Transcending Home Care: The Ubiquitous AI Episode

[00:00:00] JJ Ram: I used to have to give it a ton of details six months ago. I used to be like, go to this file, write this thing, do this exact thing, and now I can just kind of explain to it this is what I need to get done. And then it asks me the right questions and can just get it done within minutes rather than days. I think the cool part, to me at least, is it’s filling in the gaps more and more. And that’s where there’s just less you need to explain when you’re interacting with the AI.

[00:00:35] Tony: Welcome to Transcending Home Care. I’m Tony Kudner, the Chief Strategy Officer here at Transcend Strategy Group. We might call this episode the ubiquitous AI episode. I think if you’ve been following me or following Transcend for any length of time, you know that we go to Home Care 100 every year and that’s where a lot of, some of the bigger players in the home-based care space tend to be as well as some of the sort of more cutting-edge technology firms and this year was absolutely no exception. Some really interesting developments and I had the fortune of sitting down with a gentleman while I was there at dinner one night named JJ Ram.

[00:01:22] Tony: He’s the CTO and co-founder at Claim Health. I'll let him talk about what Claim Health is, but he said some things that resonated with me and maybe appropriately blew my mind is the best way to talk about it – as somebody who is not living and working in Silicon Valley or I’ve never been a coder or dealt with all of the infrastructure that goes into electronic medical records or associated tech.

[00:01:51] Tony: He said some things that just made me think, we are just in for a heck of a ride. So, I reached out to him after the conference and I said, JJ, why don’t you come on the podcast and talk a little bit about the state of AI from where you sit, which is much closer to the beating heart.

[00:02:06] Tony: So, with that, JJ Ram, welcome to Transcending Home Care.

[00:02:10] JJ Ram: Thanks for having me. I feel also very fortunate. I think we sat next to each other at two different sessions and it’s kind of a crapshoot. So lucky to sit next to you.

[00:02:20] Tony: Yeah, absolutely. So, can you maybe just start by talking about your background and how you came to be in the home-based care tech space?

[00:02:29] JJ Ram: For sure. I went to Stanford for computer science. I always knew I wanted to start a company, but instead of that, I joined Facebook and now called Meta for a year there. Really cool, like some of the smartest people I ever worked with, but it was just this huge company. You learn a lot of fundamentals, but you’re not really close to customers as much.

[00:02:49] JJ Ram: You’re not building fast, you’re not getting to build full-scale things. So, I left that company after a year or so, and then since then I've been at a bunch of small, really fast-growing AI startups. All across healthcare. I worked at a construction tech company. I worked at a private equity AI search company right when Chat GPT came out, which is really cool. And then my co-founder and I worked at this dental tech company back in 2020. We were really early employees. That company went nuclear. I think I joined like when it was 12 people. It grew to let’s say 1,200 people in the two years I was there. It was crazy growth.

[00:03:26] JJ Ram: But my co-founder and I were good friends from the early days and stayed in touch and he helped grow that sales team out. And I was an early engineer and then we went our separate ways. And then he joined a home-based care start-up, a tech start-up building for nurse incentives. Like caregiver rewards and trying to incentivize nurses to do value-additive things in the workplace, like EVV compliance, submitting notes, etc. So, he was leading sales over there, and he basically poached me to join that company. So that’s how I got my foray into home-based care.

[00:04:02] JJ Ram: It was really cool. It was really interesting that we were doing nurse retention, but through that we’ve learned that all these agencies were incentivizing nurses to speed up revenue cycle and get that super tight. And then, with my AI background and Kevin’s sales background, we were like, there’s just so much opportunity to help agencies out more directly with revenue cycles. So that’s how we got started with Claim Health.

[00:04:26] Tony: Awesome. Can you talk a little bit about Claim Health specifically and what the product is that you offer?

[00:04:32] JJ Ram: For sure. We started Claim Health about a year ago, and our hypothesis was we can build sort of like a billing company but AI native so we can help submit claims, monitor denials and help with authorizations. And we were doing that for a bit. It was going well, but every single denial that we worked, every single claim that we worked was due to issues that stem from intake, authorization and eligibility verifications and making sure the patient didn’t change insurances like two months into their stay with an agency. So, a lot of our customers led us to start building more upstream.

[00:05:14] JJ Ram: So now we build a bunch of AI automations all the way from intake, whether it be from a fax, an email, a marketer, texting something in a referral portal. We run eligibility automatically and we basically validate if that patient’s a good fit for the type of insurances an agency takes, where they live, some of their co-morbidities, some of their clinicals. Is this something you have clinicians that you can service and then trying to make a decision within a couple minutes and take that patient in. And as a patient is getting care from the agency, making sure that everything is in line to get that claim paid. So, a bunch of automations all across the board over there too.

[00:05:53] Tony: Got it. Helping with Oasis and stuff like that as well?

[00:05:57] JJ Ram: So, we do less on the Oasis and notes side. We’ll do everything up until that. So chasing down orders, getting orders signed, making sure all the documents are in order, then we hand that off to the clinician to do.

[00:06:11] Tony: Okay. Awesome. Thank you. 

[00:06:12] JJ Ram: Yeah.

[00:06:14] Tony: So, in the past few months, it feels AI disruption has accelerated, or at least the news of it. I think everybody is trying to separate fact from fiction, hype from reality. But there was, I posted this on my LinkedIn a couple weeks ago, there was a widely circulated blog from a tech worker named Matt Schumer that went viral for just voicing what felt to me at the time like a pretty clear-headed view that just says we’re not ready for the level of disruption that’s coming. And everyone’s been talking about Claude’s new models and that the older models are now helping write the Frontier models for Claude and I don’t think that’s unique to Anthropic, but across the board and the feeling even since we were at Home Care 100 is that this stuff is just accelerating right now.

[00:07:07] Tony: Can we get a gut check from you? First of all, you live in and around Silicon Valley, right? You live out on the West Coast. What’s the gut check? Is this hyper reality?

[00:07:17] JJ Ram: I actually used to live in the area. Now I’m in New York, and it still feels like that. So, the gut check is I think accelerating is the right word to use. To give you a sense of the way I view AI, I use AI for my work almost every minute. So, I code using AI. I have all these automations to help me respond to customers and surface their requests and make sure that nothing is falling through the cracks. And a lot of coding I do now is just like asking my coding agent to code for me. And the really cool thing, this is like more of a coding lens, but accelerating is the right word. The interesting thing to me is that especially, like in America, the amount people know about it is just so different, kind of all across the board. It’s just moving at such a fast pace that even people who are looking into it every day can’t keep track with how fast it’s moving.

[00:08:23] Tony: It’s interesting you talked about how you were using it six months ago. I hope I articulate this the right way, but six months ago, you think about how McDonald's has optimized all of the various tasks that are required to get you a hamburger. And if you are a line cook at McDonald’s, you literally … it’s take a burger, put it on for 30 seconds, put it in this pass-through drawer. That sounds like what you were talking about with AI six months ago. But what you’re talking about now and how you structure a conversation with it, and it can intuit the right parameters, have a conversation with you, and then go do what you need. That’s white-collar work, right? Like that.

[00:09:09] JJ Ram: Yeah.

[00:09:09] Tony: Before you would be talking to a junior engineer and you’d say, go play around with it. We have time. We’re not in the middle or at the end of a sprint. Try to build me a version of this module. And it sounds like Claude can do that now, if I’m hearing you right.

[00:09:27] JJ Ram: Yeah, I think so. I think the limitations are Claude doesn’t really understand your customers. They’re not in the room interfacing with your customers. So, being able to digest what needs to be built is, I think, the current gap. I don’t want to say it’s not going to be solved by Claude eventually, I think it probably will be. But being able to translate that is what the skill has become. I think the really cool part is, and I try to tell this to everyone at our company and my friends, and I’ll tell it to you. I think it’s kind of like electricity. It’s like this building block where you can get super, super creative with it.

[00:10:09] JJ Ram: Like when electricity started to be harnessed. I don’t think people expected you to have an outlet or you can just plug in a toaster. It started with lights and you can really apply it to everything. So, I’m coding with my AI, and then I can tell that same AI: Hey, I want you to create a document that I can send to my engineer to continue on with this piece of code or something like that. And you can really apply it to anything. I think that’s the really cool part. And that’s also how to continue on and how jobs are going to transform. It’s like you have to be super, super creative with it.

[00:10:54] Tony: Yeah, that makes a lot of sense. And I think that’s where, speaking just for myself and probably some of my co-workers, I don’t know that we think it’s going to supplant us, but we’re darn sure going to be keeping as close track of it for the level of technical sophistication that we have as we can. I’m using it, probably not as much as you, but every single day I’m using AI for something, for some part of my job at this point. But I’m not sure that’s true of the larger home-based care community. And that’s the next thing I’d like to pick your brain about. You’re straddling this world using AI for 90% of your job or every single day to manage your workflows and at the same time, home-based care is probably some of the later adopters of this sort of tech. It’s very human capital intensive. It is highly variable in a lot of ways. It’s a person-to-person business. It’s not something … we’re not making widget. And because it’s in the home and different environments, it has historically been very slow to change.

[00:12:04] Tony: How do you reconcile the world you live in when you’re at home versus when you go out to a client site and you’re trying to spin them up an instance of Claim Health? What are the discrepancies you see there and how do you all deal with that?

[00:12:19] JJ Ram: Yeah, that’s a really good question. It’s really hard because a lot of the technology that we’ve seen when we’re on-site is from the early 2000 or late 1990s. And it’s difficult to go from using that software to using really, really cutting-edge new software. One of our jobs is bridging that gap for sure. The way we do it is we try to map out with our clients and prior authspecialists that are on-site, intake coordinators are on-site, people doing the actual job to map out what workflows they want. I think automated is the wrong word. I think it’s more like multiplied. Is there a specific action that you have to do every single day that if you didn’t have to do, you could call more patients or you could interact with more families and things like that. So, we spend a lot of time with every single client just mapping things out.

[00:13:17] JJ Ram: And going back to the electricity analogy, you have to get creative with it. It’s also making these onboardings like brainstorming sessions and make it a creative experience where, you know, I’m not working as an intake coordinator every single day. Now it feels like I am because I’m helping run intake for some of our customers. But, these are experts in the field and they have awesome ideas and it’s kind of bridging the gap between what they do day-to-day and what’s possible, and it’s kind of explaining that to them. So, demos work, showing them things that other customers potentially have used works, but it’s more so just like getting their ideas. And then now that we can code so fast, we can just build it, right? So, I don’t know if that answers the question, but it’s like trying to do a lot of brainstorming with our customers.

[00:14:09] Tony: I think it does, and I want to just clarify one thing because the paradigms are shifting so fast here, but it almost sounds like the historical EMR or EVV or rev cycle management model is … we have a product and we’re going to come in and we’ve got a team of engineers and if you need a couple dashboards built or a particular workflow or a change order, there is a very structured way that happens and it’s measured in months to years, depending on the level of customization you want, those sorts of things. But I’m hearing you say that, like on a client-by-client instance, you are basically workshopping Hey, what’s your pain point? And then going back and figuring out the solution to that pain point leveraging AI. Do I have that right?

[00:15:00] JJ Ram: 100%. And the way we think about it is we’ve built building blocks like document processing, web automation, desktop automation, all that sort of stuff. And the only way to, and who knows how scalable this is going to be, I think it can be very scalable with how fast AI can write code now, but you have to bridge the gap. You can’t really build a 90% or even 80% AI product, it has to do that last mile. That’s really the hard part and that’s where you drive the most value. My co-founder, Kevin, was on-site with the customer for two days this week literally in boardrooms, just mapping out every one of their workflows to see where their opportunities were where we can jump in. And now we have this awesome plan. Where we’re going to help this hospice provider save a ton of time and, more importantly, make sure intake and auth don’t slip through the cracks because we could see exactly how they model their workflow, how they think about their workflows. And they don’t really have to change that much. We don’t want them to have to change. That’s also a benefit of AI. It doesn’t have to be this exact cookie-cutter thing. You could really bend corners to fit the workflows.

[00:16:12] Tony: That’s going to keep me up at night. That’s crazy. It’s almost the new classes of biologic drugs that are attuned to a patient’s DNA. It’s almost like you’re doing the AI tech stack equivalent of that for home-based care providers.

[00:16:29] JJ Ram: Yeah, it feels like it. And I think it helps us for eventually when we have thousands of customers, right? Hopefully. But by being so close to our early dozens of customers, we’ve learned just such an immense amount. Like when we’re in these boardrooms talking to execs and then we sit next to the desks on the same floor with intake coordinators and non-specialists. We have very, very quickly become like pseudo experts in the industry. And I think that’s also the best way to build our business, too.

[00:17:01] Tony: Yeah, absolutely. It’s just like hospice, it’s got to be individualized to the patient in the instance, the plan of care. So why would the tech stack be any different? So JJ, just to wrap up, this has been fascinating and I really want to thank you for your time, but where do you think the major, this is a loaded question, but what is the next disruption from where you are sitting and as close as you are to the pace of innovation? What’s the next big thing that’s really going to be that bomb that drops in terms of innovation and change?

[00:17:34] JJ Ram: I think this is already happening, it’s happening in coding right now. It’s the AI instance that you don’t have to prompt, it just is running and it can just spin off new AI to go for it. So, it’s not really the term AGI that people talk about where they’re like, oh, this is general intelligence, but it’s like an LLM that can connect to a bunch of sources, read a bunch of stuff, understand, and then go and schedule a call to a patient or submit an authorization or an intake comes in and it can just automatically go and write it into an EMR and then read from that EMR. And it’s just this one intelligence that can do that, like that exists in coding right now. And I haven’t really seen it at crazy scale outside of that. I think that’s the next thing that’s coming is just AI that can spawn more AI.

[00:18:36] Tony: Yes. It’s going to be AI all the way down. We’ll see if that happens. I think it’d be really interesting if there was just a super mecca intake AI that just needed a little bit of the human touch and compassion to be its coach but interested to see if that happens.

[00:18:53] Tony: So, JJ, thank you so much for your time. Thanks for joining Transcend’s podcast. And best of luck with Claim Health. Thanks for sharing your insights for all of us luddites out here in the home-based care space. Really appreciate the time. 

[00:19:07] JJ Ram: Thanks, Tony. Appreciate it.