← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Venkat Mocherla, Founder of Midstream, and Dr. Feby Abraham, EVP and Chief Strategy and Innovations Officer at Memorial Hermann
00:26:46
Loading summary
Transcript24 lines
- [00:00]
A
This is Scott Becker with the Becker Healthcare Podcast. Thrilled today to be joined by two brilliant leaders, two of the most innovative, smartest people I get to visit with. We're visiting today with Febi Abraham. Febbie's the Executive Vice president, Chief Strategy Officer and Innovations Officer, Memorial Herman Health System. He's also a former McKinsey partner, a really, really gifted person. We also with us today, Venkat Mokurla. Venkat is the founder of Midstream. He's also a former operating partner at Dreeson Corwicz. Also one of the brightest, most entrepreneurial, interesting people I get a chance to visit with. We're going to talk today about automation in healthcare and what they see on the horizon and a lot more. Before we get started, Beby and Venkat could ask each of you to take a moment to introduce yourself. Fabian, let me start with you and then I'll turn it to Venkin.
- [00:57]
B
Thank you, Scott, and good to be here. As you mentioned, I'm the EVP and Chief Strategy and Innovations Officer at McMoel Herman, a leading health system based in Houston. We serve one of the most diverse populations in the country and we've built a strategy rooted in the health of our communities. We've been here for over 100 years and we do that through disciplined innovation, a commitment towards taking care of the health of humans and not just sick care, and delivering that through partnerships and alliances that improve how care is accessed, experienced and delivered. We are also very deeply committed in aligning innovation with our operations. We do that through innovation piloting concepts early, especially as we talk about AI and some of the technologies we're going to visit with you today about and then scaling them wisely so that we are really rooted in our mission, but also delivering on all the reduction of the waste in the ecosystem and doing what's right for the communities we serve.
- [01:50]
A
And Febi, I have a couple questions for before I turn it to Venkat for an introduction. You are did a PhD at Rice, you did your undergraduate at the Indian Institute of Technology. And again, for people that don't know, some of the brightest people I've ever worked with have graduated from the Indian Institute of Technology. So I have to ask you the question. You don't have to answer, but I'd love you to what is a better school? The Massachusetts Institute Technology, MIT or the Indian Institute of Technology Bombay, which is the better school?
- [02:24]
B
Well, you put me on a bind, but I'm going to take a stand. It's the Indian Institute of Technology. No doubt.
- [02:30]
A
There you go. There you go. I expect that answer and I was hoping for the answer. But what I cannot even tell you the amount of brilliant people and entrepreneurs and engineers that I have met from there over the last decade to two decades. So just fantastic in both magnificent schools. But I love the answer. Thank you. Venkat, take a second and introduce yourself and tell us about what you do and what you're doing today.
- [02:53]
C
Yeah, yeah. And first of all, I should say that my dad went to also went to IIT and he reminds me that every day. So thanks Scott for bringing that up. But I'm Vicki Butcherla. Thank you Scott for having me on the podcast along with Febi. I'm the co founder of a company called Midstream Health. We are focused on applying AI for automating and augmenting financial workflows and financial decisions for some of the largest health systems in the world. I just believe this is the most interesting, important time to build technology to go. Both serve patients, clinicians, administrators alike. And I am really glad we're having this conversation today because so often I think in health care especially, things get painted as zero sum and I think we are fundamentally in a world of abundance. And I think Febi and I, along with you share this sort of very optimistic, exciting viewpoint and I'm excited to jump into the conversation.
- [03:57]
A
It's interesting you say that because I am as optimistic I've been in a long time. I do see some big structural changes happening in and hopefully they'll be for better, for worse. I see some challenges, but I am optimistic and I do agree with your overall perspective. Febby, talk to us a little bit about. You're one of the greatest health systems in the world, chief of strategy and innovation. Where are you seeing the automation curve going? Where are you seeing impact and where are you unsure of where there's going to be impact?
- [04:28]
B
Yeah, I mean there is a huge potential for automation in the industry. If you think about 5 trillion and growing in terms of the overall expenditures, we're obviously lagging many other OECD nations in terms of outcomes and have a come down in terms of both quality adjusted life years as well as the overall lifespan. So we have a tremendous opportunity both in outcomes and total cost of care. And I think automation can be a huge accelerant in how we do that. The way I would sort of attack that is to sort of really look for components of waste that needs to be taken out and you can sort of see that across all the workflows that you See in how healthcare is accessed, experienced and delivered. So you could think, let's start with the administrative, right? You have a tremendous opportunity in sort of taking out all the friction when it comes to things like how a visit is scheduled, the payments are made, the post visit follow ups are done. The whole administrative side is a tremendous opportunity on that. The second area is, if you actually look within the operational side, repeatable tasks. There are plenty of them and we are living in an era of a supply short labor, supply short industry. You can see the headlines around shortage of nurses and doctors and the several hundred thousand across the world and particularly in the us and we also talk about how automation can take away human jobs. I say that's the wrong question because you are dealing with a time where there is a graying of the population, there is a shortage of staff, many of those staff are not rejoining the industry, there's not as many people graduating into schools as are retiring. And we need to create a mechanism whereby 50, 60% of the day is spent now, today in sort of low value added work that are quite repeatable, where technology can take off, where would it happen? What would be the art of the possible if you can sort of free up that and capacity for the workforce and therefore see a larger slew of patients? So I think there is a huge opportunity in the operational and clinical workflows. And when I say the distinction I draw, that is clinical workflows are where a doctor or a nurse is seeing a patient and in the case of an operational workflow is how do you sort of run the hospital operations in a way so that you can sort of take care of the patients more immediately rather than having them to wait for a long period of time. And finally, there is a, a tremendous opportunity when it comes to patient education, patient engagement and getting immediate care at the right side, right side of care at the right time for the right patient. Patients are already engaging with the likes of ChatGPT and sort of how healthcare is accessed and experienced. But we need to sort of now make that more integrated into the workflows so that you know, they are being directed into the right appropriate care, is being funneled through great clinical brands. And the same experience that they get with their financial experience with the bank or with the retail setup, such as an Amazon, is what we can deliver in the healthcare setup. So those are, in all of those areas we can see significant transformation happen. But as we made the point in one of the articles we wrote, Venkat and I authored at Becker's, this would be a multi year journey and we really need to sort of focus on how we set up the roadmap in terms of the levels of automation. We should sort of go around and look for opportunities where there is more augmentation and low hanging fruit and picking some of the waste in areas where there is joint conviction across the stakeholders and then slowly cherry pick our way into some of those areas that are even more complex with regulatory barriers. And the patient safety and other kind of safety and risk factors need to be weighed in.
- [08:12]
A
Fabio, let me ask you a micro question and a macro question. The micro question is where specifically are you seeing the rate of change have the best positive impact right now? Is there a specific area you could point to and say we're really seeing progress here or there? So that's the micro question, the macro question. When we talk about all this complexity, sometimes when I hear people talk and I see so much complexity between the amount of payers in the country, how healthcare finance and everything's organized and I see the contrast of things moving back to things like dpc, direct primary care, I think that all these tools and this is the macro question are being used within a system that we have so many flaws in organizationally that we at some point need to come back and figure that out. But assuming that's not going to change, there's about payers organized versus health systems versus providers, then we better do all these things automating and operationally, well, we try and figure out if we could ever solve the bigger problem of supply and demand in the organizational structure of our system. But on a micro level, On a micro level, give us just one or two places where you're seeing the biggest impact currently.
- [09:25]
B
Yeah, I think the main areas are where you're seeing the workforce get inundated with lots of repeatable tasks which are automatable. So some specific examples of that include ambient listening where the doctor is visiting with a patient, and the traditional workflows involved taking a lot of the notes where they're not even able to keep an eye contact with the patient. And the joke was 9 to 5 they visit the patients and 5 to 11 or 12 in the evenings they're spending time sort of taking all those notes. And by the way, those notes need to be appropriately coded and input for them to get paid. So this was leading to a significant dissatisfaction among the doctors. The same thing you could sort of put in documentation where there are multiple sites of care involved. They've got to go through the longitudinal mapping of the patient journey and look for what intervention to put relative to the disease that the patient or the symptoms that the disease the patient is presenting. All of those things particularly are things that you'd say are automatable and we are sort of really rapidly accelerating that. The same thing can be said about when a patient is sitting in a bed. Can we sort of keep an eye towards predictive technology that will measure their likelihood of falling or you know, where we could sort of provide the sitters in a virtual basis and automate even that journey where some of these things are done through activation of signal events as opposed to somebody sort of monitoring and therefore improving the capacity of the workforce to do more immediate tasks. Another area that we are seeing a lot of application is on the revenue cycle. So you know, as you know from the point you made about the macro about payers, the payer provider interface has created significant dysfunction in the, in the US health care ecosystem. You need a patient to be pre authorized ahead of many visits. There are denials of the claims. Multiple different health care entities, including specialty care and the facilities touch the patient and they all need to go be circled back up across the visit. The risk scores of the patient needs to be appropriately coded. All of those things in the revenue cycle are fraught with interface challenges where the payers and the providers have different fact bases. And there's another area where you could sort of see a lot more automation come in. The same applications we could sort of see within the supply chain categories as well. So there, there's a lot of those near term, call it 6 to 18 month priority areas that we have mapped out that, that map into those categories, such as the ones that I described, that present themselves into the value creation of bringing more joy to the clinician by taking out some of the ways making the productivity and the yield of the revenue cycle in the back office representatives really go much better and the same thing on the supply chain. We're also starting to experiment a lot with automated technology in, in some low complexity surgical types where you have robotics and vision based tools to guide and sort of make sure that there are no errors and that the consistency and quality of the surgery can be much more accurate.
- [12:46]
A
Thank you very very much. Just fantastic. Inventkit, what are you watching currently? What's top of mind for you and how are you seeing automation make a difference?
- [12:56]
C
Yeah, I think, you know, I mean Febby brought up a bevy of use cases that I think are spot on. I think, you know, to me, I think so often when we think about automation we go into the clinical workflow. I think the ambient use case comes up so often, but I don't think we talk enough about both the administrator and the patient. I mean, I'll tell you, you know, one of the most fun moments I've had is when I've had my parents in a waymo in San Francisco. You know, in fact, we just took one like last night, and it's that moment of like infusing joy again. I think what we've had is over, you know, from the advent of the modern hospital, which, you know, if you go back to the orientation around like Hopkins, we've had these like specialists, not just in clinical care, but even in administrative. You've, you have specialists who've just got into their siloed world. And I think we don't talk enough about burnout, not just in clinicians, but I think across the space, patients are frustrated, administrators are frustrated. So when I think about these repeatable tasks that everybody does in their pajama time, not just writing notes like my fiance does in epic, but I'm talking about every single person who's leading the managed care department, who's in supply chain, who's going painfully PDF by PDF, contract by contract, or a patient who's had to fill out their forms 45 times is repeating the same question 85 times to tell their frustrating longitudinal journey they've had with the health system. I think for the first time, it's just humanly impossible to alt tab 85 different tabs of your life, to stitch together what the heck is happening, whether it's a form inside a health system or a record for a patient. I think my fundamental excitement is this kind of intelligent assistant that is just there for everybody. Maybe sometimes it takes an analytical form for someone who's an administrator or a leader or an executive. Sometimes it takes an assistant form for a patient. But this kind of intelligent person that's just there watching your back, advocating for you and analyzing for you in sort of stitching the, connecting the dots across the board, that capability. Now we're seeing in not just the use cases that Febio mentioned, but increasingly we're seeing it across the board. And for me, the fun part is how do you stitch all these together? I mean, if you ask, you know, I'm sure the hardest job right now is to be an executive like Febi, who's probably getting 100 pitches a day on every AI application under the sun. Maybe, in fact, I'm, I'm underestimating the hundred emails. Maybe Febby. It's a thousand. But, but I think the, the frustration that a lot of people have, like executives, like Febi is all of this is not connected. And I think for me the excitement is like actually this super intelligence that actually gets to connect all the dots. On behalf of the patient, the administrator and the clinician.
- [16:07]
A
Thank you. Let me ask you a follow up question. In terms of people being able to easily use different AI tools, what does it take to get to the spot where in different roles, different places, people can easily use AI tools? Where it becomes such second nature to them, not where there's such a learning curve and they have to figure it out. How do you see that and what do you see?
- [16:31]
C
I mean, I think that's probably one of the most fun parts about interacting with Software today in 2025. I think the frustrating part, a lot of whether it's clinicians or administrators have is when they're within the four walls of the health system or a payer, they seem like they're interacting with tools and technology that was built like before they were born. Yet you come outside and you interact with ChatGPT, you hit a car and a car, ours is driving by itself in Waymo or whatever and it feels like, you know, one, you know, @ work it's Flintstonian, you know, at, you know, in your outside of life at Star Trek or Star Wars. And I think, you know, part of the excitement, Scott now is, you know, whether it's nlp, whether it's these sort of multimodal interactions, you know, voice is, is kind of, kind of incredible. The way that these voice models are developing is that there is now a much more human, intuitive, simple way to ask something and in fact predictively, prescriptively get something even before you ask it. And so I think, Scott, a lot of these questions you ask about ease of use and sort of last mile challenges, change management, et cetera, I think the technology breakthroughs itself make it much more easier. I don't know Febi, if you have a different, different counterintuitive point there.
- [17:54]
B
No, I think that's right. I mean the complexity of the fidelity of some of the generative AI based solutions have led to a level of adoption which were not seen in the purely AI world. Right. Part of it was it becomes, it has become very much more intuitive for somebody that's engaging with your favorite GPT out there to sort of start to say, well this actually worked. I asked a question and I got a pretty good answer. So the conviction and trust that is developing with the specificity and the quality of the. Just the output that is coming up is leading to a level of trust.
- [18:33]
A
Febby, take me from that. Because that's exactly what all of us are seeing is in our daily life. Wherever we're getting our AI answers on stop. We plug in an answer and now instead of getting 30 articles to look at, it's already condensed this for us. How do we convert that from there to operations at scale in health systems? We could all start to see the positive impact. Now what does it take to implement those kinds of things at scale in health systems? From sort of the individual person doing it on their own. But we're, you know, to how do you make that actually operationally scale in health systems? That's, that's I guess my question.
- [19:14]
B
Yeah, so I think there is three parts to it, right. Part one is really picking the use cases in a way that is going to move the value creation for the system, both clinically, financially, operationally and bringing sort of satisfaction across our workforce. And to do that there is a framework we use which is like titled drivers around value creation itself, but also looking for the maturity of the technology and seeing sort of like that Goldilocks or that sweet spot of areas that we should go after. That's in my view, the where to play question. The second part of it is then proving that it works. There is no better way to do change management than the actual users community being evangelists to each other. So one of the things that we've pioneered is to have something where we have an innovation hub like capability where we trial the problem statement that we're trying to solve with a specific AI solution in a contained setup. So we'll pick one hospital or a handful of clinics, or as Venkat was saying, in some of the administrative use cases, we'll pick a function to sort of trial it out. We will sort of run that pilot for a trial period and show that the specific value creation drivers across a few key performance indicators are bearing itself to these. The promise of the AI is actually working out in the real world. We understand some of the barriers to adoption because hey, technology has already always existed. Even before the generative world, it was a barrier or adoption. So we will really dig under the hood and sort of unpack that a little bit and sort of really look for the pilot results to sort of frame that. And then that leads us to the third point which is then sort of going into scale, which is now I've got a pilot that worked. The demonstrable results not only showed financial Value creation. But, you know, in the case of, for example, ambient thing and a lot of the other things, improve pajama time, improve net promoter scores of both the clinicians and the patients approved a holistic value proposition. And now we can sort of scale it across the enterprise, across all our hospitals, across the entirety of thousands of physicians. But as you do that, you've got a set of evangelists that are actual users, like doctors, who have trialed these solutions and are able to educate the rest of their community and their colleagues on how best to engage with the technology. Finally, I would just sort of make the point that now if I periscope up from all of these, across these points, it is to me an innovation flywheel also at the end of the day, which is that you start to see some of these technologies integrated and adopted and show the magic of its specific value creation and then it creates a halo that drives pull through for other use cases over time. You're starting to see that quite significantly in our innovation hub, where some of the early applications like ambient listening are now scaling across the platform and then more and more demand is creeping up for other technologies because we've shown that things work. And as they say, success is many mothers, right? So there is obviously that sort of change management driver and the halo effect that drives the flywheel.
- [22:30]
C
I just, just to add to that, I think, you know, and this is like, man, I wish I heard this podcast when 10 years ago I was like starting a company, right? Which is basically the. The secret behind any new technology, Scott, which Febby is very much hitting on the bullseye here, is you only get one chance to establish trust with the end user, with the end executive, with the end clinician who are the person on the other side is using it. What we found with leaders like Febi, or we work with systems like Common Spirit, is that there is a sort of a magical, very intentional methodology in how you make sure the data is accurate, that whenever somebody uses some piece of technology, it stitches together to their workflow. And at the end of the day it just works because you only get one chance. And once you know, the danger of all this is that there's so much fatigue to new solutions that if it just doesn't work that first time that end user is out, they've quietly quit on your new application. And now you've burned the trust for every new application from there on out that you want to establish. And so I think, you know, especially for founders and builders, you're listening to this, listening to folks like Febi on like how do you create that trust is incredibly important for your success to land and expand these applications.
- [23:58]
A
No, thank you so much. I think that's so right because if the first impression is bad then people won't work with it to try and get it better and better and scale with it and work with it. I think you're exactly right if it doesn't hit right to begin with. One last question. We've got 30 seconds each. Febi final thoughts that you would give to a health system leader and sort of trying to automate integrating AI. You only got 30, 45 seconds so your final thoughts on that. And Venkat, your final thoughts along the same lines, advice that you would give to leaders in health systems.
- [24:31]
B
Yeah, for me the thing is just stock bust with the problem you're solving. AI is a solution in the toolkit that solves the problem rather than be a hammer in search of a nail. When you're talking to a company, be articulate about where the pain points are and the unmet needs are that you're trying to tackle. That's point number one. Point number two I would say is that think bold and big. Don't try to sort of optimize the broken workflow with AI. Reimagine your entire workflows now that you have a benefit of technology that can sort of fundamentally redo the entire workflow and create a lot more efficiency and stickiness. And then the third is think about this as a long term roadmap. You're in for a long marathon of multiple changes that will collectively add to a big transformation. Don't expect silver bullets out of the first opportunity that's going to solve every one of the world's problems just with it.
- [25:23]
C
Yeah, that's incredibly well said. I think my my things that I would add to that real quick are one is in the same way we've outlined this in the article in how self driving cars have a six stage framework on going from Lanesys to Waymo plus it's think about this in a sequential journey as you lay out these use cases that Fabiot mentioned. I think about it like Pac man and how do you start to eat one bite at a time and then you slowly get to see the dots connect in the user's eye. Second is it's got to be worth ROI and roe. It's not just the investment of the money but it's the time and effort it takes to do something. And the third thing I'll just say is it's just all about the last mile. The last mile, the last mile. I can't say that enough. Think about the workflow, think about the change management needs and you know, if it's all done right to Febi's point, you have a, you have an opportunity to really create that magical moment for the first time for a very burnt out end user.
- [26:30]
A
Febby and Venkat, I want to thank you so much for joining us today on the Beckers Healthcare podcast. Both brilliant leaders. Thank you so much for joining us. Look forward to talking to you further. Thank you very much.
- [26:41]
C
Thank you Scott.
- [26:41]
B
Thanks for having us.