# How Ambos Security Turns Milestone VMS Alarms into Autonomous Drone Missions - Event: NestGen ’26 - Published: October 2, 2026 - Length: 28:08 - Watch: https://nestgen.org/library/watch/UCNQMsDDZAc/ > The transcript of a recorded NestGen session, for reading or for use with an AI tool. It comes from the video's > captions; automatic captions can mishear names and terms. ## About this session Drone response too often runs as a separate system from the video management platform security teams already operate every day, forcing operators to work across disconnected tools during the moments that matter most. In this NestGen showcase, Ambos Security and Milestone Systems will share how autonomous drone response runs directly inside Milestone's VMS, and what security teams should think through when building a drone response layer natively into their existing video management infrastructure. ## Transcript **[0:00]** Welcome back to another session at FlytBase at Nestin and I am really really really glad that all of you were able to join in. Um thank you so much Roy and Marco for joining in as well. >> But what are we thanks Marco and Roy but what are we trying to address today? So we have like how obvious security turns Milestone VMS alarms into autonomous drone missions. Sounds huge but why are we addressing this today? So for years the security industry has talked about connecting everything together the cameras the alarms the response systems and many more. This session is one of the first times we seeing it actually done through Marco as well. Ambo security connected milestones game is directly into an autonomous drone talk. and allow fires, a drone launches and there's a live aerial feed of the situation before anybody has picked up a radio. Now let's try and understand why this is helpful from the experts of the field of course Roy and Marco they'll be letting us know how **[1:05]** they've solved this and Marco has built the system as well. So he'll have a lot to contribute and tell us throughout this will be a nice Q&A session where we'll be discussing and talking about how we've built this entire thing. So do drop in your questions in the Q&A tab and you'll also be receiving the recording of this later. Also any part of the session where you find it nice if you did like to take a recording or a screenshot and if you're posting on social media please do hashtag Nestin and also tag me Roy and Marco as well. So thank you so much. Let's get started. Gentlemen, are you ready? That's great. uh Roy uh for your round of introduction. So what is PL? What is Milestone? >> Yeah, Milestone about Milestone. Um yeah, Milestone is um one of the market leading uh video management uh software companies uh established in 1998 **[2:05]** and uh yeah we offer uh actually uh a couple of products of which our main product is uh is Milestones expert uh solutions which is the video management solutions that u connects cameras uh drones and robotics uh all kinds of sensors basically uh it's it's an open platform datadriven platform platform and you can find us in a lot of verticals like cities uh in retail, healthcare, government, education, airports, you name it. Everywhere everywhere where a scalable video solution is required uh you will typically find a milestone solution. Yeah. So that's in short about milestone. [snorts] >> Great. And what about you Marco? About you and Ambos? >> Yeah. Um thanks for the introduction first. Um yeah, we at Ambos we create um solutions for the customer security solutions which uh involve all different kind of sensors and IoT devices. So our goal is always to provide the centralized platform **[3:08]** where everything comes together. Um like like we also use milestone of course um that's why we're here and uh in those platforms then we have the possibility to yeah gain a lot of synergies just between all of those systems when they come together towards the operator. We have huge value um just by doing so. [laughter] Great. So let's quickly get into the topic. So we have seen that drones have entered the physical security um you know space and it's already changing how sites are covered and how teams are responding at this moment. And today we're going to talk about what happens when this drone becomes a part of the larger security operation and how platforms like milestone brings everything together where AI will fit and what the future of security looks like. So Marco like you've been working with Ambos you've **[4:08]** been doing a lot of things you are a solution architect yourself and uh you have seen a lot of drone based security solutions for your customers. So what are the common challenges that they come out with in the first place and why are they adopting drones? >> Um so what what we always target for our solutions that that scale so that scale with use cases that scale with the the size of the customer with the customer itself. I mean the customer also grows and so the requirements also grow and when we now want to scale on a for example parimeter and also want to scale with the response times um yeah drones are the solution for this. So this this helps us to perform a fast validation of events and um also with this um yeah we we we reach systems which are more proactive instead of reactive and this is exactly also the goal what we also wanted to **[5:09]** pro provide as a value for the customer that they have a system where they can easily perform verification of the events that occur even on a large scale Right. Uh what I can understand is drone was clearly used to solve a coverage and response problem, right? Like time and scale, right? >> But um if you're creating an entire drone solution, uh what comes to your head like what are the few things the entire solutioning package that would look like for uh using drones for your customers? >> Um so for our customers they don't have just a drone. So we we really target systems that involve way more. So in the end the problem is every customer is different is unique. They have their own requirements. Um but what all our customers have in common is like one uniform operational platform and in this we integrate **[6:10]** yeah the the the technology that we that we need. This can also involve brownfield environments and um yes so basically a system really consists of multiple sensors IoT devices cameras drones autonomous robots that we also deploy or integrate into the systems and uh also a process around this like what what happens also in the event cases. So this is also all designed together with us with the customer >> right I mean all I can see is that when you give an entire security solution it's not just um security using drones it's just you're trying to solve security as an entire problem and it involves a lot of elements inside it and Roy like as a VMS platform like in milestone as well um there must be a lot of elements that you guys are adding in and there's a lot of changes that you might have seen like maybe from 10 years ago go of a VMS platform to right now. What do you see um you know in people's ass or the way people are integrating it into your platform? **[7:11]** >> Yeah. So the the way we look at these things I mean our platform expert is is very scalable. So you can use it for six to 10 cameras but up all the way up to 10 thousands of cameras. So the amount of cameras is typically not the limit anymore. So that that's no problem. It's uh just a matter on how you process all that information. Um and and that's where the limits come in because we can easily add cameras. But yeah, you also have to have people and and systems that can uh interpret the data, review data and [snorts] and deal with it. So there is the real limit. You cannot just throw new people or more people at at a system and a solution. So you need technology to to evolve and to grow uh further and to do more with what you have. So this is where uh yeah physical AI and and AI assisted video systems uh yeah become very important and you know in order to scale and to expand these limits, **[8:13]** >> right? I mean, so basically just trying to double up coverage or trying to increase your coverage by adding more drones or adding more cameras is anyways going to make it work into silos. The entire story of it is trying to bring all of it together into one system so that each of them are able to interact with each other and that way you're not uh you know doubling your spend in terms of the people that you need to hire to process all that data in the future. Uh I mean that's a great input Roy and I think that is where we all heading as well and silos has always been an issue right now. So um is this something that both of you have seen across the market and is there some solution to it or have you guys seen uh any solution to this? Yeah, this we often see exactly those silo applications at the customer sites. Um, when they approach us, they often have different systems for access management, **[9:13]** intrusion detection, all like decoupled basically and uh yeah, no interaction between them at all. >> Right. uh and I think at this is what we at at least uh FlytBase and a few webinars with uh Roy as well we call this entire solutioning as the blended security solution where we feel that um you know it's not um you know replacing anybody we are trying to augment it into your current system so that everything talks and works and you know happens all together so I mean I would move it I mean give it to you Roy like if you would like to explain more about what blended security is and what we are thinking about it in the future. >> Yeah, so blended security is basically where uh a AIdriven video solutions meets physical AI. So uh we get a lot of input from sensors and other technology uh and and fixed cameras on our system that are also AI powered that generates a trigger and of course then the physical AI is **[10:15]** blended into what we traditionally have uh to respond to incidents and to uh reduce response times for example >> right and Marco any input from your side I mean how is blend do you see blended security happening in your sites and have you seen any considerable impact out of it or anything as such. >> Yes. So we this is our our core core DNA basically to have um blended security. We want all of those systems in one platform which then uh is accessible by the customer and uh yeah this this we see we see a huge um huge value in this also for the customer when when they also for the operators themselves in the end they don't have to search for data and fragmented UIs and saves the costs for all those single systems you just have one system >> right >> we always splendid security. This is our goal >> right. So in the end we are talking about having so even though you have so many **[11:21]** hardwares in the end the platform that we I mean the view that we are coming to is that we need to have a platform layer or a place where all of these things are coming together and it's not just so fragmented and it's not just so siloed and bringing it together into a platform is not an easy job. A lot of people are working on it. So my question to uh Milestone is is Milestone still a VVA VMS or are you thinking way further? Is there something that um you thinking of in the future? >> Um yes [snorts] well Milestone is is still a VMS. Uh the only thing is the role of a VMS has changed quite significantly in the past years. So when we started in 1988, it was a pure requirement to uh connect all the different brands of of IP cameras. So IP was the cameras were transforming from analog to to digital back then. But now uh cameras and devices have become smarter AI powered and uh people **[12:22]** expect more from video surveillance systems than uh 10 years ago. So yes, we are still a VMS but we can uh go beyond security if necessary as well. So with all the smart technology that we can put on top of video, we can also assist operational processes for quality inspections for health and safety. uh and and that's why uh platforms so the statement uh from the slide before is is quite crucial because if you look at how fast technology evolves uh the platforms you choose is becoming quite critical and uh if you look back two or three years ago what happened with AI for example that speed of innovation is so fast um if if you start deploying a larger scale system that takes maybe a year to write a tender and to put onto the Then before the first camera is mounted, easily two years have been passed. So considering the right technology, the right platforms and benefit from each ecosystem uh is very important. Uh FlytBase has that **[13:26]** ecosystem with drones and detection. Milestone has the camera ecosystem and if you combine that, that's very powerful. [snorts] >> Exactly. It's all about integrations at the end of the day. Right. When I think of a platform Yeah. When I think of a platform, I think the first thing that comes to my mind is integrations. And it becomes very important when you can bring all of these data together and integrate with multiple of these platforms in order to create this entire layer. And I think it also separates the concerns of different people. So for example, uh a security person need not know how to fly a drone whereas a drone person need not know like how much of the security VMS systems would work for them. So it's two different separation of concerns but understandable Roy that's that's a great input but Marco you've built this in person and we do have a video of you doing it so if you did walk us through or before that Roy if you were about to say something **[14:26]** >> no no go ahead >> okay yeah Marco uh I'll just go ahead >> yeah so we have a a showcase here of a of integrated system where we have a fence intrusion combined with a drone response. So we see basically now a person breaking in. We're now in milestone where this fence alarm will reach the operator. On the same time this GPS coordinates or like from the events can be utilized to like guide the drone where to go and now everything comes together. So the drone is is uh performs this automated mission towards the defense and all the streams now come together in in milestone. And this is this is really really really crucial also for the operator and for all the documentation afterwards because in the end now I have everything linked together to this one event. I I have first of all the information what's going on. **[15:28]** The operators they see quite clearly what's what's happening. they can then check what's what's going on and also trigger an intervention team. Um here now the drone lens directly but in the end you can also use the drone of course to guide the intervention team and tell give them updates about what what's what's happening >> and yeah it's a very very short this one >> right we can also have uh the intr I mean the person interception team on the ground have the alerts on the mobile as well right like that is something maybe we can >> that the the MVMS feed getting into them, the video feed getting into their mobile devices will help them redirect to this encryption as well. >> Yeah, it's a huge advantage also of just one platforms. You have one access point where you can reach all this information, >> right? Uh moving to the next one. We spoke about uh drones uh and that responding in your VMS **[16:34]** platform, but I've seen that you've gone way further as well. So would you like to talk about your robot response? One second. I'll just have the video played. >> Yeah. So, as I mentioned, we also have a robot and >> for for us it's always always the goal. Maybe you can silence it, please. >> Yes. Perfect. So, we here we see you see our our robot. >> Sorry, >> it just undocked. And now again, all the camera feeds come together. We have a little software from by ourself where you can perform all the scheduling, the missions. You see the mission control you can just uh define the times when the robot will go on autonomous patrol and um now a person is detected was hiding one camera already caught the PT set camera that is also installed on the robot will add or will will record additional data. **[17:36]** We will see this alarm again in milestone. Everything happens in milestone here. Also the the camera feeds and so on, the streams, they are all recorded in milestone. And um now this interaction is that the robot always knows where he is. So uh we can now forward also this data to FlytBase where again we can dispatch the drone um and the drone operator could now or can now perform a visual or get visual overview what's going on collect more data more information perform verification also trigger an intervention team if it is really required and support the intervention team um on the same time and basically the robot's task is done. So the robot will just continue its patrol if it's not required to use it any further. So you can just it depends really on the use case how you really want to handle this incident. **[18:38]** So the robot just will continue with the patrol. Yeah. And in the end the drone also will play back and we so this is for for us important to show also this because we have this whole process from detection till the end everything now recorded and documented which is very very important for us also to provide this to the customer that they have seamlessly >> right you can create uh reports from these also at the end of the day right just to let them know how you've been able to secure the crowd >> exactly >> yeah and Roy actually it's your ecosystem being used over here and I'm pretty sure it's your um you know ecosystems APIs and integrations um that Mako has been able to um utilize and thanks to the open ecosystem that all of us have at this point a lot of these drone enthusiasts have been able to explore into um this platform as much any comments from your side >> yeah I think uh what you saw now is is a nice example of how an open open platform facility facilitates **[19:43]** integration and I often say uh integration is the key to the future uh because you never know what's the next technology that's going to be available. So openness uh in in that sense is key. Um so staying open uh to be able to adapt to to new technology because uh what comes next is maybe now we see these uh wheeled uh robots but what comes next may be humanoid type of robots pretty soon. So you want to be able to integrate that as well and uh yeah that that's ultimate flexibility for end customers that are looking for solutions. >> Glad to know there's someone else also trying to explore into the physical AI era. I mean we've also been trying hard to experiment and try get into the physical AI era with humanoids and robots and everything on these life fields. But um actually uh that that fits out well. But uh one thing one important part is that people just don't stop with one dog or one drone. Right? Now they expand. They **[20:45]** have 10 dogs, 10 drones and there's a lot of data, there's a lot of people and now let me tell you what people do is scale and then scale the people again and again it is trying to double their spend because while scaling they are indirectly trying to double their spend as well because that much of data needs to get processed at this point. Um so I think this is where I feel AI comes into the picture and it comes into drone operations to process this huge volume of data. Something that we discussed earlier as well in the presentation but few words uh from your side like what is it where you think that you know AI is giving out an impact at this point Marco like in your operations? Um so we use AI on multiple places and in multiple multiple steps and and also on multiple devices in the end because what so when when we bring technology to a customer we **[21:47]** always work together with with technology that already brings AI with it on on the edge devices like smart cameras, smart fence sensors and um this brings like a first pre-processing of data so that we don't bring all events towards the customer already have a pre-f filtering but also it is very important on on the platform level we require those um AI um this this AI technology because um when we are now in this platform we have the possibility to create rules and very very clever rule logic behind all the devices so we can utilize different AIS and then even on this layer can bring AI into the game for additional processing, finding correlations, maybe figure out what are anomalies before we even bring then those data towards an operator. So the operators themselves, they should receive **[22:48]** qualified incidents in the end. This is the goal and this is possible with the technology that we use. >> Right? So a lot of people actually uh talk about AI I feel as something that just detects incidents or like just raises alarms and you know that that's more about AI but I feel like you just mentioned filtering out the uh noise learning from it and then understanding the situation and giving uh the person the right awareness required is where I think the entire AI insecurities is heading to and uh Roy on a platform side as well you must have seen a lot of these things. There are right now VM VLM summarizers as well which summarizes incidents for people. So what do you think? >> Yeah. So what we traditionally are good at is object detection uh based on AI uh in 202021 access and IQ site formerly known as Bosch already started to implement object detection AI models into their cameras but uh that is now are becoming a commodity. **[23:52]** So what we see now is the the rise of generative AI solutions where we can do really video summarizations but also support uh control rooms with actually analyzing uh with vision language technology to see what actually is going on with video summarizations and to filter out uh the incidents uh that are false positives or or false negatives plus uh the uh AI assistant uh operations in control rooms. So AI agents that then go on top of a VMS interacting with the subm modules to control or to uh interact with uh for example uh exoscroll systems, robotic systems where the whole yeah interfacing between a VMS uh changes uh how an operator uses it. In the past still up to today we're used to using a mouse to click here to click there. But that whole type of interaction with machines is is going to change uh **[24:54]** thanks to generative AI and agentic AI. >> That's true. I mean it's gone way beyond but since uh we are a little bit uh short on time uh and we are on AI and the last topic that we had kept for ourself as well. I just wanted to quickly know from the both of you like where do you think the future of security is heading with physical AI in picture? So what what's your take a quick take on your future of security operations? >> Um so so first of all I really also think that the um the interaction it will change. It will be more expectations also from from operators to talk in a natural language with the systems and we will definitely see more more crown robots, humanoids, drones in the field like collaborating with each other. Um but also I I'm pretty convinced that the sec security offers themselves they will collaborate with those systems when there are interventions **[25:55]** and then this happens of course with natural language. So this this interface will change >> and I'm also >> quite sure that like the way how security officers will receive data and information will ex expand into maybe also more augmented way just to to have like for example smart classes with with live data. >> That's a nice perspective augmented AI in the future. Roy your perspective. >> Uh I fully agree with bit Marco. Uh in other presentations I usually say I AI changes the way uh machines interact with humans. Uh because traditional in in in it we are used to a human machine interface but this is going to be changing 180 degrees where actually the machines provide humans with input and then you have to make decisions. So the whole interaction like Marco said is is going to change and the way we communicate with them is also going to change massively. So I fully agree **[26:56]** with that. >> Right. I'm actually looking forward to more of like a machine to machine and then machine to human and human to machine. It's going to be like a triangle that's going to keep just going on and on. And I'm actually very looking forward into that future where a machine can interact with a machine and get jobs done and we can just sit back and look only for important u you know instances. So I I I'm really glad I could host you uh Roy and Marco the few visionaries of the security industry and I'm looking forward to another session with both of you. Uh but thank you so much for joining in. Uh since we're short on time, we'll not be taking any questions. But for any questions you guys have, I've attached their LinkedIn in the chat box. So feel free to contact them. I hope it's okay, Roy and Marco. They'll be contacting you with more questions. But thank you so much for joining in. >> You're welcome. >> Yes. Thank you. Bye, guys. >> Thank you. Thanks to the audience. Byebye. **[28:03]** It's great.