up at all. The concept of proactive video monitoring is to deter, de-escalate, and prevent the event from happening before it happens.”
Displayed in Guardian Alarm Systems’ central station is a poster recognizing the serious responsibility operators have of responding to alarms as they unfold and potentially preventing crimes.
Jim Kopplin, senior vice president, centralized operations, at ADT Commercial, Irving, Texas, concurs with Wilmas about the deterrent effect. “Leveraging AI to detect a person and triggering a response to deter the person from breaking into the building helps reduce property damage and can help stop break-ins before they occur. These deterrence events can then be mechanized to trigger lighting and/or automated message alerts, or notify call center agents who can directly engage through audio solutions,” Kopplin says.
ADT Commercial offers video monitoring with analytics in a limited configuration, Kopplin says, adding that ADT is actively working to expand its capabilities to offer the solution to a broader range of customers.
It’s their extraordinary ability to filter out unwanted alarms that makes AI-based analytics an enabler of video alarm monitoring. Not only can they accurately identify that an alarm was caused by a person — not an animal or something blowing in the wind — they also can provide ample details about each detection. As an example, an AI analytic can identify the number of people violating a secured zone, what they are wearing or carrying, the direction they are moving, whether they are loitering, and if they begin to position themselves in a way that could indicate a potential crime, such as getting underneath a car to steal a catalytic converter.
Video analytics, especially those enhanced by artificial intelligence (AI), can be challenging for a security dealer or security integrator to set up for their customer. To achieve optimal performance from the technology in the central station, there are many things to consider — some of which may be contrary to your experience.
“My message to dealers is if you’ve never done this before, pick something that is workable and get good at that first,” says Morgan Hertel, vice president of technology and innovation, Rapid Response Monitoring, Syracuse, N.Y., and president of The Monitoring Association. “So you want to do video verification; if you do a lot of intrusion, pick a camera system that’s compatible and run with it. If that works well for you, go off and do something else. If you’re the kind of commercial company that does a lot of outside work, then pick a camera that has a good analytic in it and start using that to do detection outdoors. A lot of money, a lot of market share there. But get good at that.”
Hertel emphasizes that security companies should take it one step at a time. “Don’t just jump in and start this gigantic video program until you really get good at it, because there are a ton of processes that have to be done first: the sales process, the installation, the verification of the installation, how to set it up for monitoring. But also don’t ignore this, because you’re going to get steamrolled by it if you don’t.”
David Erickson of Guardian Alarm Systems says perhaps the one area dealers and integrators need to be mindful of is ensuring that their detection areas are defined as ones that a client wants protected. “When you get that alarm signal as an operator, you have a responsibility to either get in touch with the client or dispatch the police. So, when an operator receives a video alert of activity that is happening next door to the protected property because the detection area is too far-reaching, that will create a lot of noise for the operator,” he says. It also could create unnecessary expense for the integrator if the monitoring center charges by operator time.
“Many integrators and dealers will set up these very broad detection areas — I’m looking for people, but the camera is overlooking a sidewalk or a public area. And every time there’s a person there, that’s going to the center as a real-time alert. Think about an operator at a center who now sees someone on the other side of the fence or maybe they see a couple walking down the sidewalk; the dealer thinks they’re protecting the client, but really what they’re doing is creating unnecessary noise as far as alarm traffic and potentially doing a disservice to the client because they’re just getting too much information,” Erickson describes.
Doing a walk test of the detection area is compulsory, says Kevin Lehan, director of national sales and marketing, EMERgency24, Des Plaines, Ill. After set up, study what your cameras are seeing, making sure that your perimeter lines actually are on the inside of the sidewalk instead of the outside of the sidewalk, he explains. “Make sure that your cameras aren’t pointing at things that are going to cause unnecessary triggers, because you are paying per activation after you get through the basic threshold, three videos after a month,” he says.
“Only when you’re ready and once you have your system calibrated so you know you’re not generating bogus events, then you can turn on the system,” Lehan says.
When installing the video cameras for a monitored system, more important than having a pleasing image is ensuring the best detection possible. “I’ve seen a lot of well-deployed ones and I’ve seen a lot of not-so-well deployed installations for monitoring purposes,” says Nik Gagvani at Kastle Systems. “Consider what the camera is going to see not just at the time you’re installing it, which might be in daylight, but what it might look like at night.”
In addition to positioning the camera for a well-contrasted image under all lighting conditions, confirm nothing is blocking the detection area. Consider seasonality, meaning that if your installation is occurring in the winter when there are no leaves on the trees, spring leaves could block half the view of the scene, Gagvani says.
“Even the best analytic systems are not plug-and-play. They require a significant level of adjustment and training, and it is a time and labor investment to get your systems to perform to their maximum potential,” says Anthony Iannone at Affiliated Monitoring.
He offers two important pieces of advice. First, learn to manage customers’ expectations. Affiliated Monitoring once had a partner that wanted them to monitor a very busy public parking area, where people were present 24/7/365, to look for muggers, thieves and people damaging vehicles. “The problem is, I can have an analytic that detects a vehicle or I can have an analytic that detects a person, but we don’t quite yet have an analytic that can detect a thief or a mugger. I think that’s still a few years away. Managing customer expectations and educating the customers about what is and isn’t possible is very important,” Iannone says.
Second, don’t skip all of the necessary steps required by an AI analytic installation. Even the best analytic will not work well without careful thought to the environment, and then careful tuning and set up of the analytic. It takes patience, perhaps even a couple of days at the site, Iannone says. “Skipping those steps is a sure way to have a less positive customer experience, driven by additional false alarms, or potentially missing real activity,” he says.
Ray Robertson, senior vice president, commercial product and innovation at ADT Commercial, advises that optimal performance can be achieved by accurately matching the analytics to the items that should be monitored. “This can be accomplished by conducting testing of the video and analytics solutions to be used in a controlled space, where monitoring will be performed to ensure that they are fully optimized and working to spec. It is important to ensure that the correct hardware and analytic feature set is used to accomplish the task at hand. The better the quality of the video image, the more effective the analytics can perform, and certain analytics work better with specific types of cameras, as is the case with color versus thermal imaging,” Robertson says.
Just as critical is weighing the value of edged-based versus cloud-based programs, “depending on the processing power and bandwidth needed for the analytic,” Robertson says.
Be especially thorough when choosing and positioning your analytic cameras for monitoring, says Justin Wilmas at Netwatch North America. “Video analytics is something that has to be set up correctly in order for it to work. A lot of customers will say, ‘Great, I have a camera right there on the corner of my building. Can we utilize that camera with the analytics that are there?’ That camera may be in a great position for video surveillance, but may not be in the optimal position for the video analytics to work effectively. Or it may be a fisheye camera or other type that distorts the image view. That also can have an effect on the accuracy of the analytics,” he says.
Netwatch has a Proactive Video Monitoring (PVM) support services team at the central station that works closely with security integrators on system design. “We can help them build out coverage maps of where to put cameras for the proactive video monitoring to work effectively. We also have tools for the onboarding of their customers, where we do a three-day test of the cameras and the analytics before we take it live to make sure that the alarms are coming in correctly, the cameras are in the right positions, if there are any changes that need to be made, and all of those types of things,” Wilmas says. “They’re there specifically to work with our dealers on the design and the onboarding and commissioning of the systems into our central station.”
“The newest and the most advanced analytics do, in fact, utilize AI,” says Anthony Iannone, director, Affiliated Monitoring, Union, N.J. “It is almost a quantum leap in quality and usability from some of the older analytic technologies — where we may have been plagued with excessive false alarms or the inability to really discern more than just the difference between a person and an animal.”
AI’s ability to learn and apply that intelligence to make analytics more accurate makes them more and more useful to the monitoring center because they reduce false alarms. “The alarms that we do get tend to be very high quality,” Iannone says.
There is a wide variety of analytics that address different applications; and with artificial intelligence they become even more perceptive, discerning and acute. Through deep learning, the AI becomes better over time at deciphering situations. (For more on what differentiates analytics from AI, click on the links to audio interviews found throughout this article.)
“I think it’s really important for the security integrator to understand the problems they’re trying to solve and try and then match an analytic provider that’s strong in those areas,” advises Simon Morgan, chief product officer of SureView, Tampa, Fla. SureView is an independent provider of software that improves the ability of security command centers to manage and respond to events.
“There are unique products that are better tailored to different types of environments and different scenarios,” Morgan says. “I’ve seen analytic providers that grew out of the military and, hence, they’re quite good at distances and those types of things. But that might not work for a campus, because those are not the types of things they’re trying to detect; therefore, I might be buying something that’s way overpriced or way more feature-ish than I’m ever going to use.”
Very good analytics, however, detect exactly what you’re looking for, such as loitering, or a certain type of vehicle that doesn’t belong in a scene, or a certain number of individuals — and ignore everything else, Morgan says. “The best example is a storm comes through and if you had regular motion detection even if you tweaked it as much as you possibly could, you’d just get flooded with nonsense alarms,” he says. They are very strong in outside environments, which are historically a very challenging place to monitor, he adds.
Kastle Systems’ Video Operations Center, opened in 2020, complements its traditional central station but is dedicated solely to video monitoring. It is UL listed and Five Diamond Certified.
Kastle Systems, a security integrator headquartered in Falls Church, Va., whose business in video monitoring is so robust that two years ago it opened a new Video Operations Center dedicated to video monitoring, relies on the “meaningful insights” provided by video analytics, says Nik Gagvani, general manager for video services. These insights go way beyond simple motion detection.
“What AI brings now, in addition, is the ability to take it a step further by being able to determine how many people there are and how they are positioned with respect to each other. It’s not just, ‘Hey, there is a person here. There are four people here. Two of them are outside, two of them are inside and they’re moving in a certain direction.’ The AI adds a layer of being able to do more sophisticated interpretations of the same video, while at the same time bringing down the level of confusion. Now the AI is able to distinguish between a set of pixels that truly is a person, versus a set of pixels that may look like a person because they’re moving like one, but it could be a shadow,” Gagvani explains.
And these kinds of sophisticated solutions are coming “down-market,” Gagvani says. In years past, AI was too expensive, he says.
In early forms of motion detection cameras looking for movement in the field-of-view could trigger false alarms from elements such as sunlight, blowing trees, shadows, and weather, says David Erickson, chief operations officer of Guardian Alarm Systems, a central station security dealer based in Shreveport, La. “AI is an advanced set of internal rules with learning, enabling the camera to say, ‘This is not a person, it’s just a shadow. I can tell by the movement and the shape and the various algorithms.’... The camera is still looking for those pixelation or motion changes; but then the camera is layering on a level of intelligence to ask, ‘What is this motion? If this motion is determined to be a human or a vehicle, now we’ll create that ONVIF event message trigger to the VMS platform,’” Erickson explains.
There are many applications in which AI-powered analytics are solving real and expensive problems — even some alarming ones.
The National Insurance Crime Bureau stated that since 2019 it has seen a 1,200 percent increase in catalytic converter thefts — “an astronomical increase driven by the cost of precious metals.” These can cost around $3,000 each to replace. At car lots, security professionals say video analytic monitoring with audio talk-down is a welcome solution to deter theft of these automobile components.
Loitering is also becoming a popular application, particularly at places such as cannabis dispensaries and financial institutions where vestibules with ATMs are open around-the-clock and attract loitering and “bedding” among homeless people.
Construction sites, equipment yards, recycling facilities and critical infrastructure sites are other locations where video alarm monitoring has become the norm.
One application you might not expect is at dumpsters and trash bins, to determine if trash is being picked up — or worse.
“The problem that was identified was a large multi-tenant building that could potentially operate 24/7,” describes Morgan Hertel of Rapid Response Monitoring. “In the back of this particular facility, which had no gates so no way to limit access to it, there were a number of trash cans. Things were occurring with the trash cans — everything from body dumps, to trash cans on fire, to bodies and then set on fire, to stolen car parts, you name it. There was a plethora of things that were either happening to or ending up in these trash cans that shouldn’t be happening.”
Being a 24/7 facility there were also people dumping trash legitimately in these dumpsters. “It would have cost them $20,000 a month to staff somebody around-the-clock to look at that. And even then, there would be a question of knowing if someone really belonged there or not.”
Hertel worked with the dealer and an AI company to design a video solution that looked at the scene from a number of different angles. They determined they needed to track things such as the direction people will come from, the times of the day when they typically come from that direction, the types of things in their hands such as a trash can or a trash bag — looking for normal behavior, whatever that happens to be, he says.
“And we’re going to track that for a couple of weeks. After a couple of weeks, we’ll go ahead and turn it loose, and we’ll start curating what appears from the AI to be out of normal. So, it started looking at people that were coming from different directions, at times that they weren’t normally doing that, and also carrying things they weren’t normally carrying during those normal times.
“And lo and behold, we started catching people dumping stuff in the trash can, lighting it on fire, doing this and doing that, and the analytic worked very, very well. By the time it was done, gosh, I think we had almost $25,000 invested into setting up and training the AI to do that.
“Because you can’t throw something at it and say, ‘Hey, look at this for a little while and then tell me what’s different from what.’ There has to be a very defined process to make that work. But that’s a really good example of where AI can and should be used,” Hertel says.
Due to the potentially high cost, AI isn’t anywhere near everyday security, he says. “Video is really hard because it’s super subjective. It’s not binary; there are a lot of shades of grey when it comes to video. I think it’s going to be a while before you see significant work on AI hit the general burglar alarm marketplace,” Hertel says.
AI analytics offer huge benefits to operators processing video alarms. They provide the tools to respond quickly and accurately, says Jeremy White, founder of Pro-Vigil, San Antonio, Texas. Pro-Vigil’s Surveillance Operation Center (SOC) provides remote video monitoring for both direct customers and a network of channel partners.
Pro-Vigil uses industry-standard analytics as the “initial event creator,” then runs the alerts through its AI engine to filter out false alarms. White says the benefit is faster response time and less operator fatigue. “We found that by each operator handling fewer events the accuracy improves, resulting in operators who are more accurate at determining threat, risk and assessment. For central stations, the benefits include quicker response time by handling fewer false alarms, better accuracy, lower headcount and operating costs.
“To the security dealer, the benefit is that this technology and the AI are camera-agnostic so security dealers can choose which cameras they are comfortable with installing. That’s huge because you don’t have to worry about upselling new cameras. This flexibility results in higher adoption rates,” White notes.
Gagvani says Kastle’s approach to subscribers who are using AI-based video analytics either in the camera, in the cloud or both, has been to only send alerts to the central station that have been filtered by the AI, so we’re not flooding the central station with things that an operator may not ever look at. That’s all done upstream. The central station is only receiving things that are actionable.”
That is the reason video alarm monitoring is successfully growing: because AI analytics ignore the false alarms that otherwise would “flood” a central station with unnecessary alerts and make monitoring unmanageable. Thus, its value as a tool to prevent crimes before or as they are happening is realized.
“AI is wonderful and the accuracy of AI is improving every day, but if nobody is looking at it in real time and potentially responding to a crime in progress, they’re losing out on the real value of AI,” Erickson says. “To me that’s where the security monitoring industry has to go to.”
This shift in monitoring paradigms today is occurring because the recurring revenue for security companies and monitoring centers derives from alarm monitoring; and video is now trying to establish a similar status in the central station, he believes.
This analytic/AI rule trip from a camera create a central station event. Binding boxes are shared for the operator engagement. Triggers can be physical sensors paired with camera feeds, or through edge-based intelligent rule sets.
“The industry does find itself in a unique place. What we felt was necessary was now that a camera has a really accurate detection methodology, who is responding to that information in real time and how are they able to respond? Alarm monitoring centers are staffed with operators 24 hours a day, 7 days a week, and they’re the ideal place to push that AI data in real time for a real-time response,” Erickson says.
Today when Guardian Alarm’s sales team sells a security system and a video surveillance system, it’s always video monitored. Their brand name is Guard Watch. “We will explain to the customers, ‘We don’t sell a blind intrusion alarm system anymore; that’s kind of an antiquated methodology. We only sell video alarm monitoring solutions by Guard Watch.’ When our Guardian sales reps are able to talk to businesses in this area, and even consumers who want security in their home, that makes sense to people. ‘Yes, if an alarm occurs and you guys can see it, that’s the service I want,’” Erickson says.
AI enables proactive security, the ability to protect people’s exterior property without waiting for a door to get broken into or a glass-breakage sensor to alarm before initiating a response, he says. “You’re detecting someone outside before they get to the building or you’re even protecting outside equipment such as tools or vehicles,” Erickson says. “Now Guardian can give a service to customers and say, ‘If somebody’s on your property we’ll detect them. We can deter them with audio services like audio talk-down messaging and lighting control, and maybe prevent a crime from ever happening.’”
Erickson emphasizes that these types of services have the potential for high recurring revenue. “A plumbing contractor that may have spent $50 a month on his alarm system to protect his building might be willing to spend in the area of $200 to $250 a month knowing that we can protect his trucks and all his gear outside; because that’s where his valuable assets are and that’s where his productivity the next day is. That’s the proactive video monitoring space — that’s where high recurring revenue value is available and it’s not at price points that don’t give value to the customer. It’s really a great balance of service and revenue for everybody involved,” Erickson says.
In order for central station operators to respond to AI-analytic alarms they need to have a platform designed to optimize the processing of video alerts. There are a variety of video monitoring platforms now used by central stations — some of which are commercially available and some proprietary.
An interesting example is the solution developed by Guardian Alarm Systems’ COO David Erickson and founder and chief executive officer, Wes Usie, who is on the board of directors of The Monitoring Association. Their philosophy as security dealers was that video alarms should be routed through the same automation platform that operators use to process traditional alarm signals.
“We said if this was going to be a service of the alarm monitoring center it had to become a part of what operators already do every day. If an operator is sitting in front of a MASterMind platform or a stages platform, their trained operation procedures shouldn’t change, Erickson says.
Erickson and Usie believed the true value of AI in security could be realized by providing actionable video alarms and presenting them to a central station. Thus, about 10 years ago, they formed a solution that uses the OPTEX Video Bridge powered by the CHeKT Visual Alarm Monitoring platform, and it is available to any central station alarm company or wholesale monitoring station to provide remote video alarm monitoring services using existing alarm systems and camera systems.
“With CHeKT we allow companies like a Guardian Alarm Systems or other dealers to continue to install their alarm system — their DSC, their DMP, their Honeywell systems — and install video systems — their Avigilon, their Hanwha, their Axis, their Bosch — and then combine those two systems with the CHeKT platform so when an operator receives an alarm, they’re processing it all in one solution and can respond to that alarm accordingly,” describes Erickson.
The platform allows security companies to create an alarm video event from the existing alarm panel zone. For example, by pairing a door contact, motion detector or photobeam with a camera, when a door opens or the motion detector is activated, the operator can see exactly who opened the door or caused the alarm.
“But one of the things we did about two years ago that I think was really unique was to be able to convert AI data into an alarm sensor transmission, as well,” Erickson says. “So now analytics or AI on an Avigilon camera, as an example, is converted into an alarm signal like a sensor and it goes right into MasterMind, right into BOLD, and it’s processed by the operator just like it’s an alarm zone of a DSC alarm panel except it provides video evidence to that site for an operator.”
Chris Brown, CEO of Immix, Tampa, Fla., believes AI offers security integrators the potential to provide many new services they otherwise wouldn’t have access to selling. Key to that is defining a menu of services, and then identifying analytic or AI products that enable those services, he says. At Immix he has seen demonstrations of nearly every AI product applicable to security.
Brown mentions some applications where AI — particularly its ability to do deep learning on a scene — is well-suited for generating revenue by creating video alerts. One example: At construction sites, zones can be set up where supplies can be delivered and then builders can be notified when the delivery has been made, while the rest of the site stays protected. “A lot of times deliveries happen on the weekends and evenings. To disarm an entire construction site just to take a delivery is a risky thing to do. With AI, you can have it report that there’s a delivery in progress and manage the detection of that delivery,” Brown describes.
In this thermal analytic detection, the thermal imager is presented to the operator to establish the target dimension and location of threat in the field-of-view.
“AI gives you the ability to exclude things in a scene while leaving it protected,” he says. By using an AI analytic that can distinguish between colors and teach it to disregard anybody in a fluorescent vest, for example, then people could move around a site wearing a certain identifying color and that color could be excluded from detection. But if a person walks into the scene that isn’t draped in that color, they get detected as a person on the premises. The same can be done with vehicles,” Brown says.
“It opens up more possibilities and applications with the more granular detail that AI can provide, as opposed to what analytics has traditionally provided, which is just detecting people, detecting vehicles, detecting direction of movement and loitering conditions,” says Jason Caldwell, director of marketing and business development, and GuardForce Accounts, at Immix. “AI can do a lot more.
“The ability to now go beyond just, ‘I’ve got something that is in the shape of a vehicle that I’m going to say is a vehicle, and I’ve got something in the shape of a person that I’m going to say is a person. … now we can dig into really classifying those individual targets and identifying them in great detail,” Brown says.
Video and especially AI analytics represent the future of the alarm monitoring industry. “The rate at which we’re seeing new opportunities for new services emerge, you can’t even get your head around that,” Brown says. “The evolution of this is just accelerating at such a rate. It really won’t be pixel management anymore; and that’s what it used to be. It is now real science behind the scene…. I think AI is going to revolutionize, as it already has, the video-monitoring industry. The menu of services and things you’re going to be able to offer, limitless. Absolutely limitless.”
Netwatch’s technology filters approximately 50 percent of false analytic alerts, allowing NMC to monitor 600 cameras per one full-time employee. With the Netwatch platform, the company says it generates 98 percent fewer dispatches to police than with traditional monitoring solutions.
Caldwell credits the advancement of analytics, and particularly AI, as being the main catalyst that has fueled the growth in the video monitoring and remote-guarding industry over the last several years. “It has absolutely been one of the core things that has been essential to our success,” he says.
“We envision a time in the not-too-distant future where the PIR motion detector is relegated to the past and cameras — because of decreased cost and increased capability — are the new primary detector of intrusion,” Iannone adds. “Using video to solve problems that conventional intrusion technology cannot solve is perhaps no better way for security dealers and integrators to strengthen their relationship with their customers, while generating additional RMR. If you can solve a problem that could only historically be solved with a live guard, and do it at a fraction of the price, you have a great opportunity to give your customer a huge ROI on an investment that is maybe only a fraction of the cost of a guard service, but can be hundreds or even thousands of dollars a month in RMR for our partners.” VMT