← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Balancing AI Innovation and Governance in Revenue Cycle Management
00:20:17
Loading summary
Transcript19 lines
- [00:00]
A
Hi everyone, this is Brian Zimmerman with Beckers Healthcare. Thank you so much for tuning in to the Beckers Healthcare podcast. Today we're going to talk about the path to responsible AI adoption and revenue cycle management and how health systems can balance innovation, governance and change management to drive real impact. Joining me for today's discussion is Kanar Kokoi, founder and CEO of Chirac Health. Kunar, thank you so much for being here today.
- [00:24]
B
Thank you for having me and to.
- [00:26]
A
Get us going here and help listeners appreciate your perspective. Can you, can you share a bit about yourself and your work healthcare?
- [00:33]
B
Sure. Thank you. So I'm Kokoi obviously, and founder and CEO of CH Health. Really my work in healthcare brings together the operational and the technical sides to create real lasting solution. That's what I've seen myself doing in the past 25, 30 years in this industry. I've seen myself as an NFT innovator and a solution maker, someone who helps organizations strengthen their revenue while improving the way care is delivered. I have a strong background on the technology side as well, where I've designed and applied tools that improves accuracy, efficiency and financial performance. At the same time, having the understanding of the operational challenges the that many healthcare organizations face every day has been a crucial aspect in terms of my experience and understanding that with many of the healthcare organizations that I work with today and that has allowed me to bridge the gap between strategy and execution. One of the most important thing that matters to me is really impact. I truly believe provider burnout is one of the biggest issue in the healthcare industry and my focus in the past several years has been on building solutions that eases that administrative and operational burden so providers can really focus on patient care. And lastly, really, innovation for me is not just about technology. It is about creating a system that is financially strong, operationally sound and supportive for the people who deliver the care.
- [02:21]
A
Scott, some great points there. And that being financially sound too is a part of what makes the rest of that mission you laid out possible, right Kinar, correct?
- [02:31]
B
Yeah, it's very important to have a lot of those intuitives together and we.
- [02:38]
A
Know there's some numbers out there already and of course we talk about AI in healthcare all the time now. But I believe the stat I have in front of me is over 70% of health systems have already deployed AI pilots or full solutions in areas like finance, RCM or clinical care. From your perspective then, what is the right pace of AI adoption in rcm? Specifically, what risks are health systems facing as they try to do this work and what Are the risks, I guess, of going too fast or too slow? How do you find that sort of happy medium there where you're going at the right pace?
- [03:11]
B
Yeah, that's a really great question. So, you know, when we talk about the pace of AI adoption, especially in revenue cycle, I think the key is balance too fast and we expose ourselves to risk around accuracy, governance and unintended consequences. And if you go too slow, I believe we risk being left behind. Especially as I'm sure you guys are seeing, many of the private payers are already leveraging AI aggressively and often to regress reimbursement to the me. So in my opinion, the right pace isn't really about speed for speed's sake. It's more about how thoughtful the adoption has been with guardrails and clear strategy. So I just want to break down a few key points. What I mean by guardrails and strategy, the very most important thing is regulatory and audit environment for any adoption. Because right now we are in a situation where rules are inconsistently enforced and audits are increasing from the payer aspect. What this means is that many of the health systems are under more scrutiny than ever. So if we implement AI without the right control, we could end up amplifying errors at a scale. On the other hand, if you hold back completely, you'll be outpaced by payers who are using the technology to their advantage, which essentially puts the providers at risk. The second is how do we respond? And that is really through innovations and technology development or through lobbying and political advocacy? In my opinion, I think it's both, honestly. The third is the question on who you going to partner with. And this is going to be really important because do you go with big players or smaller players? Obviously if you go with big companies, those guys, they bring in scale, some resources, but the smaller companies usually bring agility and laser focus on solving specific problems. So in my opinion, it's not really so much about the size, it's more about the governance. Do you have the right structure in place to evaluate and monitor and hold those partners accountable? The fourth thing is the testing and the transparency. And this is where the biggest gap is right now, which is there is a lack of standardizations and framework around the testing. Many of the AI models that I have tested personally in the past couple of years especially, you know, they're against one another, everyone wants data and obviously not, not much has been shared as a result. This putting a lot of health systems, you know, their testing tools are in silo or they don't have any visibility into how a solution is compared more broadly. And then the last thing which brings me to the most important is the human in the loop. Unfortunately, you know, fortunately, I should say AI has incredible potential. But in healthcare this is where the unfortunate part comes in. Especially in rcm, we cannot take the human out of that process. I do see many of AI technologies claiming that they'll go 100% tech proof and no human oversight. You know, this just cannot be really a function within revenue cycle. Whether it's coding, whether it's documentation, reimbursement, decision, there has to be a layer of human expertise, oversight and judgment. Otherwise the risk of errors, the bias and the unintended financial consequences that comes with that is just too high. I think the right future is for a collaborative one. AI does the heavy lift, but the human guides and reviews and ensures the output is safe and it is accurate. And it aligns with both financial and compliance. Again going back to the payers, creating more scrutiny around audits and increasing those audits. This is where the human in the load becomes really important. So just to sum up, obviously the right pace in my opinion in the RCM is a measured one, fastest, not fast enough to stay competitive. But you have to make sure it is a careful enough to ensure accuracy and fairness and making sure that there is a balance approach between both the AI and the human in the loop. So that way the outcome is really successful.
- [08:18]
A
Yeah, and to your, I mean some really important points there and the human in the loop, clearly the you got to have that because the stakes are so high. But I want to come back to what you touched on there terms of governance, because that is to me when I talk to folks where many, many are struggling and I think that bears out in some of the information we have on terms of governance. So large scale adoption, we've seen that. But really the number I have in front of me, 17%, only 17% of organizations report having a mature governance structure. So can you speak to that gap and what that really reveals about how organizations are approaching AI and what governance practices should be prioritized? Why is this so difficult? And this might be one of those a place for you to talk about the importance of agility here and finding the right partner?
- [09:07]
B
Yeah, absolutely. Even though the AI adoption is happening at a scale, it was surprising or it is surprising to hear about the 17% of the organizations say they have a mature governance structure. And really to me this shows that a lot of organizations are jumping in enthusiastically, but without the framework needed to manage risk and ensure sustainability. Obviously, without the proper governance, they're exposed to what I call the AI bubble. You know, many of the companies that are providing AI solution today may not exist tomorrow. So the question for many of the leadership is to think about if they're relying on them without safeguards or you could suddenly lose critical capabilities. You know, another common issue is the resilience. You know, the operational resilience. Many organizations don't have the archiving, the vaulting, or the data portability processes set up. So what this means if the AI vendor goes out of business or their model changes, you really have no way of transferring that experience or that data to a newer system. So essentially you're back to square one, you know, so governance is not just so much about formality, it's about protecting your investment and making sure that the AI continues to deliver the value over the time and preserving the operational knowledge. Strong governance framework is what really separates many organization that can scale AI safely from those that risks disruptions every time a vendor changes or disappears. And so in my opinion, having a balance between both large and small company and creating that agility, perhaps, you know, some of the smaller companies bring in to focus on some of the smaller solutions, while the larger companies are at a scale to deliver at a high level, perhaps, you know, might be the right balance. And at the end of the day, you know, the AI integration is just as much about people as it is about technology. And the roles like training and management and oversight and governance, all of that has a big impact as that decision is being made. So while that number seems to be low, even though the AI adoption, like I said, it is happening at a large scale. But it really means that we just need to take a slower pace in making sure that the governance and the structure is set up right and that there's communications and intuitive aspect is in place for each area, for sure.
- [12:03]
A
And want to get back to also the human in the loop comments. Right, because that's so important and obviously crucial in your earlier comments. But let's zero in and talk about those people, the humans in the loop. So what role does training and change management play in ensuring both human teams and AI models are working to improve workflows in the patient experience? So can you expand on that a little bit about how you keep the humans in the loop but also give them the tools and the training necessary to be effective?
- [12:33]
B
Yeah, I mean, one of the things that I have heard over the past couple of years especially is, you know, many healthcare organizations, they talk about information taking Information and rendering a judgment, then communicating that recommendation to the patient. And obviously there's a lot of tools that's been developed in that I believe in the healthcare, in the healthcare world that we're living in, everything really comes down to one thing which is taking information, rendering a judgment and communicating that recommendation back to the patient. That's the core of what providers do every day. The AI and other tools can help, but adopting them requires more than just technology. It requires cultural change. For clinicians, it's about trusting and integrating these tools into their workflow without feeling like the machine is replacing their judgment. Unfortunately, this is one that I see from a couple of just in the past two months I've seen three different technology companies indicating, oh, the machine will be able to predict what the diagnosis are and here's how you have to advise the patient. And in my opinion, if a physician can be replaced by AI, then that is not a good physician because I do believe that patient care is just something that cannot be replaced by technology. You can develop the technology to enhance the patient care, but replacing that and replacing their clinical judgment, it's a tough call. And you have other folks like administrators, which is they need to understand how the AI can improve their operational while ensuring safety and compliance and quality aspect. So you know, at the end of the day, one of the biggest challenge is that the machine has to be taught tone, it has to be taught context and culture. AI isn't naturally content aware. So this is why the human in the loop is so important, because it needs that continuous feedback and iteration. For example, one of the tools that we have worked with, the initial output didn't meet what the client's expectations were. They continuously saw that same line of improvement as to when they implemented the technology. They needed the human expertise to come in and review that technology, audit it and see how well it's outputting. And based on those findings, we identified that there was quite a large number of conditions that were being missed by the technology. In this particular example, it was a technology that was developed more on the value based side where it predicts diagnosis and its resurface diagnosis at the time of care. At the end of the day, the outcome was you really need to have the human on top of the machine to be that overlay to continuously train what the machine is outputting to cross verify and make sure that information, that data is accurate. Is the machine still at the same pace as it was when we first engaged with this customer? Absolutely not. In my opinion, the machine has become 10 times smarter but it needs to have that continuous feedback for the human in the loop and the human overlay to ensure the output is effective. And you know, again, implementing AI is a responsibility which requires constant validation. So and you know, most often you need that root cause analysis when something feels off and you always want to ensure you're not breaking any law or regulatory requirement. So it's not, you know, the AI adoption isn't just about building the smarter model. It is really about embedding it into the right culture, teaching it context and continuously refining it to support the human decision.
- [16:48]
A
Yeah, to your point that yes, the technology has gotten so much better, but perhaps a part of the reason that technology's gotten so much better is because that human's in the loop, the human is in the loop, that the culture is help feeding, improving the technology, correct?
- [17:02]
B
Yeah, absolutely, yeah. And more and more I'm seeing, which I am glad to see this, that more of the AI technology companies are coming to this consensus that I need to have the human in the loop, I need to have the expertise. And we should be all thinking about that. I think we should be not promoting this idea that the technology is self aware and it can make decisions and it can be just implemented and not worry about it. I think having that type of a thought process can be a bit scary just considering all the scrutiny and payer audits that are increasing day by day.
- [17:42]
A
So yeah, completely, we're just about at time. But when I ask you one final question, is there anything we didn't touch on that you want to say or maybe there's something you want to reemphasize for listeners before we let you go? What would you like to share to close out here?
- [17:56]
B
I'm just. Final thoughts is really I've just been a big advocate in like coaching, mentoring and advising and even assisting all of my clients in making sure that they understand that, you know, AI in healthcare is not just a technology problem. It is really fundamentally about people, process and culture. Oftentimes we focus on building the tool, but the real work is how they are adopted, how they're integrated and governed. And really without that strong governance and that human oversight and ongoing feedback, you really haven't developed any advanced tool and it will not deliver any sustainable value. As part of chirok, we've actually tested this through. So far we have conducted an audit and verification over 30 different AI technology. And from anywhere small to large AI technology, whether it's been on the revenue cycle or focus on the value based side and we have come to the same consistent output. My experience goes back to all the way where we had computer assisted coding and NLP and still to today we still have the human in the loop and the oversight for that. Because at the end of the day it is a technology that you have to make sure that the human is kept in the loop. The AI can accelerate workflow, it can improve accuracy, it can reduce administrative burden, but it cannot replace judgment, it cannot replace empathy, it cannot replace context. My best solution of everything that I have worked with in the past several years is really the best outcome comes when the technology amplifies human expert rather than replaces it.
- [19:56]
A
That's a great place to land the conversation. Thank you so much Kinar for coming on the podcast today.
- [20:02]
B
Yeah, thank you for having me. Really appreciate your time as well.
- [20:06]
A
I want to thank our podcast sponsor Chiroc Health. You can tune to more podcasts from Becker's Healthcare by visiting our podcast page@beckershospitalreview.com.