IoT and Agentic AI Power Tomorrow’s Smart Hospitals

So this week, we are talking healthcare and not just healthcare. We're talking how to turn the operations of, in this case, a hospital, in fact, several hospitals, into a production line with a level of automation that assembles an automotive production line or an Amazon warehouse. And if that sounds crazy, with all of our personal experience of the healthcare system, no matter what country you're in, then I would say keep listening. My guest today, is Dr. Burak Bardak. He's the CTO of a company based in Türkiye. They operate globally called Borda, B-O-R-D-A, Technology. And he's got a deep IoT background, but they've taken IoT for healthcare and have turned it into a system which resembles, and you'll hear us talk about this, the world's biggest Sims game where you simulate usually, it's not simulating. You're showing the real time collaboration between assets in the healthcare, the staff in a hospital, and people, i.e., the patients, and then putting Agentic AI on top to define the processes, optimize the processes, change the processes, and give comparative KPIs across different hospitals within a country or within a healthcare group. It's a very interesting story. It really links IoT with AI, which is a big subject right now and gives some great examples of what Agentic AI can do. Without further ado, let's get going with my interview with Burak Bardak, who's the CTO of Borda Technology. Enjoy. So, Burak, good morning, and welcome to the IoT Leaders Podcast. Good morning, Nick. Great. Now this is a very interesting podcast because healthcare is a really important area, and we're going to get into the whole healthcare vertical. But before we do that, you've also got a very deep and long experience in the field. And one of the themes of the IoT Leaders Podcast, especially the recent ones we've been doing, is increasingly IoT and AI. And I think this one in particular really is a blend of those two. So to start us off, can you just give me an overview of your background and the different areas you've worked in before you arrived at Borda Technology as the CTO? Yeah. Sure. Thanks for inviting me. Okay? I'm originally from Cyprus. Completed my undergraduate studies in electronics and communications in Istanbul, Türkiye. I worked with for wireless communication during these studies, and this was back in 2007, actually. Later that, I moved to England for my Master's in Computer Engineering. Again, I focused on, algorithms or image reconstruction enhancement back in 2008. In my past, it led me to location technologies, which is both indoor and outdoor positioning and navigation, which was my PhD research was focused on. I did my PhD in London at University of Westminster, worked on a Galileo project called Insight Project. It was a collaboration between four different universities and industry. And my research focus on optimization, system integration. And, for my postgraduate studies as a postdoc, I work on edge posters for cameras, develop power processing, parallel processing algorithms, and compiler implementations to run AI at the edge, which is within the cameras, up to 600 core processors. Everything is within low, very low latency. And then after nearly a decade, and I transitioned into the private sector, working in highly regulated industries like banking, finance, healthcare, and payment systems. That's very relevant and very impressive. Essentially, that's eighteen years of experience in the areas. And, from a time well before it was called IoT, it was machine learning. But, also, 600 processors in a camera. I believe from when we spoke originally, that was in military settings. We can't go into that, but that's incredible power. Edge processing is a great example of edge processing and AI image enhancement, and I know that was in the biological field and you were looking at human cells. Yeah. I worked on image reconstruction for microscopes. We use different phase information to enhance the image. And in order to do that, we train the AI in the background, which is, in the old days, we call them the neural networks. Now the name is the Neural. That's right. They were neural net self healing neural networks. Yeah. Yeah. Microsoft used to brand them. And we we enhance the images, get the 3D we convert the 2D images to 3D ones with the help of AI and then, basically, these things build up. Small things, they build up, and it came the AI that We know today. They want Yeah. Changing weekly as we find on this podcast. Yeah. So you're now at Borda Technology, and you're based in Türkiye. And most people who listen to this, I think I'm right, probably won't know anything about Borda Technology, although we have listeners in 95 countries. So there will be some people in Türkiye who know all about you and the other countries in which you operate. But for everybody else, can you just give me a high level overview of Borda Technology and what they do and its relation to the background the deep background you've got in IoT and AI? Do you? About six years ago, I joined Borda Technology. I live work on next generation at location technologies. In simpler terms, we build the next generation technology for indoor tracking. So interact entities in enclosed environments, indoor and outdoor environment, actually. And we built the technology to be able to do this in a very high precise and low latency way. So, basically, we can achieve up to ten centimetre precision with power powered tags. Okay. So let's go into the tech a bit. We're going to talk a lot about business processes and get into agenda. But on our journey, let's start with the bottom rung of the tech. Asset tracking, indoor and outdoor, specific capabilities, which is very relevant to your background, of being able to identify an asset within ten centimeters, I think you said, which is Yes. Not very much at all. That's pretty industry leading. And it uses is that BLE technology from a sensor that's attached to? So it's multi technology. We use ultra-wideband. We use BLE as our base, and we use BLE Angle of Arrival. So for ultra-wide band, you can achieve to the ten centimeters. With the BLE Angle of Arrival, we can achieve to the submeter level accuracy, which is 40-50 centimeters. And then with the other side, you get a kind of a zone level accuracy. So you're using some form of triangulation. So if it was in a hospital, you're putting sensors, battery powered sensors on devices using triangulation to identify them. And the battery life of one of these sensors So we call them tags, actually. And these tags, they have five years of battery life. Usually, it's around two to five years depending on the usage. And our gateways, which we call them locators, and they are used to calculate the envelope, the transmitted signal from the tags. They are called locators. And the power through the ethernet, power over ethernet, being sold is on sale, and then we have these tags. This is a sample stuff tag, which has five years of battery life. We can track the stuffs or persons who wants to be tracked, and for various reasons, can be tracked less than to eleven, let's say. Again, to help people unpack this You held something up there, it looked like a credit card, but most people are probably listening. Yeah. The taxonomy here is assets, people, and patients. Now we'll now in the healthcare environment, a hospital, for example. Yeah. And we'll talk about a case study. But if we just so first of all, the assets are the tags that get physically attached to, asset. Yes. So the asset tags are very small. The tags that we have again, I'm showing you, but it's a very small 3x5 centimetre tag, which has five years of water life, and we can track the assess location along with the some inertial some with these tags, they also have some inertial sensors, so it can sense the vibrations. They can sense temperature. They can sense the movement of the others. So you build a picture of all the assets in a hospital, and we've had a a couple of healthcare examples on the pod. Tracking assets in hospitals is a major issue because people tend to pick them up. If it's a trolley or something, they they roll it down a corridor, they take it somewhere else, and nobody ever knows where it is. So tracking the assets. Now the doc, that looked like, for those listening, you were holding up something that looked like, I would say, like a very thin domino. That was about the size of it. Yeah. Domino piece. Then you said about the staff. Now the staff was the credit card one, so we're tracking assets and triangulation. The staff are the nurses and the doctors, I assume, and the business Nurses, doctors, healthcare workers. Yeah. This is the credit card size device. And I think you mentioned the word safety when you Yeah. So what's the use case for the the staff? And then from there, we'll get into the patients. Sure. So, basically, unfortunately, there can be attacks on healthcare workers in hospitals. And our job is to prevent these attacks, for stress purposes, staff stress. We call for product called staff direct stress prevention. So whenever a healthcare worker is basically is being attacked or someone is basically detention is increasing that particular situation, they double click or maybe it can be programmed that or they long press on on our staff tag. And this basically triggers an alarming background with the exact position of that particular staff. And we can even connect CCTV cameras and get some screenshots and send that them to security. Right. Then security will know which staff is being under attack by phone and where. Sec there to prevent this attack be to the healthcare work. Okay. So primary use is security of stuff, and it is a big thing in A&E. All over the world, it's a big thing. There's a lot of people who drink a lot who tend to end up A&E with mental health issues. But, of course, you a side effect, of course, is you're also getting the location of the healthcare worker as well. So now you got the location of the assets, you've got the location, and whether they're moving or not or vibration, you've got a location of all the healthcare workers and their identity, so you know where they are. Now let's look at the third aspect in this triumvirate, which is patience. So talk me through what happens if I'm going to a hospital in Ankara in Türkiye patient. How does this system track me? What happens what happens when I check-in as a patient? Okay. After we check-in as a patient, basically, we get the admission information through the entire patient, through the EHR, to our system, and the nurse knows which patient to be needs to be tagged, and they go next to the patient, and they take the patient. They and the patient means something like a small pouch. They have very small tag. I will show what they can a wrist tag or something? Oh, there you go. You have to give it wrist there. And we did our response, which is our tags are reused. They attach this wristband to the hand of the patient, and we can start tracking the patient's location. So, again, for those of you not on video, because Burak has been holding up several devices, so we're not on video. Yeah. What I would say is if you've been to a Coldplay concert and you've been issued with a multi a wristband that changes colour, it looks like that. Smaller, though. It's still in monitor. I'm not using the video for the kiss cam. That that's another story. But it's smaller than that. But, essentially, what we've got here is if you think of the picture, we know the location of all the assets and their moving and what assets they are. We know the location of all the healthcare staff and who they are. And the patient know the who the patients are. And I know in Türkiye and across the Middle East, and I'm very jealous as someone who operated out of the Middle East, and I didn't know this, but I'm very jealous as someone who's in the UK. You also, the systems, the data that the patients' healthcare records, it can be integrated as well from the GPs, the doctors, nationally and across the region. So that really helps because then you've got all the patient's history in the data. So what you've got is, like, this model with the patient history, and then above it, you've got assets, stuff, and you've got patients. And now you've got the data, and, again, in my mind, it's like a big Sims game. It's a very complicated Sims game because it's a hospital, and everything's moving. The staff are moving. The assets are moving. The patients are moving, and a lot of information about everything. Okay? So that's the best. That's I think, obviously, it's a very big Sims game. And this seven twenty four operation, the operation doesn't stop. And you have to be on the wall all the time, and you have to collect all the information that you can get from the sensors, from the or from the HR, and feed in one big system and start working on your magic to optimize it, optimize the make it more efficient, making track turnout turnaround times, make them faster, more efficient. You have to manage your capacity. So it goes on it's like a very big system, very complex system, and you need all of the information you can get to feed in that model that you created and start on the optimization. So you're visually, I encourage people to think of it like a stack of bricks. You've got the bottom level. You got the basing data from patients' health records, the GPs, around tech, or even visitors. Are you fortunate to end it up in a in a house this hospital? Now you've got you've got the the assets. You've got the big Sims game of the assets, the people, and the the staff and the patients. And now you're collecting a lot of KPIs, which is what you just talked about. You're collect you're collecting, I would imagine, tens of, if not hundreds of thousands. I mean, over time, millions of data points, KPIs into a system, which I guess comes to now what Borda Technology do, and it's going to lead us into the AI piece. Yeah. Because now Of course. You have a solution now for not just reporting the KPIs, but actually looking at the processes and the process optimize optimization. Is that right? Yeah. That's certainly correct, actually. So we for office citizens, we track staff in patient and the assets. And asset is a very big part of the operation inside in the hospital. I mean Because obviously, medical devices, they are very expensive, and it's very hard for hospital management to measure the usage of these assets. And they somehow invest and procure more asset than they need because they don't have any way to measure the utilization of these assets. What do you mean? That means our system provides them the visibility to reorganize more the assets to the right places, and that with with doing that, they can reduce the asset. They can increase asset utilization. They can reduce the investment they need to make for the assets for the hospital. Basically, they over purchase by twenty to thirty percent because they don't know where their equipment is. They don't know how are being used. They don't know how if they are maintained, if they are up and running, if they are calibrated. So our system basically solves their problems. They can reduce the investment they have put in by thirty percent. Thirty percent. Just basic information, like finding sometimes it's simplest it's the simple stuff, which has the biggest payment, ROI. We did a we did a podcast recently on which is also IoT and AI with the guy who runs IoT for an AI for Volvo, a hundred and forty factory, and he's, tracking millions of components. But he said one of the biggest paybacks is when they make a truck and they put it out in the field, they're all white initially before they get painted. And they do that in stages, and they have to call them back into the factory for the next stage of assembly. And they can't actually they don't know which truck is which, so they lose trucks. They lose trucks in huge fields because they're all white, and they have a team of people who run around zapping barcodes on windshields to find a particular truck that needs a certain extra added to it because of the configuration. But this is a similar concept is you're finding x portable X-ray machines, ultrasounds, important bits of IV pumps, patient monitors, wheelchairs, pads. Whatever that hospital they use for within the hospital, we track them. And you say it it's down the corridor in the room on the left. Yes. That's KPIs. Now what's interesting, certainly when I spoke to you previously, is that a lot of systems that we see stop at the KPI. They say, here's an asset tracking system. Look at the data. Look at the KPIs. You can get an efficiency by finding assets quicker. But what I know you're doing is you have a series of AI tools. I'd like you to because you're now taking it up the next layer of the stack, which is the Agentic AI using the KPIs, using the data, because the AI is only as good as the data you feed it. So you're collecting step one, collect all the data from everything time, and then you're feeding it up into an LLM, and you're using Agentic AI. So let's break that down for the listeners in terms of what It's a very interesting subject. Yeah. It is. Yeah. So we are building agents in the background. So we do have the system to collect the data, as you said, and we which is the most precise data you can get from as a as location wise and the sensor wise, we obviously have the ability to collect them. We collect them, restore them, and we feed them to an AI agent in the background, such as proactive asset management. There is an AI agent in the background that monitor current usage in real time. And so staffs are going to check manually. This AI agent can check and predict and device are underutilized. It recommends reallocation. This will basically increase the utilization and prevent new purchases. And, also, operational workload as well. Okay. Oh, obviously on the work. That every asset that we within the healthcare sector, actually within any industry, almost requires annual maintenance or maybe weekly maintenance, calibration, cleaning, and storage. And these are all operational costs. And these are couple of hundred dollars per device per year, let's say. And this is basically this can be prevented by using an AI agent in the background, which is offloading the operational workload from staff to the AI agents that nurture real data. Again, workflow optimization, we do have in the background. We meant we basically follow the workflows. It can be patient or staff moments. It can be interaction between staff and asset, or it can be interaction between staff and patient or asset and a patient, which is quite common. We can tell which asset is used for which patient. So this interaction is very the the data in this interaction is very rich, actually. We can extract the information if this assess needs to be maintained, needs to be cleaned, disinfected before we use to another patient. We can put in a work order for cleaning to the bionic team. Can you can prevent using this particular asset. If it is not calibrated, it's not good for use, you can prevent that. And you can flag all the bottlenecks in the background. You can put in beds if a certain patient is waiting for a room to be cleaned, you can basically in the background, you can optimize his workflow. You can put in a cleaning order, and you can reduce the wait time. Everything is in the data, actually. And It helping agents Excuse me. Because you're saying some really interesting stuff, and I just want to make sure because most people aren't doing this. So you're living it. You're living it and breathing it. So I want to really make sure people are understanding this. Because the Agentic AI, which is fed by the data, and that's a series of, as you said, workflow optimization capabilities that run-in the background. And so the idea of a patient is finished with a procedure and needs to, I don't know, be put in a room for a period of time, but the room's not clean. And the person who normally would be wheeling the patient on the trolley down the corridor doesn't know the room's not clean, gets to the room's not clean. So then we've got to find another room, people end up in corridors. Lots of examples. I can think about a personal example. Occasionally, I've ended up in A&E, probably when I've done something stupid or something's happened to me. And I'm thinking, why does it take seven hours? And I'm talking from a UK perspective, but I think it's pretty bad over the world. And if you actually look at why it takes seven hours and sometimes it takes seventeen hours, Most of that time is you're waiting for the next step in the process. You're sitting on a plastic chair in a room with a terrible TV and a coffee machine that doesn't work, and you're waiting for the you get triaged very quickly. That's quick, and you oh, this is going to be great. And then you wait, and you try and work out who's ahead of you in the line. And then you see the doctor, and then the doctor wants you to see somebody else, and then they do some tests, and then they have to wait for the results. And the other and then at the end, they say, okay. This is what it is. You just go back to your chair because your medication we need to get your medications. Don't leave until you get your medications, but then you wait again, and you wait again. Yeah. And so there's all these process inefficiencies. And what you're and that makes up the bulk of the seven hours. I think it probably been Yeah. Six of the seven hours. What you're saying is that that, I think, keep me honest, is that as a result of collecting the data, you're not only getting the KPIs, but what the agents are doing, because they're reacting real time, and they are changing. Agents are dynamic. Then it's not like SAP workflow or any of these ERP systems where you define the processes, and the processes never change. What you're saying is that the data is dynamic. The agents, they could be a procurement agent, be behaviour agent, a support agent to take support from reactive to proactive. Let's do an update because we know it's going to need an update, and that'll all be automated. Lots of different agents. These agents are actually then able to map out the dynamic again, keep me honest here, but the dynamic picture of the process for this operation, this procedure, could be a cataract operation, it could be what happens in A&E, it could be a kidney investigation, a kidney operation. So you're now getting optimized process information for what a hospital really does, which is execute collaboratively with people and assets and patients across a series of processes. And so you've now got this the Sims game is now much more than assets. It's a series of processes that are interacting with a big dashboard showing efficiencies. This is actually a series of very complex processes, and AI agents learn from this process, and they interact with each other. They update, and asset agent can update the workflow agents, so the workflow agent can create another action. So everything becomes hopefully, obviously, it will become like that. Everything will be connected to each other. Everything will talk to each other. And then the adaptive AI will solve the problems where you need this an actual human controlling the actual healthcare worker. It can be a nurse or a doctor. It needs to be It frees them up solve the problem. Be doing as a good to the admin. It gets rid of the admin and some of the bureaucracy. Yeah. Yeah. If I'm a general manager or a manager of a hospital, this dashboard is incredibly important to me because it can save me money, as you point out, and make sick save thirty percent. But it also enhances the patient experience. And the quality of the healthcare service that you are getting as a patient Yes. That's very important. Wait time, The quality of the asset that you that is being used with you is well maintained, is up to date, is calibrated. Imagine an IV pump which is not calibrated, and instead of pushing one liter per hour, it should push you three mils per hour, and it's a disaster. So they get a better healthcare experience Yes. Is ultimately what it's all about. But the other question I got is there's more than what you have a lot of hospitals as customers. In Türkiye, for example. Canada, the US, I know their areas. They In Canada, Middle East, and Türkiye, obviously, is our most of our customers are in, but we do have some customers in Europe as well. It's all around the world, actually. All around. And If I'm at where I was going, Burak, if I'm a if I'm a CEO of a hospital or the CFO or whatever, management team of the hospital, can I can see the data on my hospital, and I can I can I get the AI agents should allow the specialist to focus on what they were really trained for, which is to give healthcare? It gives a better patient experience, gives better healthcare, more importantly, and it saves a lot of money, eliminates waste, etcetera, etcetera. Can I actually see how well I'm doing in my hospital versus other hospitals? Yes. We can compare them. We do have customers that big hospital groups. They own multiple hospitals. They manage they track the processes with our system, and we can actually compare two different hospitals, buildings, managed by two different management team. And we can compare the PURE processes, their asset utilization rates. Even we can compare their performance on the operations, OR rooms, the OR rooms utilization rate. We can comp the length of the durations, the duration of the operations for Different hospitals. Different hospitals, different doctors for the same procedure for different patients. Obviously, we can compare with the post diabetic therapy, why it took two times longer than the other one, and those stuff. It usually is an operational failure. The operation didn't start on time because the disinfection team didn't arrive. They couldn't clean it, or maybe the asset was missing. They waited for the asset. Since we tracked the asset, we know the asset has been delayed, and they waited that time for the particular asset as well. So yes. Yeah. So there's even an HR advantage. And if you run six hospitals, you're part of a hospital group, you can look at the you can do performance measurement across the hospital. So it's a very interesting area. And as I said, TKI is something which, in our experience, it is just people are just realizing what it actually is. It's not ChatGPT. Only that. Yeah. But people think it is. Actually, because we do have a technical audience, not exclusively, but a lot, I mentioned at the beginning of the pod, you're using most you're not just betting everything on one model. Are you use can you just maybe describe at a high level which model? Yeah. Obviously, healthcare is a highly regulated environment. You can just put, get the information about staff location or access calibration data and put in a public LLM, actually. It's you do have compliance with GDPRH in the background. So we do use we do have our own datasets. We do have our own LOM in the background, and we use some public LOM as well. It's OpenAI is another favourite that we have. But for interactions with the end users, which can be a little bit more challenging because interactions can be nurse can be from a different country. They can use different words. We use particular limbs. But that interactions that require the data access to that particular hospital's data, we use on prem or maybe in cloud, the private alarms at home. Private agents. Okay. Yeah. A series of different ones, which is consistent. People say it's not one AI. You use different tools. Small. Yeah. And, of course, the humans are still interacting, but they're looking at the data. They're interpreting it. They're checking it with their knowledge. And they have the ability to A variety. Correct the one to all of them, actually. So we allow them to feed the give feedback to their particular AI agents so we can improve it. So it can be improved, which is how they do the accelerated learning. The robot's not yet in control. But on the let's finish because it's a idea of it's a very interesting area. And as you said it very briefly, as you said, this is healthcare, but, actually, there's been a lot of people listening to this saying, that's very applicable to the industry I'm in. But where do you think because you going back to your beginning, you've been in research. You've been eighteen years in what we now know as AI. You've done a lot of different things. You've got a bit very technical. You're I should have mentioned you because of your PhD, you're a doctor in the field. Where is it going? Because it seems like you're defining the whole operationalization of, in this case, the healthcare is it's almost like it's catching up with what would be normal on a very advanced manufacturing line, say, in automotive or Amazon warehouse. Would that be right? Yeah. Actually, yes. It can be an automation production line or Amazon warehouse. Let's say if a tool or part goes missing in a production line, the production stops. It's the same as in a warehouse. And if you don't know where the items are, you end up searching them. And, like, the everything is delayed. You're out to processing time is gets very bad. And the hospital faced the same problem, actually. Instead of those packages or delayed bars, these patients, the care is up to us. The waiting times just get longer. You end up waiting an ER, six hours for a small procedure. Tell us and they have to bring us the same precision and efficiency into healthcare and may we make sure that the staff can and and patients are always in the right place at the right time, in the right border of the order for workloads. And as you were speaking, I could actually make the case that it's more complex than a production line because you're automating. You're taking the philosophy of the just in time production, the production line, which is well known in manufacturing, or the robots working in the factory for Amazon with the humans just looking at screens and being told what to do next. But you're taking that you don't have members of the public wandering around the production line of your factory. Yeah. You you don't have people wandering around picking their own boxes off the shelf at the Amazon warehouse. You've got thousands, tens of thousands of patients. It's the patient component. It's the what I call the triumvirate. It it And the other thing is also you have the time component. With the wait time goes along, the patient's situation can change. It increase the risk of workflows to be disturbed or has to be changed. And you need an something in the background that adapts to the workflows, adapt to the changing situations because the problem that you are called you are trying to solve keeps changing, and you have to keep track and update yourself. You can just say, okay. This is the predefined workflow we have to follow it. No? The situation can change. And then that the the agent modifies, and it it can actually say, okay. Based on new data I've got, now this is the new workflow process. Okay. Listen. Let's finish. We could talk for a long time about this. Let's finish here. I want to come back to my I talked about the Coldplay analogy. I think you told me you've seen Coldplay recently. Yes. Right. Couple of weeks ago. Okay. I've seen them as well. It was amazing. The patient gets the the wrist the wrist bracelets. Now when you leave the Coldplay concert as you come down the stairs at Wembley or, I assume it was Wembley, that you there's people there collecting them. Please hand them back. Is it the same system for you? Do do you collect them back off the patients, and what percentage of patients hand them in? So, basically, the process is there. If a patient is going to be discharged, the nurse should go there and take the bracelet out and discharge the patient. If they don't do that, they can't be discharged. If there's an alarm, they also because they also don't enter the patient for their own safety. Yeah. Like taking the cannula. The yeah. Yeah. I mean, the patient shouldn't be leaving that part of the room. Maybe it's an infectious patient that has to be current. We also track those, and they can't take out their bracelets. They have to cut them, and then if they cut them, We monitor the bracelets condition, and we get a warning in the background. They are trying to tamper the tech. So we don't use the techs. Obviously, like, maybe one or two percent is being lost or being stolen, but we don't use the text because using text mode using the patient, we don't lose the patients as well. We make sure the texts are attached to the patient from admission to discharge. Okay. And and they're taken off and they throw throw cut them off, throw it away, and because they need to be sanitized. So listen. I think we should stop it here. It's a very interesting area. Just to recap, the company your company is called Borda Technology. And I know you've got on your website quite a few case studies of hospitals, healthcare institutions that are using the system. And I also know that the opportunity in this particular area is huge. Healthcare is, I think, is the single thing every government spends. Healthcare and defence, but in is where they put most of the money in it, and it's just accelerating as we get older and more demand for treatments. But it I do believe also it is actually it's not healthcare specific. The concept of what we talked about of the IoT, the data, in in your words, the assets, the staff, and the people or the customers is actually generic across many industries. And the operationalization of a huge complex entity like a hospital and turning operation operate operationalizing through technology a hospital where the staff are following the instructions, like the Amazon warehouse where it's on the screen that says do this next, is a very interesting idea and has huge ROI. I want to thank you, Burak, for being on the pod. I think it's a particularly interesting one. And for anyone listening, please visit Borda Technology, and you get a wealth of information there in the case studies. And as I say, we do have listeners all around the world, and I think every country where we have listeners wants to solve problems like this. Thanks again for being on. Good to talk Thanks, Nick, for inviting me. You're welcome. It was a lovely talking to you. Yeah. Hopefully, the listeners will get the same. I hope so. And while they, and I know we don't want to talk about the NHS in the UK. We can't go there. Because then we I'll really get frustrated. But one day, let me just say, one day, I really hope we have a British case study that actually talks about this. Yeah. Yes. I mean, we are looking for that, actually. I'll fully able to do very good use case with San Andreas in here. Because, boy, do we need it. Boy, do we need it. Anyway, thank you, Burak. Really appreciate it, and, thanks for being my guest on the IoT Leaders Podcast. Thank you. Thank you for your time.