What's Up with Tech?

The future of Asset Management: Cognosos on AI, IoT, and Real-Time Tracking in Healthcare

• Evan Kirstel

Share episode

Interested in being a guest? Email us at admin@evankirstel.com

Curious how AI and IoT are revolutionizing healthcare? Join us as we chat with the Chief Product Officer at Cognosos, to uncover how they are transforming asset management in hospitals. Learn about the latest advancements in real-time location systems (RTLS) that leverage AI and machine learning to solve the age-old problem of locating mobile medical equipment. Discover how these cutting-edge solutions not only streamline operations and reduce costs but also ensure that critical medical devices are always accessible exactly when and where they are needed.

In this episode, we also dive into the implementation of an innovative overlay system featuring Bluetooth beacons and a proprietary 900 MHz communication backhaul, promising seamless asset tracking with minimal maintenance. We share insights on how this technology parallels asset management in automotive manufacturing, demonstrating its versatility across industries. We also explore the potential of crowdsourced infrastructure technologies, such as Apple AirTags and Amazon Sidewalk, to further amplify tracking capabilities in both consumer and industrial applications. Don't miss this compelling conversation on the future of efficient, technology-driven asset management!

Support the show

More at https://linktr.ee/EvanKirstel

Speaker 1:

Hey everybody, fascinating chat today, diving into the world of asset management, staff, safety and hospitals and beyond with an innovator in the field Cognosus. Hey Deerian, how are you? I'm great, I'm great. Thanks for having us on. You're doing that's kind of at the intersection of AI and IoT and other technologies.

Speaker 2:

Before that maybe introduce yourself and your role at Cognosys and the mission of the company. Yeah, so I'm Chief Product Officer CPO, which I often think just stands for Chief Prioritization Officer honestly trying to juggle the various demands on our roadmap, which is always fun. But yeah, I've been doing this. I'm in my 25th year here of finding things with technology for various reasons, so I've been at it a while and my mission and Cognosus' mission kind of came together. Cognosys mission kind of came together and in fact, that's how I ended up at Cognosys, where my background has been in like super high-end location in ultra-wideband and not just ultra-wideband but the high end of ultra-wideband with the previous two employers and cool technology, huge amount of value, but huge amount of cost associated with it. So it's kind of a niche technology that has its applications.

Speaker 2:

But what I was seeing at the time that Cognosys kind of came calling, what I was seeing was that the RTLS industry had kind of let the market down in a lot of ways. It was far too expensive and difficult to deploy these things. It's been the kind of the bane of IoT, hasn't it that? It's a huge amount of value, but, gosh, how do you get started? And here was Cognosys with the mission of how do you make it easy, how do you create the high performance of some of the legacy systems, but without all the incumbent cost and complexity and just difficulty of buying and owning these things. So yeah, at a time where I was getting a little jaded, here came Cognosys with a fresh approach, and that was four years ago, and I dare say there are many more in front of us together.

Speaker 1:

Fantastic. Well, let's dive into the solutions and the problems you're solving. Yeah, in particular an area near and dear to my heart healthcare, healthcare, it, yep, we've all seen hospitals overwhelmed, overloaded, all kinds of challenges today, and you solve for asset management. How does that work and what's the problem statement and solution you guys bring?

Speaker 2:

to bear and solution that you guys bring to bear. So this is like the age-old problem of RTLS, real-time location systems and even back in gosh, probably the mid-2000s, one of the very earliest adopters of this technology were healthcare systems to solve this particular problem. So it's not a new problem. There just hasn't been a very good solution so far. So the problem is simple Hospitals have a lot of mobile medical equipment.

Speaker 2:

When you need it, you need it. So if it's not on hand, there's no time to go searching, and what hospitals have historically done is just purchased way more equipment than they really need. And if you buy enough stuff, you can pretty much trip over what you need in any hallway. That's the thesis. But it's not cost effective and in a time when hospitals are struggling I mean I think there was a report I forget the exact numbers, but coming into the year, coming into this year 2024, think about at least half the hospital systems in the US were losing money. So it's not a new problem, but there's a new focus on how do you better manage assets.

Speaker 2:

And again, the history of this marks a little bit of my kind of personal belief in the value here. The history was gosh. If you can't find something. What about giving nurses a thing finding app? What if nurses could pull out a phone and search for what they need and find it and go straight to it?

Speaker 2:

Well, that's not really fixing the problem, is it? That's a band-aid. Wouldn't it be better if everything was just always where it needed to be, right when you needed it? So you know, the new era of asset management in hospitals is very much focused on how do we help people improve the process of managing, cleaning, maintaining and staging all those mobile medical assets right where they're needed. He got that picture of the operating room right and the surgeon reaches out his hand and says scalpel, and the nurse knows exactly what he needs next and he brings his hand back with the scalpel in. And wouldn't it be great if a nurse could reach out a hand and said wound vac and boom, there it was. That's the mission. How do you make sure everything is at the right place at the right time, right when?

Speaker 1:

Oh, that would be quite a dream. And you don't just you know locate devices and items, medical equipment, med tech but you also provide the intelligence via AI around the process of putting that device into the right room, into the right patient. How does AI work and machine learning in your solution?

Speaker 2:

Yeah well, so it's kind of interesting. This is what intrigued me when I joined Cognosys. This was the thing that sold me on the mission, the mission being all the value, none of the cost, which you know. You've been around in technology for a while, you've heard that a few times and you think what exactly are you guys doing different? So the thing that cognizance is doing different is ai, right in the core, right in the senses, actually, and um, traditionally, what's happened is you've got a few different classes of locating technology and what you really need in a hospital is room level accuracy. That's the kind of the standard of usefulness. Can I tell you when to go and replenish a clean room when it's getting low on IV pumps? Well, you know what? Yes, I can, if I can track them and I know where they're in the room and I can count how many are in the clean room and let you know when it's falling below a threshold, right. So room level accuracy is this kind of threshold performance limit. And previously, the way you got that was a huge amount of infrastructure, all kinds of installed in the ceiling, tracking infrastructure, some with batteries that needed awful maintenance scheduled, some that you had to run cables, to goodness me, which is incredibly cost prohibitive in a hospital. So what we do is we use AI to actually create that level of accuracy.

Speaker 2:

We use Bluetooth sensors, which is very common now in healthcare. Bluetooth is cheap. It doesn't work that great. I mean, its accuracy is maybe oh I don't know 20, 30 feet or so. That's kind of cool. It's not that great when you think that 20 feet goes up as well as a long. So you might not even know what floor something is on, so okay. So you might not even know what floor something is on, so okay.

Speaker 2:

But what we do is we take AI and we listen out for Bluetooth signals from our network of beacons and you kind of fingerprint a room. The machine learning algorithm learns I think of it as fingerprinting all the smell of the room or however you like to think of it, but it learns the room, just in the same way that if you feed a bunch of pictures of cats to an AI and say these are all cats, and you feed a picture of your dogs and you say these are all dogs, and then you give it a picture and it says ah, that picture is 95% in agreement with the model I've developed of catness. Therefore it's a cat and that's a machine learning algorithm doing classification. And if you teach it room oneness and room two-ness and room 273-ness, then the AI can recognize those rooms and it can do it with very high confidence and it can do it with very little input and it can do it with very little input data. Ai is really really really good.

Speaker 2:

We've all kind of noticed it come of age it happened last year. It's doing amazing things. It makes very high quality inferences on not much input data. What does that mean? High performance, room level accuracy. Not much input data is not much infrastructure, just a sparse array of Bluetooth beacons that is super easy to put up there. They last forever on batteries, don't even worry about maintaining them. Job done. So it kind of broke the paradigm. It broke the high cost, high performance versus low cost, low performance paradigm. You've got this high cost, high performance, lowcost solution that AI is delivering. It's pretty new and pretty cool. It made me excited it sounds like it.

Speaker 1:

Well, it sounds very Star Trek in terms of what it can do.

Speaker 2:

Ten years ago, you wouldn't have believed it possible, and when the recruiter called, I didn't believe it either. I'm not going to lie, and from a tech standpoint.

Speaker 1:

What about locating people and not just, uh, fans? Is that part of the equation here and um, how does that work?

Speaker 2:

yeah, um, so, so it works in exactly the same way, um, and that's a different. It's a different kind of form factor. You know, if you're trying to track an asset, you want some kind kind of tag that's going to kind of double-sided tape onto an asset or something. If you're tracking people, you want it more like a more like your ID badge, you know, your access badge.

Speaker 2:

So, but inside same technology, same AI, same AI principles. Actually, we actually developed a new AI model for people. Just because people move differently than assets, do you know? Assets move and then sit for a while, and then move, and then sit for a while. People tend to kind of be moving around. So that's the beautiful thing about AI it can detect what's moving, it can understand oh, this is a nurse walking down the hallway. This is a patient laying in a bed. Let's throw the right algorithm at this situation to do the best. Locating that we can, all those decisions made up in the cloud where you can throw giant-sized brains at it, which is cool. So, locating people not so very different than locating assets, but the use case is very different and unfortunate. But we're proud to help here.

Speaker 2:

Which is workplace violence. Medical staff, and this in particular, are more prone to being attacked on the job than in any other profession, including the police. Wow, they don't give nurses hand to hand combat training and sidearms and bulletproof vests, right. So it's a real problem. It's a growing problem. It's a. It's an awful problem, um, and we've applied our technology to giving um medical staff, whoever it might be, the.

Speaker 2:

The frontline workers are key, but? But everybody in a hospital is vulnerable to this um a badge where they can very discreetly call for help, where the incident response team knows exactly where they are. If they move, they know where they're moved to. There's no point responding to where they just were. That works in rooms and hallways and stairwells and basements and parking lots and on the campus and the pathway between the building and the parking lot, which is not a fun place to be at night if somebody's after you, you know. So full coverage of the entire campus and beyond, actually, with high-quality incident management and pinpoint location response all built in. And that's something we're very proud to bring and we feel like it's an important part of what we do, for sure.

Speaker 1:

Yeah, what a wonderful mission, and it must be challenging to implement, in a hospital environment in particular, given all of the legacy IT infrastructure. We call it technical debt. It's basically a spaghetti system of outdated technologies. How do you integrate?

Speaker 2:

into that environment? We don't. We stay the hell away from the spaghetti for good reason, right? So, first of all, hospitals have trained, they've tuned their spaghetti of technologies to achieve the mission, and adding new use cases in is, uh, is never a good thing and it's also a very complicated, time-consuming and often costly thing. So, um, so we have a complete overlay system.

Speaker 2:

Actually, um, we, we deploy, as I said, a fairly sparse network of bluetooth um beacons in the ceiling. You know, think every, you know, 30 odd feet down a hallway, a beacon and some critical rooms you might put a beacon as well, but that's it, and those things last gosh eight years, I think, plus on a battery. So, you know, maintenance, not that big a deal. And then, in terms of data, obviously we have to have a data link back to our cloud-based ai and we have our own proprietary communications backhaul that connects to the devices, connects to the tags and badges worn by staff and that operates at 900 megahertz. It operates over very long ranges with very low power consumption, as you need in these IoT systems, right, everything's battery operated. And it came out of a background of satellite communications, the poster child for long range, low power consumption communications. That's the technology we've encapsulated there.

Speaker 2:

So that goes back to a, you know, a device of our own, a gateway device, and that goes straight to cloud and we have we have no touch points with with hospital IT. Basically, all we all we require is for the hospital to provide firewalled internet access for our gateways and that's it. So very simple to deploy, very quick and easy to deploy. It's not a giant IT project, not a giant construction project, and doesn't get entangled in all those other things, so it's very quick and easy to overlay. We just did a I think we just did a three and a half million square foot facility Took us 12 weeks start to end. Effort facility took us 12 weeks start to end, and you know that's traditionally a year-long install for one of these. So, yeah, that's the value of not getting tangled in the spaghetti, as you say.

Speaker 1:

Very nice, almost unheard, of a three-month project these days. Well, we're changing the game you are and your technology, of course, is being used for hospitals, but also other industries. Talk about other use cases, industries. The world must be your oyster.

Speaker 2:

Well, yeah, and even more and more these days. So we do a lot in logistics and that focus hitherto has been on manufacturing and we do a lot of business in automotive manufacturing, especially on their outbound logistics. You know, you kind of think that you order your car and it comes off the assembly line and it gets inspected and gets a big checkmark and they ship it right to you. And well, that's not how it happens at all. It goes through a highly complicated dance of all manner of different locations and trains and trucks and accessorization and gosh, all kinds of stuff. So hugely complex process and we help monitor and manage that entire process. Monitor and manage that entire process. You know it's not. It's. It's wildly different from managing assets in hospitals. But then again it's exactly the same it's. It's wildly different because you know it's a different environment. It's outdoor, it's 100, 200 acre yards of 20,000 vehicles parked out there, rain and shine, day and night. But the problems are the same. What's asset management problem in healthcare is how do you pre-stage the correct mix of devices in a clean room so that nurses can always lay their hands on them when they need them? Well, think about a marshalling yard for a car factory? How do you pre-stage the right mix of vehicles in the rail loading buffer so that, when the train rolls in, the loaders have the exact cars they need to put on the train? It's the same problem, which is why it makes strategic sense for us to do both of these things, and we're extending that into inbound parts and parts distribution, and we're extending that now to reusable transport containers, racks and totes and all the things that kind of just go around a loop the large container that takes eggs from the egg farm to the bakery right, the racks of containers that take finished windshields from suppliers to car plants, and so it goes on. So the whole world of logistics. We apply the same AI technologies to get great performance at a very low cost, easy to deploy systems.

Speaker 2:

Yeah, and they're kind of merging too, evan, you know hospitals loss prevention is one of the use cases for asset tracking. Let me know if they're heading out the door or down the laundry chute or out in the trash. That happens a lot. You know, you clean a room, you wrap up all the bedclothes and you don't realize you're wrapping up a telemetry monitor in there and it goes down the laundry chute and there's a few thousand dollars out the door. Well, where did they go? So tracking assets out in the world people care about, that's a logistics problem. Home health is becoming more and more of a strategy. Well, if you're sending assets home with people and having people make home visits, you'd better be able to track assets anywhere and you'd better have a safety system that works anywhere. So our logistics and healthcare not just use cases, but technologies are converging as well.

Speaker 1:

So it's a really interesting mix what we do, oh, fantastic, fantastic and given the rise of you know technologies, like you know Apple tracking devices and Android tracking devices kind of the public consciousness has risen a lot and now you have you know 5G and satellite tracking. Where do you see the technology heading in the next few years?

Speaker 2:

What kind of things are you doing to stay ahead of the curve here? We're going big on crowdsourced infrastructure. So Apple AirTag you mentioned that's a tracking problem with crowdsourced infrastructure and the infrastructure of iPhones and that's very much a consumer application. You can't rely on that for the kind of industrial and enterprise applications that we're talking about. But there are other technologies out there. There's things like Laura Wan, there's things like Amazon sidewalk. Um, that's uh, a really fascinating zero cost. Nationwide, 90% of the population covered crowdsourced data infrastructure. So we think those things are incredibly important in logistics and healthcare too, and we're front and center in our roadmap is incorporating those technologies.

Speaker 1:

Yeah, oh, that's going to be exciting to watch. What about you personally, professionally? Dog days of August here. What are you looking forward to the rest of the summer? What's next for September?

Speaker 2:

Right now.

Speaker 2:

I'm looking forward to hanging on vacation for a couple of weeks back to England to see the family and then fretting about leaving our Florida property in the peak of the hurricane season. What do I look forward to Just here at Cognosys? We've got some big projects we're about to get underway in healthcare and logistics, some game-changing new logos. We're growing fast. We're hiring like crazy. We're at that phase in our growth here where we're looking ahead thinking we've got to nail this and get this right. We've got to grow at exactly the right speed. Stay ahead a little bit.

Speaker 2:

You know this technology industry, evan. You've got to be able to support your customers properly and you need the people to do that. So we're enjoying hiring, we're enjoying building up the team. We're enjoying all the opportunity. We're having a lot of fun in the roadmap integrating these technologies.

Speaker 2:

And for me, you know what I'm excited about, which is this whole AI in the census thing that we're doing. It's new, it's different, it's amazing. In our first go around we've reached feature parity with the highest end legacy systems and we're just getting started. We're doing some really cool stuff in the lab, looking forward to bringing that into products. As we leverage more and more of AI and as AI gets more and more capable, our products just keep getting better and the beauty of the new. People have really bought into this. Saas cloud era now Means every time we make an improvement in how we do things, we can roll it out overnight to our entire customer base. We don't have to jump through hoops to keep everybody upgraded or go back and charge people for the latest version. We just flip the switch to the new algorithm and everybody gets better. So yeah, lots more to come from us, I'd say.

Speaker 1:

Fantastic. Well, here's to your continued success. We're all rooting for you making hospitals especially safer and more efficient. Couldn't ask for a better mission than that. Thanks, adrian. Thanks for joining and sharing the vision.

Speaker 2:

Absolutely. Thanks for having us on. It's been a lot of fun.

Speaker 1:

Thanks everyone for listening and watching. Take care, bye-bye.