Mary Sheridan, Senior Manager for the Accelerate Health team here at HIMSS interviews Chooch AI CEO Emrah Gultekin about the work Chooch AI is doing with computer vision for healthcare . In this 15 minute podcast Gultekin explains that “in healthcare, you have lots of visual tasks. Whether it’s cell counting or whether it’s patient gestures, or maybe operating rooms where actions are happening. These are the kinds of things that we have used Chooch for and is being used in production in many, many different healthcare scenarios.”
.”
Mary Sheridan:
Hello, changemakers. Thanks for joining us for another episode of the Accelerate Health Podcast this week. I’m Mary Sheridan, Senior Manager of Accelerate Health here at HIMMS, and I’m so very excited to be filling in as guest host for this special edition of the podcast. It is the beginning of our monthly showcase of highlighting outstanding startups and digital health. As we lead our way up to the HIMMS Global Conference and Exhibition at Las Vegas in August. Today, we are joined by Emrah Gultekin, CEO of Chooch AI. Hello, Emrah. Thank you so much for being our first guest on this very special series.
Emrah Gultekin:
Hi Mary. Great to be here today. Thanks for doing this.
Mary Sheridan:
Absolutely, and so, as we were talking a little bit earlier, you are the first in this inaugural series as we lead up to the larger HIMMS ’21 Global Conference, which as we know, highlights everything in healthcare information and technology. What the big thing that we want to talk about are all the great innovations, all the great startups, and you guys were the first in the series. I know there’s a lot to talk about because of all the great work you guys are doing in this space, but I really want to take a step back and say and ask you, what is Chooch AI? It’s a great name, a fun one that you don’t normally hear in healthcare. Kind of walk me through that journey of how Chooch got started and what you guys do.
Emrah Gultekin:
Yeah. Thanks. Thanks Mary. Yeah, it’s an interesting name and Chooch means it’s a mix between search and choose. It also means dummy and in an old Italian dialect, so it’s kind of like a pun on artificial intelligence. We are an artificial intelligence company based out of the Bay Area, and what we do is visual detection. Basically, what we try and do is we try and copy human visual intelligence into machines, and it’s a large platform that does that. Taking everything from the AI life cycle, from data site collection to model training and to inferencing, which is when the AI predicts something, that’s what we do as a platform. We’re a horizontal platform, which means we’re in many different spaces and healthcare is one of them. Yeah.
Mary Sheridan:
That’s great. I know Chooch has been pretty successful in several industries like retail, manufacturing, safety, security, even digital media, and now healthcare. Can you explain a little bit how you guys got into the healthcare game? Did that come first or did you see that those applications could be applied to healthcare as well?
Emrah Gultekin:
Yeah, it’s a great question. First, we started mainly in the media business. Basically, like tagging images and tagging of videos. Basically, that’s what it started with. It’s a very interesting way. My co-founder who was developing this way before Chooch, did it for the care industry. He had developed a visualization of radiology before developing Chooch AI. That was for the healthcare industry. That was for doctors mainly in Asia, and places where they had to take care of like large amounts of patients a day, let’s say 100 to 200 patients diagnosed a day. We try to take that and say, “Can we generalize some of this? Can we generalize detection?” What we did was we generated that platform. It did kind of start on healthcare initially, and it went full circle and we did receive a lot of healthcare requests over the last two, three years.
Emrah Gultekin:
We started developing more stuff for the healthcare industry. When I say develop, it’s really, the platform itself is agnostic. It’s a visual detection, a visual AI platform that could be used for almost any purpose. When you teach the AI something, you’re teaching it to detect certain things. All these human visual tasks are being automated by this platform. In healthcare, you have lots of visual tasks. Whether it’s cell counting or whether it’s patient gestures, or maybe operating rooms where actions are happening. These are the kinds of things that we have used Chooch for and is being used in production in many, many different healthcare scenarios.
Mary Sheridan:
I think that’s wonderful, especially how you used it in other industries to apply to healthcare. Like you said, it came full circle. There were other uses you found after being in healthcare, going into other industries, and then coming back around to find other uses, because we talk a lot at Accelerate Health about how you said the larger consumerization of healthcare, how people are trying to get more access, get using things that they wouldn’t normally think from healthcare, but we’re looking to other industries to see how we can improve upon healthcare. I think that’s extremely timely and very cool. What would you say differentiates Chooch from other AI companies in the healthcare market space right now?
Emrah Gultekin:
Yeah, so we wouldn’t consider ourselves a healthcare service provider, just to healthcare service. The basic thing here was how do we make the machine mimic human visual intelligence initially? Very, very basic things. When we took that problem and we try and solve it. It became kind of like an agnostic platform. If you don’t do that, it doesn’t matter what you do in healthcare or what you do insecurity or what you do in other places, you need to take care of that capability first. We’re like, “Can we solve that piece?” Then you can augment other issues on top of that, whether that’s in healthcare or security or other verticals as well. Basically, the flexibility of the platform is really crucial here, the way it takes the entire life cycle.
Emrah Gultekin:
If you divide AI into three main processes, three main buckets, the first bucket is data set generation. In data set generation, you are creating a data set where you will start teaching the AI to do something. That in itself is a very important aspect of the AI life cycle. It’s also crucial that you’re able to do that on a single platform. The second bit of that is the second bucket, or the second process is training the model, so generating that model. That has to be done with with certain elements and with your data set that you had initially generated. The third process and the third bucket, which is probably the most important one and that’s what everyone sees is inferencing, which is when new data comes in, how does the AI predict what’s going on?
Emrah Gultekin:
That part of it is also part of our system. This whole package is really crucial for the acceleration of AI in any space and also, healthcare and the scalability, the reproducibility of that is really crucial as well. That’s what we’ve got. We’re not healthcare experts, and we never said we were, but it is being used in healthcare. We are an enablement technology for smart ORs, we’re enabled technology for cell counting, we’re enabled technology for public safety and security and so forth. Unless you have that entire platform, which is flexible, reproducible, scalable, you’re not going to get anywhere with any type of AI project that you have. It’s very crucial to have it. We’re a part of that ecosystem. Let’s put it that way.
Mary Sheridan:
Yeah, absolutely. Yeah. We work with some startups where they have something really great, but if they don’t know how to apply it correctly, it doesn’t help the user. It looks cool but in the end, it’s not sustainable enough, but it sounds like what you guys are building will be, is incredibly sustainable. That’s exciting. That’s the type of innovative tech that we definitely need. I know you touched a little bit on smart ORs. Can you talk a little bit more about that? Is that the biggest scene you guys have developed within the healthcare space, or are there other items you guys are exploring or have already done?
Emrah Gultekin:
Yeah, so smart ORs is our major, let’s say product differentiator that we have. Basically, it started out with understanding of the actions in the operating room. There are dozens of actions happening at the same time. Doors opening and closing-
Mary Sheridan:
Yeah. Apply it to that.
Emrah Gultekin:
We had a panel the other day. It was like applied AI isn’t that sexy, but it’s important. Many times the doors open, or ho walks in, who walks out? When does incision start? When does anesthesia start? When does it stop? When is the operation over and when do the cleaners come in? These types of actions are things that we log on our system in the operating room. What that does is it helps send alerts to the operating team. If there’s something that’s wonky going on, like if there’s things are not timed properly and also allows the nurses and the surgical team to focus on the patient, instead of trying to like do different, trying to use 20 different softwares to log things into.
Emrah Gultekin:
The third thing is they share the best practices then with the other surgical teams. You might have a surgical team performing really well, but you don’t know why they’re performing that well. If you’ve recorded all their actions, then you understand, maybe it’s a timing thing. Maybe it’s the way the incision is being made, or it might be something else going on. Those best practices can be shared between different surgical teams as well. We’re at the beginning of this. We’ve launched it and we’re enabled to move forward, but we’ll see how it develops over the years and how deep AI can be integrated into those activities as well.
Mary Sheridan:
Right, and like you said, it might not seem super fancy or what people consider AI, but it’s pretty meaningful to the people that are operating on you and ultimately the end user, the patients, the individual so it’s, yeah.
Emrah Gultekin:
We think of AI in a more science fiction type of thing, but we’re taking baby steps towards that mission and towards that vision. Applied AI today, we do things, it seems simple like traffic lights or smart license plate reading, or you could do mask detection or those types of things would just seem very, very mundane, but that’s really where it starts. You got to get those right first until you try and change the world,
Mary Sheridan:
Right. No, that’s great. Amazing stepping stones. What else do organizations need to know to adopt computer vision with Chooch AI or to just adopt your platform in general?
Emrah Gultekin:
Yeah, so we are at the beginning of this transformation and enterprises. The first thing that we help our clients and our partners do is define their use cases. What visual tasks do they want to automate, or what visual tasks can they think that a machine, if a machine does that it would be more efficient for them to do? Those use cases have to come out and we help our clients do that. We really define a use case, which is low-hanging and easy to deploy. Start with that, then you can use some of the pre-trained models that we have, or if you want to build custom models, we also do that on our system where you can do custom models. It’s really important to define a very particular use case that you have, and then say, “Okay, can we teach the machine to do that?” Then we take it from there really.
Emrah Gultekin:
It’s kind of open, but it’s also collaborative with our clients and with our partners. We encourage them to use our pre-trained models just to see how it works. Then how they integrate that data into their systems, because this is new data. This is data that did not exist because it was all humans doing it and humans, they can tag things here and there, but you can’t really tag 24/7 feeds that are coming from different machines that you have. Really what do you do with that data? Then how do you send alerts and how do you do the data mining for it is really what we work on with our clients. Really define what you need. It’s something new, you’re not replacing it with anything else. It is kind of a new area for these enterprise and we work with them to define those use cases.
Mary Sheridan:
Are you guys having them start in the OR for now, because that’s where a lot of your models apply easily? Obviously, those are the stepping stones where they could build that foundation applied to other areas.
Emrah Gultekin:
That’s right. The OR is a very specific field, right? The OR, it has PII and all that kind of stuff. You can start from the OR, but you can also start from the lobby, or you can start in another location as well. I mean, you could basically start anywhere you feel that you have the need to deploy these models.
Mary Sheridan:
That’s great. Is there anything else that you want us to know about Chooch or anything coming down the pipeline you guys want to let us know?
Emrah Gultekin:
In general, I think it’s important for us to deliver the message of AI in a positive manner. I think we’ve always been sacrificing efficiency for equity and equity for efficiency for many, many years. We’ve had to kind of make those hard decisions as we move forward as companies and organizations. I think with the onset of AI, we have the capability to not do that. We can have efficiency and we can have equity at the same time, and this goes into healthcare and like public safety, security, and all that as well. Just want to send that message to people who kind of have some doubts about AI, or have some fears and who don’t want to adopt it so quickly. There is a positive outcome for everybody for AI. That’s my message to everyone listening.
Mary Sheridan:
That’s great. That’s a great place to start. Absolutely. Well, thank you. Thank you so much, Emrah, for joining us today, giving us a little snapshot into Chooch AI and of the amazing work you guys are doing. I know it’s only just the beginning. If anyone is interested in learning more, we will include information on the podcast description, but also you can visit Chooch.AI and you can learn more information and watch out for their article coming out next month from Chooch on computer vision in healthcare on HIMMS Resource Center. We will send out the link and promote it on Twitter as well, and be sure to visit them at the booth in our startup park. If you can make it to the HIMMS ’21 Las Vegas. There’s also a digital component if you can’t, where you can also check out their resources and information what’s going on with Chooch. Again, Emrah, thank you so much for joining us. We are excited to see everything that’s coming out through this year and more, and thanks again for sharing your story with us.
Emrah Gultekin:
Thank you. Thank you for having us, Mary.
Learn more about about Healthcare Computer Vision
