Joey Wahler (Host): It's an ever-emerging revolutionary tool. So, we're discussing AI for healthcare organizations. Our guest, Dr. Boykin Robinson, Chief Executive Officer and founder of Core Clinical Partners, one of ACHE's premier corporate partners. Premier corporate partners support ACHE's vision and mission to advance healthcare leadership excellence. For more information on Core Clinical Partners, please visit the corporate partners section of ache.org.

This is the Healthcare Executive Podcast, providing you with insightful commentary and developments in the world of healthcare leadership. To earn more, visit ache.org. Thanks for joining us. I'm Joey Wahler. Hi there, Dr. Robinson. Welcome.

Dr. Boykin Robinson: Good morning.

Host: I appreciate the time. So, this obviously is a topic that's on so many people's minds these days. So to get us started, when you hear AI in healthcare, people often think about clinical tools like imaging or diagnostics, but what are maybe some overlooked opportunities for AI on the operational side of things?

Dr. Boykin Robinson: Yeah, absolutely. So, I think the number one place that AI is being used in healthcare right now is with the AI scribes. We're seeing that it took over primary care over the past 12 months or so. We're now seeing it in the emergency department. We're seeing it across the inpatient organization as well. And so, I think that's really healthcare's first foray into AI at larger scale. It's still not everywhere for sure, but it's growing by leaps and bounds, this AI scribing across the clinical domain.

But also, there's a lot of opportunity for AI in things like revenue cycle, in things like credentialing, in things like accounts receivable. So, there's a lot of back office opportunity for AI, and I think we're just scratching the surface of that.

Host: Gotcha. So, executives, of course, often struggle to separate hype from reality. What are the questions a leader should be asking these days when someone pitches them an AI solution?

Dr. Boykin Robinson: So, I think the first thing to ask is to really understand how it's going to help your current process. If your current process is broken, the AI isn't going to fix it. The AI can make humans more effective. And I think in healthcare, we really have to think about AI augmenting what humans can do and not replacing those humans. There may be some of the back office elements where it replaces humans and replaces even some jobs. But on the clinical side, I think we're always going to want humans taking care of humans.

And so, the first thing to think about when a leader is evaluating an AI modality is, "Is my process good enough for this to help it?" If for instance you have terrible wait times in your emergency department, just putting in something AI in the ED may not fix that if you've got bad process with humans not doing the right thing. So, I think the first thing to ask is where is this actually going to help me? How is this going to save my people's time so that the people can keep doing what they're doing?

I think once you get past that first question, "Is this really going to help?" You've got to, of course, think about safety. IT security folks are working really hard. They have all sorts of pitches from all different sorts of AI companies, and they've got to figure out which ones are safe and secure. And so, there's that whole aspect. And then, of course, there's cost. Because, you know, AI should bring costs down. But I've also seen some AI opportunities that look to me like they're more expensive than the time they would save and would end up not being worth it. So, there's a lot to think about. It's a pretty complicated thing to think through as a hospital executive.

Host: And speaking of cost, which you mentioned there, for leaders without a big innovation budget, what would you say are some lower barrier entry points into AI that can deliver relatively quick wins, if you will?

Dr. Boykin Robinson: Yeah. So, I think this AI scribe is the first place that I would look. Some of these AI scribe companies are actually pretty affordable, monthly fees per clinician. I've seen a couple that are priced per note. I think that might get expensive, but I think that licenses for clinicians can be pretty inexpensive. So, that's the first place to look.

I think that the really expensive AI is going to be the clinical decision support stuff. So as we get into-- and I've talked about this before-- that I can imagine a scenario in which a patient walks into the emergency department and sits down in a smart chair. And on that smart chair, there is something that reads their pulse oximetry and their blood pressure and their pulse. And so, we get some vital signs. They've already scanned in their driver's license, so the smart chair knows who they are, goes through their medical record, maybe does a medication reconciliation with them. Obviously, there'd be voice as well in this chair. So, it does the med rec, it goes over what they're there for that day, and then does history. AI can now do a very directed history based on the complaints, based on the patient's medical problems, based on the medications they're on. And then, even come up with a plan that would then be sent to some sort of clinician, hopefully, sent through a human to be approved.

That's going to be really, really expensive. There's hardware involved. There's really good software. There's clinical decision support that has to be really vetted well through all sorts of different committees and med exec. It's a very complicated thing. So, that, I think, could be the future. But that's getting into heavy expenses. I think for the leaders who are worried more about, "What can I do from an AI perspective without spending a ton of money?" I think the AI scribes are a good place to start.

I think there's also some of this back office stuff. For instance, even if you spend some money using AI to do a better job with your accounts receivable, that's going to pay for itself if you do a better job with your accounts receivable. And it is surprising how many billing software solutions. When the AR person goes in to start working claims, it's surprising how often that person just starts at the top, first claim, "let me work this one, okay," second claim, third claim, fourth claim, without any plan or roadmap for which claims should be worked first. Should we work claims that are about to timely file out? Probably. Should we work claims that we think have almost no chance of getting any money back on? Probably not.

So, AI could look through the thousands or tens of thousands of claims in the AR queue and help the people doing the AR help them to prioritize what to do and make sure that they're going after as much as they can, as quickly as they can. And so, there's going to be all sorts of cost in AI. There's going to be inexpensive things like a scribe, very expensive things like a clinical decision support tool with hardware. And then, things in the middle like AR solutions that, although they might be a little expensive, may well pay for themselves if you do a better job with your accounts receivable.

Host: I just wanted to go back for a moment to something you touched on earlier. You were addressing problem solving in AI. You have maybe an example or two of a common problem that healthcare leaders should not try to solve with AI?

Dr. Boykin Robinson: Well, I think that, right now, there may be a temptation to let AI solve clinical issues, and we just don't know that the AI is mistake-free enough for that. One of the reasons that healthcare is slow to adopt technology is that errors in healthcare affect lives. And so, you can't do as some of the tech companies do you, you can't go fast and break things, is the famous Mark Zuckerberg quote. We can't go fast and break things in healthcare. We've got to be very careful with how we do it.

And so, I think that moving directly into AI, helping make clinical decisions is something that we probably need to wait a little bit on and make sure that we're doing it with a human in the loop. I don't think we, as Americans, are ready for the computer to make the complete diagnosis and to not have a human involved. That being said, I think that AI can absolutely help the human make a better decision. So, human clinician in the loop, but with all of the decision support around them. It goes back to people worried about AI replacing human jobs. And although some of that worry is founded, I think really it's humans using AI will replace humans who aren't using AI. And I think that's what we're going to see in healthcare too. We're going to have nurses who can do things more efficiently because they've got AI in the background, and we have APPs and clinicians, physicians who can do things more effectively and provide better care, because they have all of this AI in the background.

Host: Absolutely. So as the CEO, how about the way in which core clinical partners is currently using AI and where else do you think you'll use it over the next few years?

Dr. Boykin Robinson: So, the first place we started using AI was in the scribes. So, most of our emergency departments have at least an option for clinicians to use an AI scribe. And for people who aren't as familiar with what that means, what that means is that you walk in the room with your phone. You let the patient know that your AI note taker will be assisting you. And you have a normal conversation. You have a normal conversation with the patient. You take a history. It's actually improving patient experience to use the AI scribe for a couple of different reasons.

First, the clinician has more time to spend with the patient because they don't have to go write a note. But past that, the more they talk about in the room, the better their note is. And so if you imagine what happens a lot of times when you see a doctor, they talk through the history part and then they do the physical exam, but they don't always tell the patient what they're doing during the physical exam. Because I think, you know, most physicians have thought, "Why would I bother doing that?" But studies have shown patients like hearing that that's happening. "Your heart sounds great. Your lungs sound great." And plenty of clinicians do that while they do the physical, but not all. And if you're using an AI scribe, you're much better off dictating your physical exam while you're doing it, because then that'll end up in the note and you won't have to write it.

Then, you do your history with the patient and your review of systems. You dictate your physical as you're doing it, and then you start talking about what you're going to do. "I'm worried about these three or four diagnoses. We're going to order these lab tests and this x-ray and maybe this CAT scan, because I'm worried about you maybe having this or this or the other." The patient loves that. That's great information for the patient. The patient's much happier having that level of explanation from the clinician.

But also, now, the clinician goes back to their computer and their note is perfect. And by the way, these notes don't just transcribe what happened. It looks exactly like a note a clinician would've written. The AI knows what information to disregard. It knows what needs to be in a note. So, you come back, and your history is written, your physical exam is written, and your medical decision-making is also in there. And that's a really important part of professional fee coding is the medical decision-making. So if you've told the patient what you're worried about and what you're going to order, that medical decision-making gets done for you in the note. So, the clinician, they leave the room, they go back to the laptop, and immediately a beautiful note based on the interaction they just had is in front of them. Of course, they read through it, they can make changes. They can actually make changes verbally. They don't have to type in changes. So, it's pretty slick. It saves a lot of time on the note-taking side. But what's great about it is that the other benefits, it's allowing clinicians to spend more time with patients. And it's not forcing, but it's encouraging the clinicians to talk more to the patients while they're in there and let them know more about what's going on with their care. Because the more the clinician lets the patient know about their care, the better their note's going to be, which is going to be important. The clinician wants the note to be as good as it can be, both for continuity of care, the next doc who sees that note for medical legal purposes, and also for billing. So, there are a lot of reasons to have a good note. And these AI scribes are really helping. So, that's the first place that Core started using AI.

We are now looking at AI in other back office things like credentialing. And there's a lot that AI could help us with in credentialing. There's a lot of information that has to be found online that could be more automated. I think scheduling is somewhere that will get to AI. Right now, we use a scheduling software that has a very good algorithm to make schedules. But I think, with AI, this will get even better. And then, certainly, on the revenue cycle side, there's a lot of opportunity to use AI in accounts receivable, like what I was talking about. But really, all the way through the revenue cycle, AI is going to help us do a better job on the front end, have a higher clean claims rate, and make sure that we're getting in the right information at the right time, to the right payers, to make sure that we're getting paid for what we do.

Host: You mentioned earlier the importance of continuing to have the human touch, obviously, in healthcare. So, how do you balance the push for innovation like AI with staying grounded in traditional values and that human connection?

Dr. Boykin Robinson: And that's, I think, a beautiful thing about being in healthcare is that the human connection is what it's all about. If we were in investment banking, analyzing deals, then you use the AI. And I think, unfortunately for the analysts doing it, there's probably a pretty high risk of those analysts being replaced with AI, because what they were doing lends itself very well to being completely replaced with AI and there are a lots of jobs that that might affect.

In healthcare, there is a human-to-human connection. And so, our clinicians are going to see patients face-to-face. And if we can get them very well supported by AI, but still be human-to-human connection, I think that's the ideal. And I don't think that'll go away. We specialize in emergency medicine and hospital medicine. And in emergency medicine, one of the greatest skills an ER doc can have is walking in a room and understanding if a patient is sick or not sick. I'm an ER doc. That's one of the things that I know is most important to what we do. Yes, you take a history, you do a physical, you look at the vital signs, you look at the medical history. There's all sorts of things that you need to do to evaluate that patient. But when you first walk in the room, you are evaluating sick or not sick, and it just becomes innate to you as you do more and more emergency medicine.

I don't think AI can ever do that. No matter what else they can replace, they can't walk in a room and tell you if a patient's sick just from how they look. And there's something that ER docs do that just can't be replaced with that. And across medicine, I think there's this human-to-human connection that just can't be replaced. And so, using AI to help support our clinicians is important, but we always want to keep that human touch.

Also, with our hospitals, what we do is we outsource emergency medicine and hospital medicine practice management for hospitals that want those services managed. Well, there's, again, human-to-human connection. There's a partnership there. We're working for that hospital to make sure their service line is going as well as it possibly can. And that connection is not going to go away. Are we going to use AI to make our PowerPoint slides for the monthly meeting? Sure. You know, are we going to use AI to analyze the clinical operational data? Of course, we are. But then, we're going to go as a human in-person for a monthly meeting to talk to another human about how things are going, because that's how partnerships work.

Host: Couple of other things for you. Many executives, of course, worry about the workforce impact of AI, not just job loss, but also clinician trust and adoption. So, how can leaders like yourself introduce it in a way that supports their teams and makes them comfortable with that transition?

Dr. Boykin Robinson: Yes. And there's a lot out there right now about job loss from AI, and I think the real answer is nobody knows. Over the course of time, we have had emerging technologies that people thought would take all the jobs and they didn't. We just created new jobs that we hadn't thought about before, and I think that will probably happen with AI as well.

However, I think some of the jobs that are out there right now are certainly at risk and we don't know what the new jobs are. So, it's hard to say. I think healthcare is in a unique place for a couple of reasons. One, healthcare just adopts new technology more slowly than the rest of the world, right? And that's bad sometimes. it's good sometimes and there's reasons for it, right? Healthcare is highly regulated. The price of an error is very high compared to some other businesses. And so, we have to adopt technology a little more slowly than other businesses do.

That being said, healthcare does need to move a little faster on adopting technology in general. In hospitals and healthcare, there are still records being faxed from one place to another, which seems unimaginable in 2025. So, we do need to do a better job of moving forward with technology. But I don't think it's going to be the bleeding edge of AI because the stakes are too high in healthcare. So, I think the first thing that a leader who's talking to their people about workforce can say is, "Hey, this isn't going to take over tomorrow. This is going to be well thought out. It's going to come in very slowly." And again, the AI is going to be used to help people do their jobs more efficiently.

We have huge shortages across healthcare. We have nurse shortages, we have doctor shortages, we have respiratory therapist shortages and lab technician shortages. And so, we don't have enough people to provide the healthcare that we need to provide across the United States. So if we can make those folks' jobs more efficient, it might mean that they do more, but they might not even feel like they're doing more because so much was done for them in the background.

And so, it's not that we're short all those people. We don't need them to work harder. We need them to work smarter. And the way they can work smarter is with AI support. And so, I think the way to think about it is that you're going to have AI come in like your assistant, they're going to be prepping everything you do so that when you need to go do the work you need to do, any of the back work or setup work for what you're doing has already been done. It's already been surfaced up for you so that instead of reading maybe 20 pages of a patient history, maybe you're getting the paragraph you need because AI read those 20 pages or hundreds of pages and figured out what it was that you needed to see before you saw the patient. And now, that just saved 20 minutes of reading those hundreds of pages and that 20 minutes times multiple patients a day really adds up quickly.

So, I think that the way to think about workforce is that we are not in any position to decrease our healthcare workforce right now. We're short. So, we've got to support those people who are working really hard out there on the front lines, get them as much support as we possibly can with smart AI solutions. And then, those people will be able to be more effective and more efficient so that we can take care of all of the patients we need to take care of.

Host: And then, finally, in summary here, five years from now, what role do you see AI playing in the operations of a healthcare organization?

Dr. Boykin Robinson: So, I think five years from now, we really will have figured out more of this clinical decision support stuff. We will feel more comfortable with the AI really helping make decisions. We're not there yet. It makes too many mistakes, and we're just not ready for it. But in five years, I can completely see how a patient in the hospital, there'd be an AI avatar or something that is following this patient's care along. It's making sure that if a surgery consult was ordered, that the surgeon came to see the patient. So then, the hospitalist comes back to see the patient after the surgeon did. And it's surfacing the surgeon's note for the hospitals, even summarizing it. Well, actually, they probably didn't write the notes because AI was writing the notes.

And so, I think it will be in the center of all of the care. I don't think it replaces the care, but I think it's in the center of the care and allows us to be much more efficient in moving people through the hospital from the time they feel like they're sick and are thinking about going to an ER, maybe instead of going to the ER, they're logging on and talking to AI from home to decide where they need to go. There might be some AI triage about urgent care versus ER versus doctor's office versus maybe an AI consult only if we get there. And so, I think from the time the patient feels sick, they'll start talking to AI. AI will then sort of sherpa them through the journey of: Where do I go for care? How do I get this care? The care will be very supported by the AI. And let's say they're going to the emergency department, there'll be clinical decision support tools in the emergency department. There'll be clinical decision support tools on the inpatient setting that is really the center of the patient's care all the way through their hospitalization.

And then, very importantly, the AI will be a large part of where's the patient going next. Can this patient go to go home? Or does this patient need to go to a long-term acute care facility or a skilled nursing facility? And if so, let's find one with a bed. It's a real problem we have right now in healthcare, is finding those beds. I think AI will help us with that. It'll help us get patients discharged more efficiently, that'll get us more beds, more availability in the hospitals, more availability in the SNFs. I really think in five years the journey from feeling sick to feeling well, whether that's a couple hours or a few weeks, I think AI will be at the center of that journey all the way through.

Host: Well, folks, we trust you are now more familiar with AI for healthcare organizations. Dr. Robinson, valuable information indeed. It's an exciting time to keep an eye on it. And thanks so much again.

Dr. Boykin Robinson: Thank you.

Host: And a reminder again that Core Clinical Partners is one of ACHE's Premier corporate partners. Premier Corporate partners support ACHE's vision and mission to advance healthcare leadership excellence. For more information on Core Clinical Partners, please visit the corporate partner section of ache.org.

If you found this podcast helpful, please do share it on your social media. I'm Joey Wahler. Thanks so much again for being part of the Healthcare Executive Podcast, providing you with insightful commentary and developments in the world of healthcare leadership.