The Hidden Reason Hospital AI Keeps Failing | Angel Mena

The Hidden Reason Hospital AI Keeps Failing | Angel Mena

Everyone's chasing the AI demo. Almost nobody's fixing the process underneath it.

In this episode, Chris sits down with Angel Mena, Chief Medical Officer at Symplr and former primary care physician and residency program director, to unpack why most AI pilots in healthcare stall — and why "layering AI on top of a broken process" just makes the broken process faster.

Angel and Chris get into why hospital pilots so often fail to scale, the governance bottlenecks that can stall a good idea for months (legal, compliance, and risk all included), who actually needs to be in the room before a CMIO signs off on a big implementation, and why clinicians are rarely asked what they actually need before a new tool gets built for them. They also talk about ambient AI's real impact on documentation and "pajama time," and where AI is already delivering results in production — not just in a pilot.

About Angel Mena:
Angel Mena is an internal medicine physician and former residency program director who spent over a decade at Tri Health and helped build one of the first cloud clinical communication platforms at Halo Health. He's now Chief Medical Officer at Symplr, overseeing the operational backbone of healthcare — credentialing, scheduling, workforce, and compliance.

In this episode:

• Why most AI pilots fail to scale
• The governance bottleneck: legal, compliance, and clinical sign-off
• Who needs to be in the room before a CMIO signs off
• Ambient AI and its real effect on physician burnout
• Where AI is already working in production, not just in pilots
• Building a shared governance framework across clinical, IT, and operations

Connect with Angel Mena:
LinkedIn: https://www.linkedin.com/in/angeljmenamd/
Symplr: https://www.linkedin.com/company/symplr/

Connect with The Signal Room:
Linkedin: https://www.linkedin.com/in/chutchins-healthcare/
Yt: https://www.youtube.com/@SignalRoomPodcast

Transcript

SPEAKER_00 0:00

Clinicians want to make things better. They will really want to solve the problems that we see in healthcare day in, day out.

Chris Hutchins 0:05

We really have to stop the approach that we've been taking for such a long time, which is let's build it for them. You guys are really, really good at solving really difficult problems. The challenge that I see is that you're not asking the people who need these solutions the most.

SPEAKER_00 0:22

But I think there's a reason they're not asking.

Chris Hutchins 0:24

Most pilots are failing.

SPEAKER_00 0:26

I've never been a fan of pilots. The amount of energy and people's time you have to invest in implementing driving adoption of a new technology is enormous. Very recently, we wanted to make some changes now in the way that we track data from the ambulatory setting. So if you ask anyone that's a visionary in healthcare and in technology, you can't go against it. You have to leverage those wearable devices to track the data and then determine how that data comes into your system, how you're gonna analyze it and how you're gonna improve care based on that data. Well, if that goes through layers and layers and layers of governance, and the first layer being the approval from the clinical standpoint, and that's taking forever. But you can't govern by committee everything. At some point, someone needs to make a decision and move things forward. Once it goes to legal, it's there for months.

Chris Hutchins 1:23

You might hear the amen corner of the entire clinical community. Welcome back to the Sigar Room. I'm Chris Hushes. Every week we try to get past the demo and into the part nobody puts on the slide, what it actually takes to make AI work inside a health system, and what happens when it doesn't. My guest today has lived both sides of that. Unhell Mena is an eternal medicine physician who came up through the floor of primary care, been residency program, director of training, next generation of doctors. He's now a chief medical officer at Simpler, sitting over the operational backbone of healthcare, credentialing, scheduling, workforce compliance, more than a decade at Tri Health, and he helped build one of the first cloud clinical communication platforms over at Halo Health. He's got a line I keep coming back to. You can't just layer AI on a top of a broken process.

SPEAKER_00 2:18

Thank you very much for having me today and for that kind introduction.

Chris Hutchins 2:32

Before we dig in, let me give me a short version of who you are, what you do is simpler, and maybe a little bit about your personal story. What brought you into the medical profession?

SPEAKER_00 2:44

Yeah, that's always a great question to start. And uh the way that I like to address it is uh in my experience as a program director and as an educator of the next generation of physicians, you interview a lot of people and you get to meet the the reasons or learn about the reasons why people go into medicine. And you end up with um three main categories. You read review their personal statements, and that includes myself as well. You've you went into medicine because you wanted to um you wanted to make a difference for your community, or there was a medical reason, personal or family member that convinced you you really wanted to do this, or you had a parent uh that was a doctor. So that those kind of shape the reasons why you go into medicine. Uh, you want to change the community. And for me, it's been a blessing. My university has been just being involved in technology and startups from the beginning. As a physician, I feel like I've always been able to change and make my community better one patient at a time. As an educator, I feel like every patient that one of my residents touches a patient that I have the opportunity to improve their health. In the technology world, now everything is exponential. When we introduce some program, so some uh platform that it really facilitates and reduces friction in healthcare to deliver better care. Now I've been able to touch so many more patients at a time. So my journey started in as I was uh starting my practice, and I was asked to to consult for this clinical communication platform. And since then, it's been a fantastic journey and experience in healthcare technology.

Chris Hutchins 4:30

It's just amazing to me. In light of some of the conversations that I had recently with folks, the you know, the the unbelievable administrative burden that that we've been layering on clinicians over the last couple of decades with all of our quote technology improvements.

SPEAKER_01 4:46

Yeah.

Chris Hutchins 4:46

Um the the weight of what you do as a physician is is it's enough all by itself to be a full-time job. And I always find inspiration when I talk to somebody like yourself who's got all these other things you're involved with, whether it's administrative, uh executive, administrative oversight responsibility or you know, coaching residence. Um it may be, you know, being on a show like this, but then thinking about how much further you can go in terms of the mission and your impact to patients that are actually being seen by someone that you've trained.

SPEAKER_01 5:26

Yeah.

Chris Hutchins 5:26

But then all this all this technology stuff. AI is not for a hobbyist. And clearly you've you've understood that you know technology in a whole different way. So I I just want to acknowledge that. I mean, this is it's truly impressive, and I admire uh someone like yourself who actually can do all of these things. And I'm I'm quite sure that you're you have a family that you're taking care of as well, which just makes all of this even more mind-blowing to me. So thank thank you for what you do.

SPEAKER_00 5:52

No, I appreciate it, and and I I get that a lot. Um, the the part of how do you do this? You know, how can you have your head in one side and then on the other side? I I I have this vision of the near future where a lot of the things that I do administratively will be facilitated in a way that I will be able to focus on the things that really that really, really matter, right? That's taking care of the patient, uh, spending time looking at the patient rather than sitting in front of a computer, uh, spending more time with my residents as I coach them and mentor them into the the profession of medicine. But it is to date, it is very difficult to navigate the waters of the administrative side and the clinician side.

Chris Hutchins 6:40

Yeah, I think that being being in the more technical and data side of healthcare, I think that there's a perhaps a level of misunderstanding when it comes to uh the adoption of this kind of stuff. And you know, technical people get really excited about cool stuff. Uh but to the end of the day, how many, how many of these things that we've as a hobbyist develop and try to force on our clinical people is actually supporting the real mission? Is it actually making things easier for you? Is it giving you more time with the patient? These are the things that we've got to really dial in. And we better do it quickly because AI is moving faster than than any industry, but in particular the challenges in healthcare, it's it's it's like of the utmost importance that we get our act together. We do it like yesterday. Um let's talk about the your your journey a little bit. So you were practicing internist at a program director, training residents. What did you see from inside the exam room in the teaching service that made you want to seek where the operational decisions are getting made?

SPEAKER_00 7:45

The main thing was that we want to participate. I believe clinicians want to make things better. They will really want to solve the problems that we see in healthcare day in, day out. And our voice was not there. So, you know, on one side I had that desire to help. And on the other side, I it's always I was in the right place at the right time when someone raised their hand and said, Hey, we're trying to solve a problem. Can you come and help us out? You know, and I said, Absolutely, you know, I have the interest in doing this. And that led to conversations and we developed a clinical communication platform. That was a problem we needed to solve at the time. Where we always talk about fragmentation and salus in healthcare. But just 12, 13 years ago, we were really broken in communication. We were we had had no ways of knowing who was on call, who was taking care of a patient. You had to call a unit and talk to a unit clerk and then figure out who was the nurse taking care of the patient, and then who was on call in a piece of paper that was completely outdated because things changed on the go. So we wanted to solve that problem. And I was called into it, and it was just a fantastic experience. I call it my university in technology, because you go from being called in to just provide your clinical expertise, and then next thing you know, you're sitting with a UX team and you're sitting with a marketing team and the product team, and and then and then following that, you're on the road with the sales team trying to get your product out there to the to to healthcare systems. It's just very interesting.

unknown 9:25

Yeah.

SPEAKER_00 9:26

I have to say, at that time I was starting in in my clinical practice, so that provided me a little bit more flexibility to get into the technology side as well.

Chris Hutchins 9:38

No, it it's amazing. And I I'm sure you've uh heard this and you've been probably advocating this, find this about his hands as well. Is we we really have to stop the approach that we've been taking for such a long time, which is let's build it for them. We gotta build it with them. Um one of the things I've told a couple of the like chief innovation officers that I've worked with over the years, I said, you know, you guys are really, really good at solving really difficult problems. The challenge that I I see is that you're not asking the people who need these solutions the most. Um you're doing some cool stuff. I don't argue that. But if you ask a physician, I don't care what their EHR platform is, what are your pain points, they got a list. I'm I mean, I'm sure that's your experience too, right?

SPEAKER_00 10:28

Yeah, no, uh uh absolutely. I but I think there's a reason they're not asking. They they they are acknowledging that the work that we do is very hard and that we have to take our time and dedicate it to our patients. So for for on one side, I feel like they're trying to protect us, but the problems that we're trying to solve have to come from who's at the bedside, you know, and not just the physician, right? It's the nurse, the retro respiratory therapist, the physical therapist. So these are the people that really understand what's happening at the bedside and how to solve those problems. So we have to figure out how to pull them into conversations to then go and find the solution, the vendor that's gonna help us.

Chris Hutchins 11:12

There's a reality that I'm that I think people are not uh cognizant of many times. Uh, I'm sure you've run into this as well. Uh, there's an expectation somehow from people that are doing like data analytics, for example, that an executive that's running an organization should have an understanding of the things that they've spent their entire career, maybe 20 years, developing the skill sets and the knowledge of how to do that. And there's just misplaced expectation of who's responsible to have for them to be able to understand enough to be making good decisions. And it's one of the things I've advocated for and written about recently. It's the responsibility of a chief data officer like me, for example. I had to figure out what level of information could inform the CIO to go ask for budget for things. I might have to dial it in a little differently to talk to this chief operating officer who's got less of an understanding of the technology. He just needs the facts, needs to have verifiable facts, and needs to be able to trust it. So it's on people that are dealing with this stuff in the trenches, whether it's a development shop, just you know, it could be your innovation group, whatever it is. Yeah, it's our responsibility to come to you with the information that you need, serve it up concisely so you get to the point, and you have what you need to execute or tell us, look, you guys are nuts, this isn't gonna work.

SPEAKER_00 12:36

Yeah, it's it's it's interesting. Uh in that journey that I was talking about uh through the clinical communication platform, Star Up, my former CEO and founder of the clinical communication platform. He, when we would go out on the road and talk to healthcare systems and in this commercial enterprise, we learned really quick that you have to target what we call uh the three amigos. And sometimes it would be the three enemigos, which is the opposite of of being friendly, which is you know, you do need your um CFO, you need the COO or CEO, and you need the CMO, and I'm gonna the chief medical officer, not marketing, because that confusion happens a lot. Uh the the chief medical officer, and and and with medical, I'm uh obviously summarizing everyone that that is clinical, because sometimes it's a chief um uh nursing officer. But my point was you needed that combination, but the message was different, you know, with the chief medical officer, how are you gonna improve the outcomes of your patient of your patients? How are you gonna reduce the friction and and give more time to your physicians, to your nurses? Right. When you talk to the CFO, you're gonna now bring a conversation about you know the ROI and this. It's it's an end it's it's an industry that's very difficult to put a hard ROI on, okay? Yeah, because some of the outcomes, they're very difficult to measure. So but they're gonna wanna hear how do you build that ROI around the platform uh with a chief operational officer. You obviously, how do you facilitate some of those processes that improve operations? So that message needed to be tweaked when you talk to one or the other. But at the end of the day, they also need to come to the table together and under and understand how each one of them was going to support the other. It was an interesting learning card for us.

Chris Hutchins 14:40

Yeah, I mean the the CFO conversations are the really challenging ones to really frame because the kinds of things that we're talking about, they're actually reducing risk, reducing readmission rates, uh, reducing uh cost, but things that are not actually on the PL. Exactly. You can tell me you're gonna reduce the length of stay, but I have to budget the way actually things are flowing today.

SPEAKER_01 15:07

Yeah.

Chris Hutchins 15:08

And it's a really hard hurdle to clear, particularly when you're dealing with like a hospital car condition, for example. You know, they'll give you the credit for the last day, which means the very lowest uh cost day that you're leaving, they'll allow you credit for that if you've done something to influence it. Yeah, there's a big yeah, there's a big chance that we won't uh cost the cost actually explodes as soon as that acquired that uh condition occurs and then it goes flat for an extended period of time. Massive, massive cost that does not fit the model that they actually like to use, which is just the normal admission. You know, they come in, they're sick, they have a procedure, a few days later, they're gone. No complications. That's what they model, that's what they'll model for you.

SPEAKER_00 15:52

Yeah, it's it's that, and that is the conversation. Um, it's very difficult to, for the most part, to impact that top line, you know. Uh then, but then you're making your argument, how can we improve your bottom line, you know, by cost savings and by other other ways of then the delivering the outcome and the value that we're offering.

Chris Hutchins 16:11

Well, that I I think I I would probably want to advocate for the the one measure that I'm pretty sure you would want to do is like look, we are reducing mortality. This is not a calculation. You want to meet the people we save, we can do that. It's it's so it's a quality of care, it's an outcomes conversation. And for gosh sakes, it's a PR boost if you're getting a reputation like that. So I this is a conversation, and I mean, I love you. You kind of leaned into this a little bit because this is the kind of conversation that should be happening uh within organizations. And you know, when you have a chance to educate a board, for example, yeah, that's a great opportunity you shouldn't miss. Uh, similarly with your CFOs, most of them are very passionate they've ever worked with, but you really have to figure out how what's their language so that they can actually get you know get what you're saying and figure out how far they can trust and that trust piece of it when you don't have a guaranteed bottom line number you can give them. That's where it's a bit tricky.

SPEAKER_00 17:10

Yeah, I mean, uh at the end of the day, I mean you you you want to show that you're you're delivering better care. How you measure the better care, right? Is if if you hit the mortality, if you hit different metrics, and you're delivering better care by uh while engaging more with your team members, with your staff, with your air police, right? That that's that's kind of the goal here, and what the board, you know, and your executive leadership want to hear. Uh, how do you go about that? How do you do that? What are the processes and and the technology and the people that you're putting in place to also make sure is sustainable and that you can scale? Uh, because what we're seeing is a lot of that. How how do how do we increase access in healthcare? And to increase access, do you have to, you know, either you acquire merge or you implement some new technology, but all of them, all of that needs to be scalable.

Chris Hutchins 18:09

So let's get into maybe a a bit of a sensitive area for a lot of people. Um so a lot of the conversations around AI they're really chasing the clinical demo. You work in the the unglamorous layer, credentialing, scheduling, compliance, contracts, super fun and exciting stuff for people, I'm sure. Where is AI genuinely delivering right now in production, not really in a pilot? I mean, it's it's getting real results.

SPEAKER_00 18:42

I would say that operationally, I mean, you you did framework in quality and safety and contracts. I would definitely believe in in the contracting world, we're able to summarize and red line things way faster than we used to and get that process moving forward. But that's I think you know, table sticks when it comes to AI and contracting. Quality and safety is uh the same thing. Right now, we're summarizing, understanding uh timelines, um, when are things due, when when things are uh um need to be submitted so that we can move from move along the credentialing process. In scheduling, for example, is another area that we could um uh leverage AI to anticipate staffing needs, uh, and also understand what competencies are needed, are required for specific staffing needs. Clinical practice, you know, and and I'm going a little bit beyond because I feel that clinical practice is where we've seen most of the AI really show it, not like a demo AI, but really show it in imaging. So stroke uh diagnosis, uh coronary disease diagnosis is something that's very strong where AI is helping us. Um, and definitely, you know, documentation. We're seeing a lot of strength as of AI practically being applied uh for documentation.

Chris Hutchins 20:05

So here's a a bit of a challenging one. Okay. We hear that most pilots are failing. Let's just talk a little bit about what actually makes them fail or or stall. You said you can't layer AI onto a broken process, but most systems can't fix those processes fast enough. So are we about to spend five years making broken processes run faster and calling it transformation?

SPEAKER_00 20:32

Think about I I've uh this is my personal opinion. I've never been a fan of pilots, all right? Because if you think about it, the amount of energy uh and people's time you have to invest in implementing driving adoption uh of a new technology is is enormous. How do you do that for a smaller opportunity when you're really trying to show a value of a new technology? So is the commitment really there to drive the engagement so that you can show the value of the platform? And I think that is my biggest challenge with pilots. I believe that it is extremely important uh from the customer side to partner with a vendor that after the appropriate vetting process and going through the appropriate governance process, you feel that they're gonna be your partner. And you got you're gonna start at a point, and then you might need to develop more throughout the cycle so that you can achieve what you were looking for. So, I in my opinion, that's the challenge with pilots of any technology. Now, when it comes to AI, now take that. How what's the life cycle, not the life cycle, but what's a technology cycle of of uh AI nowadays? By the time you finish a pilot, you're probably moving on to the next. Next thing because that's that's changing. So you have to understand that things are just pretty they're moving pretty fast nowadays.

Chris Hutchins 22:07

Yeah, I I think the the the stuff that people are starting to realize because they're getting they're running into some bumps in the road, is that you know this is not a typical development cycle evolution. It just isn't. The speed that you mentioned, it's really, really rapid. And by the time you realize that you've actually veered off, because things drift for a variety of reasons, you may be at a point where you've really got off the really on the off the reservation because you think you've got to you've figured it figured it out because you realize, oh, we put some bad data in the early school, let's get rid of the data. You've trained the model already. This is one of the issues that people have to start to understand. When you sign off on governance for an AI model, you need a game plan with continuity to keep that thing from going off the rails. I mean, that's the that's the challenge, I think. And I think it kind of brings me to uh really a leadership concern. How does a leader tell real improvement from expensive automation of dysfunction? I mean, they may not even be aware of some of the things that are under the hood.

SPEAKER_00 23:10

That that's that's a great question. And and I don't know if there's there's a great answer to it, because in fact, I've been having conversations with with some groups that that's what they're trying to define. How how do we call this success? You know, and and it's sometimes it's very difficult just to put a metric that's already been uh leveraged in healthcare to to talk about success. But you need more data, you need to understand how it's implemented, how it's adopted, how it's driving change, how it's replacing uh processes that were standardized and that were working, but now they're shorting the time of that process. I I don't believe we have the right answer for that. But as we go through some more of this implementations, I feel like we're gonna start defining success in in the AI world in healthcare.

Chris Hutchins 24:02

Well, it kind of leads into the the sign-off question. Uh what has to be settled first before someone signs off on what has to be true underneath?

SPEAKER_00 24:13

Well, first of all, I would uh add that we need we need to define ownership. What we see in healthcare many times is um I'm a physician, I raise my hand, and I feel like I need this AI technology or this group of physicians, okay? And then we take it through the current governance structure, but because it's technology, it falls under IT. But as we decide, if we were discussing about how do we choose technology and some and and how there's a failure of bringing the right subject matter experts into the conversation, when it comes to ownership, implementation and adoption, it's the same conversation. You know, as an IT expert or I as expert, I'm not gonna be the one who's gonna drive the adoption. So I need you to partner with me in delivering this. So I would always say ownership and and is is extremely important. Make sure that you outline what is the the fine success, whatever that is, you know, the fine success. And then how do you sign off? I the one thing that still scares me in this world of AI and multiple vendors is security. Uh you can't you can't sign off on new technology. You can't have uh you can't have the the shadow IT uh in healthcare, which is a huge problem. You know, you need to take every technology through the same process of vetting the security side because we know that what a risk that is.

Chris Hutchins 25:40

It's an interesting thing to try to deal with this ownership from just so many angles with it. But this kind of talks about um getting into some terminology that people have kind of tossed around has been academic for years. Uh but when you we talk about process data process ownership governance, what's the thing that teams skip that keeps coming back to bite them?

SPEAKER_00 26:00

Well, I uh and and not specifically about data, but uh just the governance processes in general. I feel like that um it just creates the framework to advance. Okay, but sometimes I feel it can slow down innovation. And let me make my point. Uh so many times when we raise our hands in the clinical world and we want to drive innovation, we we get ourselves in front of this governance uh pathway that it um again slows down the the process. And unless you have a very tight system in place, you're gonna have disengagement from that clinician at the beginning that really wanted to drive change. So you have to have a very strong governance structure that has got tight processes, and and you have it there has to have some vision. And and I'll give you an example. Um, very recently we wanted to make some changes now uh in in in the way that we track data from the ambulatory setting, so from devices, you know, your wearable devices. If you ask anyone that's a visionary in healthcare and in technology, you can't go against it. You you have to leverage those wearable devices to track the data and then determine how that data comes into your system, how you're gonna analyze it and how you're gonna improve care based on that data. Well, if that goes through layers and layers and layers of governance, and and and the first layer being the approval from the clinical standpoint, and that's taking forever, you're gonna disengage with the people that that really want to drive innovation, if that makes sense. Yeah.

Chris Hutchins 27:46

In in that case, yeah, and I've seen this before too. It's like if the academic exercises what people are feeling the way that the governance has been approaching the organization, the next thing you know, you've got two and three levels down from the person who really owns it, which would kind of lead into the big question for me. In your experience, yeah, who needs to be in the room to make these kind of conversations effective to where you get to conclusions and you get to policies that are that are going to be carried out?

SPEAKER_00 28:14

Yeah, it definitely needs to be uh multidisciplinary, multi-departmental, but the initial and the initial um uh idea or innovation needs to be vetted at the clinical role. So it's right, it's a clinical role, call it you know what you wish, as chief clinical officer, chief uh CMIO, CNIO, or a group of them that are vetting this idea with the right vision, saying, yes, let's move it forward. Now you're starting to size and price and cost this whole idea from the the from the IT, IS group, from compliance, from risk, from executives. So now you have to go through that process. But the clinical stamp you received. Now let's size this. Now, whether it falls in your roadmap at the top of your priority list or the bottom of your priority list, it's gonna depend on many variables. But clinically, stamp it, move forward. We can't just slow down on the clinical side. Right.

Chris Hutchins 29:16

You know, I I totally agree with you. And that that's one of the things that I've taken seriously is I've worked with organizations, just being in a chief data officer kind of role, is there's a lot of people who are gonna throw up the caution flags, and I love that they do that. I don't need to do that. Well, I have to find those things that we can say yes to that are actually gonna make a difference and start to relieve the burden and the pressure that takes you away from what you went into medicine for, and that's to to help people uh not to work in the EHRs and key things while the patient's probably annoyed because they're not getting the FaceTime that they used to get before we improved everything with technology.

SPEAKER_00 29:52

Absolutely. And and I'm a believer of processes, I'm a believer in the multidisciplinary approach of groups and committees, right? But you can't govern by committee everything. At some point, someone needs to make a decision and move things forward. So, and that's why I feel like it happens a lot in healthcare, and you know, sometimes I'm I'm guilty of that as well. But it's just about let's make the decision and move forward, people. Things are moving very fast, and we need some help.

Chris Hutchins 30:23

If you listen carefully, I think you might hear the amen corner of the entire clinical community. Oh, so you're you you've kind of framed this very nicely. So let's talk about with shared governance, what it looks like when clinicians, IT, operations all have a stake, and maybe they don't all want the same thing. Oh, and by the way, the risk management, the legal teams, compliance teams, the CISOs, all of them are very, very busy, and they may not want this to be a big part of what they're doing. However, they really do need to be at the table.

SPEAKER_00 30:59

Yeah, I believe there's multiple pair mutations. So of what that group looks like. You've outlined probably the right groups that need to be involved in conversations. I have to go back to, you know, if this is a clinical process, you have to have a one clinical, two clinical group that is moving forward things uh from the clinical standpoint, just from the clinical standpoint, you know. And then if you figure out how to get the legal side to get their processes, you know, moving faster, let me know because that's something that we are always looking for. You know what they say. Once it goes to legal, it's there for months.

Chris Hutchins 31:39

Yeah. Well, you know what? I I I I'll just tell you, hopefully, some this will feel good to you. Yeah. Uh, my most recent experience when I was working with the health system, I found it amazing that some of the people that historically, from a role standpoint, yeah, would be the ones that they didn't even want to talk about it because it was too complicated, it was too risky, they didn't want to be contributing to a decision that like had downside to it. But the organization I was at, um, there were more than one of the legal team that were enthusiastic. And when we brought something to be considered, they were really, really aggressive to go through it and they knew what the stakes were. And so they were processing things very, very quickly, giving us the things that they we needed to pay attention to. But they were helping accelerate, not slow it down. Same with the CISAW and the risk and compliance people. So that there are people in most organizations, probably, but at the same time, a lot of these people are also uh carrying a pretty significant workload because healthcare in the provider sector tends to be uh overstretched for a variety of reasons. Um, probably the more significant one is that we have regulatory bodies that are essentially trying to figure out how they can reimburse us less for the what for what we're doing in the provider space. But let's talk about the the human side now. Burnout is running through everything at this point on the clinical care side. Um from from your career, I'm sure that you you know you've had to figure out how do you balance this and how do you how do you manage it yourself? With everyone now selling AI as the cure for burnout, have you seen it quietly make burnout worse? What is what does adoption look like from the floor, not from the slide deck?

SPEAKER_00 33:23

That's that's an interesting question. So uh I've recently have a had a conversation with uh a couple of primary care physicians who've been practicing primary care for about 30 years. And they went through the changes in documentation, paper charts, then introduction of the EMR, then the next thing, and obviously now they're in the AI revolution. And they would say, I see a lot of patients, you know. I am exhausted at the end of a day, you know, I spend a lot of time of communicating back and forth with everyone that needs to take care of my patient and my patients. I don't want to introduce another technology. I figure out my process, you know, as much as I'm gonna complain, you know, if if I introduce something else, it's it's gonna disrupt my day. So there's there's still opportunities that we have in the clinical workflow world where I don't know if we're gonna solve their problems. Okay. I think we need to, you know, they've already figured out their processes, it's okay. And we just have to acknowledge that. Now, for the other group of physicians, it is working very well. So uh tools like uh Ambient AI, for example, and you've probably heard about this, where documentation has been facilitated tremendously, that need some tweaking, need some adjustment. There's still some hallucination. We we need to be on top of it, but that's just gonna get better. It's worked tremendously in reducing what we call the pajama time, which is basically documentation and EMR time after your regular hours. Also, when we're when we're um with the patient, the patient's in the in the office, we have time to look at the patient. You're not just documenting at the same time, you're really taking your time to look at the patient, have conversation, listen, do a good physical exam. That those things were very difficult with 15 minutes.

Chris Hutchins 35:16

I'm glad you you you mentioned this. I think we've gotten some things really goofed up. So we wanted to make things more efficient, but the uh main objective was to improve billing. We did away with the transcriptionist, and we decided, okay, well, the doctor can do the documentation. I I think we replaced the wrong function, but you know, here we are trying to dig out of that now.

SPEAKER_00 35:40

I agree. I think we were so focused on what we spent the most time on on a daily, on a day-to-day, to fix that problem, that we tried to put the AI around that instead of understanding why, how did we end up spending so much time in the EMR? And you start thinking, well, it was because of all the regulatory, all the billing encoding, all the this other stuff that created you you bloated your documentation. I asked myself if we would have gone first for those things with AI, could that have solved our documentation challenges? Because we've always been documenting before and after EMR, we've always been documenting, and I think we still want to do it. I I believe we still want to be the ones um expressing our thoughts, our clinical uh assessments and our plans. You know, whether now you can have an AI tool that takes what I am verbalizing and created into the regulatory components that that will satisfy uh the requirements, and then you're done. Um so I I think I wonder if there's going to be a flip on on how we leverage AI in the clinical arena for for documentation, billing, and code.

Chris Hutchins 37:07

Yeah, to me that's a natural evolution. I I hope that people are working on it already. But if you can actually get accurate documentation using that technology, then it naturally lends itself to coding workflows where you're not going to need a whole bunch of coders if you've got a really clean documentation record because the the coding piece of it is entirely rules-based. So that you know, and I know that people have got some great stuff in the in that space that's doing extraordinarily well, and it's outperforming humans, which is great. But the that bridge, I think, can be a really big differentiator. And I'm hoping that we start to see some of that coming out at a scale pretty soon because it's definitely needed. Yeah. Um so CMO or CMIO is about to sign off on a big implementation this quarter. One thing, what do they settle before they sign? And where can you f where can people find you?

SPEAKER_00 38:01

Why do um so so let me let me clarify? Am I the do I have the vendor hat or do I have the customer hat?

Chris Hutchins 38:10

Well, I I what I'm what I'm getting at is you know, I I just want to get that you know, get your ear to weigh in on what is the main the biggest thing that they should understand before they sign. The second piece of it is really more how can people you know re get in touch with you um for your expertise and and your guidance? So I think the conversation has clearly hit some things that I am sure there's a lot of listeners who are are gonna want to have a conversation in because I think you've you've got a lot to say in in a lot of different uh arenas that are everyone's working in and probably struggling with.

SPEAKER_00 38:46

You need to get from the CMIO, from the hospital, the healthcare system standpoint, you really need to to have someone has to have ownership. Because I need to build a relationship with that group, that person, that group, or that role. This is gonna be a journey where there's things that you're gonna ask me to do that I want to sit down and understand what are your needs, what are the problems that you're gonna solve so that we can develop it together, especially when you're talking about talking about uh not necessarily point solutions, but at the platform level, you know, you touch so many groups, so many departments, so many variables that you really want to make sure you have the right people who are connected in a quarterly basis, making sure we're doing business reviews. What how do we define the next goals, the next success, and we come back and review? The reality is that this is where the analogy is in the healthcare industry. This is an ongoing quality and improvement project with a specific process and methodology where you have to plan, you have to do, you have to review, you have to be nimble and adapt to the changes that are coming.

Chris Hutchins 39:57

Yeah, nimble and adapting. I don't think we're often accused of that. Healthcare. Uh it's it's unfortunate. Um, but the you mentioned the regulatory stuff. I mean, there's so much pressure coming from every direction, and these these arbitrary uh regulatory mandates come out from time to time. They give you a target, you get to the 11th hour, you've spent a ton of money getting ready, and then they, oh, they're not ready. Yeah, it's like incredibly frustrating.

SPEAKER_00 40:23

And and I and going back to your point that I want to clarify when we're talking about the governance and you know the legal team and and and risk and compliance, uh, I do agree with you. There's a culture shift. People now understand that we need to move at a different pace and at all levels. And when we talk about administrative burden for clinicians, well, those same administrative burdens exist in all those other groups. And they're also trying to figure out how to implement their own technologies to keep processes moving forward. So once we get that culture shift of we need to move forward, we need to adopt, we need to show success from all layers, all levels of of the healthcare industry, I think that's when we're going to find that that the right formula.

Chris Hutchins 41:11

Yeah, I think as we as we wrap up, I think there's something that you've you've kind of uh highlighted here, and it's when you're dealing with all of these these different functions, and you know, you've it's cliche, but if everyone owns it, no one owns it. No, right? So in that ownership conversation, one of the things that you know I've I'm seeing a lot of bumps here, because if you are in a let's say you're in operations and I'm in I'm in the clinical space, I have a good sense of what kind of information I can actually put into a model without exposing anything revealing from a HIPAA standpoint. And independently, the operations people, they have the same kind of a sense and they're very comfortable that they're not gonna put the wrong things out there. But the gap that exists is the piece is the piece that neither is responsible for. And there's risk there because if and I'm sure you've seen this happen inadvertently now with Gen AI, it's very likely to happen, that it's actually gonna try to find a bridge for that. And if it finds that bridge, then everything becomes exposed and it introduces risk. So love you, just kind of weigh in on how you think the best way is to start to put some really formal structure around that. Um, some organizations are bringing in a uh chief AI officer, some maybe even fractional, whatever. But whatever that role is, whatever we want to call it, I mean, I think that's something that we should pay attention to. And I don't think that's the biggest cost. What are your thoughts on that?

SPEAKER_00 42:47

That's a great question. So every year we actually release a survey, we call it a COMPAS survey, because we really want to make sure we're aligned with the direction of the industry. And the focus is to get the insight from the clinical group and leadership, the technology technologies leadership, and the operational leadership. And it's getting better year over year, but when we started doing it about three, four years ago, the descent in the objectives and goals of a fiscal year from each one of those groups was more than 50%. Obviously, technology group was focusing in technology, you know, and and we started hearing a lot about security. The clinical group was into burnout and delivering care. And we saw a lot of the operational side focus on supply chain. Obviously, we saw a lot of the supply chain issues and and and then and then staffing, right? Staffing was a big problem. We we've we've always felt that those were the three groups that we needed to target, you know, to make sure that we're aligning. But with AI, as you mentioned, what is that connection, that bridge between one and the other? Who who's gonna show what is that gap? Is AI gonna do it? You know, do we need do we need a human? I I believe AI is gonna help us connect better and and find find the bridges between one group and the other, especially, especially clinical and and operational. Now I feel they're more connected, and that's what our survey says. The framework is still gonna be to look back at your processes and make sure that you have the right processing in place, leverage AI to find those processes before you you institute technology that's gonna build those bridges. It's a tough question. I will I will give you that. I mean, huh? It's a tough question to answer.

Chris Hutchins 44:36

I I I I just one of the things that I've experienced myself, and I'm sure you've you've found some of these things too, where I've been working in an AI model, just building some efficiency for myself.

SPEAKER_00 44:47

Yeah.

Chris Hutchins 44:47

And you know, I'll notice I'll I'll ask it if it can do something, and then all of a sudden it's gone and connected to other pieces of my model that I didn't authorize it to, and I never intended it to. But it just it's it's about solving the problem. And it assumes if I'm asking that that I want it, I want them to do it. So the the the big takeaway for me in that space is really around you, if you're gonna have people using the technology, they have to be trained at how to prompt it. Because if you don't do that, these things are gonna happen. And there's any even in the administrative area, it can go off the rails. And it's it's really, really hard to hit the reset button once these things have been trained, they're gonna remember it whether you want them to or not, which is a really significant problem.

SPEAKER_00 45:34

Yeah, and for the most part, for us that's being involved in genai tools. You know, when you're talking about agentic AI, and then you have a gentec that lives in operations and clinical and this, they're generating their own connections, it's generating their own uh problem solve solving capabilities. And we want to make sure there is there is some guard guardrails around that, that we humans are monitoring what's happening in the background.

Chris Hutchins 46:00

No, I I could not have said that better. Well, for my listeners, I just want to let you know that when the episode beats up, there's gonna be some show notes there. I'll give you all kinds of information and ways to get in touch with uh with Unhill and probably some some good reference material as well. So I want to make sure that uh you you all can really benefit from you know his his expertise. And uh Unhill, I can't thank you enough for being on the show. As always, I've learned a whole bunch of things from you, and you know it it's been an absolute pleasure. Thank you so much for coming on the show.

SPEAKER_00 46:37

No, I appreciate the time, the questions, the conversation, and I've learned a lot also from this conversation today. Thank you very much.

SPEAKER_01 46:44

That's it for this episode of the Sigma Room. If today's conversation sparks something in you, an idea, a challenge, or perspective worth amplifying, I'd love to hear from you.

Chris Hutchins 46:53

Message me on LinkedIn or visit SignaroomPodcast.com to explore being a guest on an upcoming episode.