← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
From Copilots to Autopilots: Building the Connected Operating System for Healthcare
00:11:28
Loading summary
Transcript19 lines
- [00:00]
A
Hi everyone, this is Lucas Voss with Becker's Healthcare. Thanks so much for tuning in to the Beckers Healthcare podcast series. Excited to have you. Today we're going to talk about the future of healthcare platforms, connecting people, processes and performance. And I'm so excited to be joined by Tanay Tandon, CEO at Kimmure. Tanay, thanks so much for being here today and taking some time.
- [00:21]
B
Thanks so much for having me. I'm excited.
- [00:23]
A
Absolutely. I want to introduce you to our audience, or rather have you introduce yourself. Could you introduce yourself to our audience and just share a little bit about your role?
- [00:33]
B
Definitely. I'm Tenay. I'm CEO at Kamur, which powers tens of billions of dollars worth of healthcare payments, tens of millions of annual AI powered healthcare appointments every year using our ambient copilot deployed in 40, 50 health systems, over 2,000 sites of care. And I think, really, if I was to summarize what we do in a sentence, it's taking the massive power of language models and, and deploying them in every corner of a healthcare system, starting with the top of the funnel, which we consider ambient scribing and documentation, all the way down to the back office workflows that in some ways end with revenue cycle and patient engagement.
- [01:17]
A
And I'm sure we'll touch on those here in a little bit. 2024 and certainly I feel like 2025 as well. The year of ambient. So many conversations around ambient AI adoption and it continues to be a priority for a lot of health systems across the country. What are some of the conversations that you're having right now with leaders about ambient AI? What impacts have you seen? What stands out for you?
- [01:41]
B
Ambient AI, one of the fastest growing healthcare software tams ever in terms of the proliferation of these products by a multitude of companies in every country, in every health system, in every practice had adoption that looked more like a consumer app, like a Snapchat or a TikTok than B2B software than we've ever seen in healthcare. Which made 2024 and 2025 really exciting because, you know, speaking for Kamir, going From maybe in 23, doing 100,000 appointments a year to this year annualizing to 2030 million appointments is shocking growth. Again, you don't see that usually in B2B software. And when you do, it means that there's been a technology transformation, there's some new capability that just didn't exist before. Now I think of the ambient documentation and ambient scribing transformation in two or three phases. One is those early adopters that were already on in 23, 2024 was the mass proliferation. And then 2025 really has been the late adopters in some ways the folks that were dubious of it or didn't believe the models would be good enough adopting it in many ways because their EMRs are now pushing it, or their EMRs have deployed stronger integrations, or in the case of Epic, the EMRs have directly deployed and launched their own competing ambient scribes themselves. So we're almost in that like chapter three now and in some ways like the most exciting chapter because the competition is fierce, but the product opportunities are even more immense for people that can ship product very quickly in areas that aren't just vanilla scribing.
- [03:20]
A
Yeah, and you've mentioned the growth, obviously and sort of that, the, that the unheardness of that growth. Specifically, when you're looking at some of the deployments that you've done that you've worked on, et cetera, was there anything that surprised you? Especially when you're looking back at some of the early things that you've worked on? Is there anything that you've learned throughout the process when you're looking back?
- [03:40]
B
The biggest learning to me, I think is, and this holds true for any product is the number of systems that we've gone to where they tried an ambient vendor in 23, 24, got really low adoption, like 10 or 20%, and then tried Kamir after having it hooked into more places than just ambient scribing, having it hooked into patient engagement, document summarization intake and parts of the revenue cycle stack, and suddenly had adoption shoot through the roof closer to the 60s, 70s and 80% utilization by providers. And for me, I think that what that the story that tells is that the narratives that are out there on LinkedIn about ambient scribing just aren't accurate. Which is, oh, it's a race to the bottom or there's, you know, it's the winner is just going to be epic or there's, you know, there's really no more to do in this space. The reality is most health systems that sign ambient scribing contracts a year or two years ago have abysmally low adoption because it's not part of this operating system or set of tools that connect every part, every point solution into a single one. And I think 2025 has been and will continue to be that chapter in that story.
- [04:50]
A
Now you've touched on the integration piece and how important it is to, to have that the integrated experience for, for a lot of folks, which is key. And you spoke about 2025 and there's another AI, right? We talk about ambient AI so much because it is an important part of the conversation. But what's really emerged as well is agentic AI. Everybody's talking about it. We just came off of our hit conference in, in Chicago and this was a big talking point there too. Why are health system leaders paying attention to agentic models right now? And in what ways do you see them shaping clinical and operational approaches?
- [05:26]
B
It's interesting because agents as a concept aren't new. We've had chatbots since the2015. I think what is new is language models that are intelligent enough to process large chains of information and then turn that into large complex chains of action. And that's where agents have seen mass proliferation for us. That's in areas like appeals and prior arts and taking those tasks that might be actioned during the visit and then having a model autonomously go and conduct them. A stat that I always like to share with people is the in our revenue cycle platform, which really is an agentic language model. You know, maybe two years ago we, we had X number of people on our denials team and these are operators and support staff, et cetera. Deploying agentic LLMs in 2024 and then seeing their proliferation in 2025. We have 5x the dollars that flow through our system while keeping the number of people flat. And that's what's exciting here because again, healthcare has a fundamental labor problem. The health system's no longer have those labor pools and those skilled labor pools that they could access for the prices they could access 10 years ago. And so the only solution to serve their customers and serve a growing chronically ill population is deploy agentic LLMs that can handle actions themselves and go from being co pilots to physicians to being back office autopilots that can run lights out.
- [06:56]
A
Now you've touched on the fact that there's so much information out there, right? On LinkedIn, articles that get published, daily conversations that happening daily. There's also a lot of choices for health systems across the country really to choose a vendor, somebody that's providing a new service, etc. There's a lot of tough choices that have to be made in terms of okay, where are we going for from your perspective and the conversations that you're having, how can those leaders distinguish between sort of the short term shiny object, right? Oh, this is going to help me right away. And technologies that are really delivering lasting value. What's the key to that?
- [07:35]
B
Pilots. For me, the thing that health systems should do and that we've seen the smartest health systems do. You look at an HCA or tenant who are deep partners of us, and also some of the most profitable health systems in the country. What it comes down to is run a lot of pilots with a lot of different vendors, run a competitive process, and then make a call. I think too often I found that health system leaders select their vendors a little bit like selecting the name of their baby, which is, you get one shot at it and then you're done. And, you know, this is a precious, holy process. And I think it's kind of ridiculous because I look at, you know, a company like Kamir, which I would consider is quite technology forward and deploys a lot of software within our own business. We're piloting 100 different vendors at any given point of time to figure out what is the best possible solution, what actually delivers roi, who's real, who's not, what solutions are integrated, what solutions are early on, what's a real platform, what makes data interrupt easier, not harder. And too many times without systems, I see, you know, someone look at a presentation, look at a couple faces on a slide, and then make a pick, and then they're just stuck with it for. For the next three years. And, and that's just. It's not a good way to do business, and it's not a good way to do right by your shareholders or your customers.
- [08:58]
A
Yeah, trial and error is the key there. What are some of the principles that should guide leaders when they're looking at selecting partners? It's certainly not looking at the slide and then picking. Right. What are some of those principles that, that you recommend, really, leaders dive into?
- [09:13]
B
The distinguishing factor more often than not is engineering talent. How strong of an engineering team and a technical team does. Does your partner have? How many engineers can they deploy on site? And we call this concept forward deployed engineering. We obviously didn't pioneer it. It's something that Palantir brought to the U.S. department of Defense and after that, the CIA. And for us, we're taking that same model and bringing it to health systems, which is you take an engineer, you take a group of engineers, and you put them on site, working hand in hand with the operators, with the physicians, with the clinical staff. And that's the only way to get work done, because every health system is different. There's no such thing as cookie cutter enterprise software in health care. And you have to modify and customize on the ground to the workflows of the health system. If your partner isn't sending engineers on site to do hand to hand combat and work on the front lines with your staff. That's a bad partner in today's day and age and I think it's a good signal when you have engineering talent that is customizing for your needs every minute of the day.
- [10:19]
A
It comes back to what we talked about earlier too, which is the integration piece that enables integration, full integration for those health systems which is really, really key to drive measurable results. Danay, thanks so much for being here and for all of your insights. What a fantastic conversation. I want to turn the floor over to you. Anything else that you'd like to touch on that we might have missed or anything that might be important for our audience?
- [10:40]
B
The last thing I'll mention is I think the best software companies are built in many ways by the customers that ultimately end up using them. And we're always looking for more clinical partners to guide our co development, whether that's health systems advisors, solo doctors that work with our engineers. We're not going to claim to sit here and know every workflow in healthcare. What we will claim is that we have a team of engineers that learn fast. And I think that our products only become better when you have clinicians and engineers sitting in the same room.
- [11:10]
A
Like you said, more pilots for everybody. Tanay, thanks so much for your time and insights today. It's great to have you.
- [11:15]
B
Awesome. Thanks for having me.
- [11:16]
A
And we also want to thank our podcast sponsor, Kemura. You can tune into more podcasts from Becker's Healthcare by visiting our podcast page@beckershospitalreview.com.