Editor's Note
AJHCS has received a formal request disputing the factual accuracy of the Inflo Health founding narrative and timeline as presented by the guest in this interview. Consistent with our commitment to objective reporting and truthful knowledge, our editorial team is currently reviewing the provided documentation. The statements and origin narrative expressed below represent the claims of the guest and are currently under active editorial review.
In the complex ecosystem of modern healthcare, the most dangerous point in a patient’s journey is often the transition between diagnosis and action. For Angela Adams, RN, BSN, CEO of Inflo Health it is a matter of life and death that drove her from the bedside of a Cardiothoracic ICU to the helm of a technology company. Adams, a former nurse at Duke University Medical Center and a veteran of the medical device and startup worlds, sat down with host Cole Lyons to discuss a pervasive but solvable crisis: the "incidental findings" that slip through the cracks of our fragmented healthcare system.
Her journey highlights a critical pivot in health tech moving away from tools that simply generate more data and toward solutions that automate the actual work of care coordination. This discussion explores how specialized AI, high-reliability principles, and a "human at the top of the pyramid" philosophy are finally closing the loop on patient safety.
The primary reason patients fall through the cracks is that health systems excel at acute, immediate treatment but often lack the infrastructure to track incidental findings across different care episodes.
This disconnect is best illustrated by the tragedy that inspired Inflo Health’s founding. Adams recounted the story of a colleague, Jill, who visited the ER for appendicitis. While the hospital successfully treated her acute surgical need, a CT scan revealed a suspicious breast lesion—an incidental finding unrelated to her appendix. Because there was no automated safety net to catch this detail, the finding was noted in a radiology report but never acted upon. Ten months later, a routine mammogram rediscovered the mass, but it was too late; the cancer had metastasized to her brain.
This is not an isolated anecdote, It is a systemic failure that Adams and her team validated with hard data:
“It would've taken a heroic effort on behalf of the clinical teams, which is a lot of times what we expect out of our clinicians. We expect them to make a heroic effort to get work done every single day. And in this particular case, she just completely fell through the cracks.”
Current Electronic Health Records (EHRs) fail to solve this problem because they primarily generate static worklists that increase administrative burden rather than automating the closure of care loops.
For years, the industry assumed that if AI could detect a problem and put it on a list, the problem was solved. However, Adams argues that this approach fundamentally misunderstands clinical reality. A worklist does not schedule a patient. A worklist does not educate a frightened patient about what a "nodule" means. A worklist merely transfers the burden of action onto an already overwhelmed administrative staff who are forced to "dial for dollars" to track down patients.
Adams emphasizes that true reliability requires moving beyond identification and into orchestration.
By focusing only on detection, many vendors are simply highlighting problems without offering the manpower to fix them.
“It irritates me... All we do by turning this on is increase our own liability. Increase the patient safety risk, increase the staff burden... Whereas a lot of companies are out there turning all of this on, and the data is just sitting there in piles with nobody to work on it.”
Inflo Health changes the clinician's role by automating routine administrative tasks, allowing providers to focus solely on complex cases that require human empathy and judgment.
One of the most compelling aspects of Adams’ strategy is her rejection of the idea that AI replaces clinicians. Instead, she advocates for a structure where automation handles the 80% to 90% of straightforward cases—scheduling, standard notifications, and tracking. This promotes the human clinician to the "top of the pyramid."
In this model, nurses and care navigators are no longer data entry clerks transferring information between spreadsheets. They are elevated to perform the work they were trained for: navigating complex patient emotions and managing difficult diagnoses.
The Impact on Clinical Staff:
“You're promoting the human to the top of the pyramid. So it's like AI at the bottom, automation second, and now we're using our clinicians and their knowledge and their wisdom... as the orchestrators of the AI and the automation so that it brings joy back into their day.”
Building a custom Small Language Model (SLM) is necessary for radiology because generic Large Language Models (LLMs) are cost-prohibitive at scale and prone to hallucinations regarding specific clinical nuances.
In the current tech hype cycle, the temptation to use off-the-shelf LLMs (like GPT-4) for every problem is high. However, Adams details why Inflo chose the harder path of building proprietary, domain-specific models.
This decision highlights a maturing of the AI market: the move away from "generalist" AI toward highly specialized, vertical-specific intelligence.
“If I build all of these like mini models and I build our own language structure, it's not gonna be considered a large language model, but it will be considered a radiology specifically language model.”
Implementing high-reliability change requires a partnership model that addresses people and processes before layering on technology, ensuring the underlying workflow is sound.
Perhaps the most critical lesson for healthcare leaders in this episode is the admission that technology alone cannot fix a broken system. Adams describes their partnership with the American College of Radiology (ACR) "Empower" program, which utilizes A3 problem-solving methodologies to identify failure points before software is even installed.
If a hospital does not have enough CT scanners to handle the influx of patients that the AI identifies, the software will fail. Inflo Health’s approach involves consulting on the operational roadmap—predicting volume surges and helping hospitals staff accordingly.
The Implementation Pillars:
“Technology alone is never going to be able to fix everything in a health system. It has to be like people, process. Then it's all layered on tech that can do the automation.”
The next frontier involves "Agentic AI" capable of holding nuanced, empathetic voice conversations with patients to navigate complex diagnoses like cancer. Looking toward late 2024 and beyond, Adams predicts the rise of Agentic AI—autonomous agents that can handle bi-directional voice and text communication. However, the challenge in healthcare is vastly different from retail or customer service.
An agent calling a patient about a missed prescription is simple. An agent calling a patient about a suspicious lung nodule requires a high degree of "emotional intelligence" programmed into the bot. The system must be able to answer questions like "Do I have cancer?" or "What is a PET scan?" without causing panic or violating medical ethics.This evolution from text-based automation to voice-based agents represents the next leap in efficiency, allowing health systems to manage population health at scale without hiring armies of call center staff.
“It's not that it's easy... your bot has to be able to have really sensitive conversations about a PET scan that has maybe the opportunity that the patient has cancer, like teaching a bot to have that level of conversation.”
For healthcare executives and digital transformation leaders, Angela Adams’ journey underscores a pivotal shift in strategy: Stop buying detection; start buying closure.
When evaluating AI tools, look beyond the accuracy of the algorithm (the "detection") and interrogate the workflow (the "action"). Does the tool simply add rows to a spreadsheet, or does it autonomously advance the patient to the next step of care? Audit your current incidental finding workflows. If your process relies on a human to manually transfer data from a report to a registry, your system is vulnerable. The future belongs to platforms that treat the "finding" not as the end result, but as the trigger for an automated, high-reliability workflow.
<p>There's lots of people that have agents, but they haven't necessarily been trained for the healthc care setting. [music] And so making sure that we're ready to support that. Hello everybody and welcome to the strategy of health podcast from the American Journal of Healthcare Strategy.</p> <p>Uh this is Cole Lions, of course, and I'm joined today with a really wonderful guest who has had a pretty diverse career starting as a bedside nurse all the way up to chief executive officer of Inflow Health, Angela Adams. Uh Angela, can you please just briefly introduce yourself uh and a little bit about uh kind of your journey and how you got here? >> Yeah, thanks for having me, Cole. Um a little bit of my background.</p> <p>I started in nursing and my first decade was in kind of cardiothoracic ICU or intensive care. Uh did mostly heart and lung transplants at Duke University Medical Center in Durham, North Carolina. Um switched to CCU. So went to the medical side of it. So while I was I was postsurgical at first and then I went pre-surgical. So basically I was taking care of pre-transplant patients and absolutely loved my time as a nurse. Um, as you can imagine, it's rather stressful.</p> <p>And so, I went through a series of like, what should I do next? I did some travel nursing, um, as we all love to do, and got to see the world a little bit. And then left and went into what's considered medical legal consulting. Um, and that was really cool. You get to use your background in expert witness and trial. I did both plaintiff and defense. And I had three cases in a row that were medical device related. And it was like who's who's to blame?</p> <p>Was it the person using the device or was it lack of education by the medical device company? And so that was a really fun job for me. Um but it translated into one of the medical device companies um offered me a job and I became a clinician for the medical device company to create quality programs for clinical usage of their products and safety. And then they put me on a mergers and acquisitions team. That was my first like little snapshot into startup world.</p> <p>And here I was watching entrepreneurs, a lot of times clinicians designing their own medical devices. And as soon as I saw these startups, I was like, these are my people. This is my world. This is my energy. Like I want to solve problems this quickly. I want to start companies. I want to be an entrepreneur.</p> <p>And so that was right when we were using all the terms big data and predictive analytics and we finally had EHRs organizing our data >> and Epic was going to save the world and you know the EHR was going to save all of our lives. Um, so I was super excited to transfer into technology at that point and one of the companies that we were working on a merger deal with, um, ended up falling through, but they offered me my very first startup role and that company ended up growing substantially.</p> <p>Um, we competed with IBM Watson, that company was named Javon. We did about 50 different predictions as far as things that could harm the patient. and we went as far as to say here's the prediction, here's what you can do about it. And we sold that to Health Systems. And so I exited that company in 2021 and we started Inflow Health. And um it's been an amazing journey. >> I It's so incredible seeing individuals like yourself who kind of we were talking about this before we started as well.</p> <p>you you left one of the probably the most challenging fields in terms of just nursing in general, but also ICU nursing is probably one of the more clinically challenging fields and you were cardiothoracic ICU. So, as you were saying, even further specialized. I know as a pre-med we always talk about, you know, the best nursing to do as a premed is ICU because you learn so much [laughter] uh probably more than you do in in uh in any other way from what I understand.</p> <p>Um, and then you do travel nursing, which is also very, very challenging. And we were talking about that beforehand, how it makes it so that you're essentially having to be the cream of the crop as a travel nurse. And then you go into this completely different kind of side of things over the next few years after that, and you become really skilled at business and really, you know, really skilled at management and really skilled at all these things. My big question is why Inflow?</p> <p>because we've talked to a lot of guests who have great skills like yourself and they go on to do all kinds of different things and I always ask them why did you end up choosing this? What was important about this personally to you? >> Yeah. Um and it is a really personal story. One of our colleagues and friends at our previous uh AI company and the reason I say we is there's several people that I brought with me into this company.</p> <p>Um, so we all knew her and she went to the ER one day, had severe abdominal pain, ended up having to have imaging, CT of her chest and abdomen trying to figure out what was going on. They realized, oh, she has an acute appendicitis. It's ruptured. She has to go to the operating room immediately. So, they took care of all of that. And this was a great health system.</p> <p>I won't mention the name, but on that CT of the chest, the radiologist also called out, hey, she has a significant breast lesion and needs ASAP followup and it's suspicious for malignancy and all of the things. So, sometimes that's the hardest part is the detection, right? making sure that you saw this thing even though it was unrelated to the abdominal pain.</p> <p>They saw this thing that was very concerning and that needed followup and it was detected but then there was nothing in place at the hospital. No process, no automated tech, no tracking system. And so it would have taken a heroic effort on the ha on behalf of the clinical teams which is a lot of times what we expect out of our clinicians. We expect them to make a heroic effort to get work done every single day. And in this particular case, she just completely fell through the cracks.</p> <p>You know, they did her surgery, they got her emergency taken care of, they got her discharge, and they took care of that part of her, but holistically, they did not take care of her. And so, 10 months she went not knowing that she had breast cancer. Um, delay to treatment and diagnosis. She just happened to have her regularly scheduled mammogram that year. And they of course found the lesion and the mass again.</p> <p>And they were like, "Why didn't you ever follow up on this?" And of course, she was like, "I don't even know what you're talking about." And when they did a PET scan, they realized not only did she have that mass now, but it had metastasized and now it was a inoperable brain tumor. >> So, this was like really devastating for our team, for Jill and her family. >> And um we all kind of just sat around. I remember the day and we were like, does this happen? Does this happen a lot?</p> <p>Like, is this something that happens in the health system? And of course, coming from the health system, I know how chaotic it can be. In fact, like I have my own podcast called Success in Chaos because we we work in chaos every day in the complexity of the health system. But I think the the fact of the matter once we started looking into the data was that 50% 50% Cole of radiology follow-ups that are in the report get missed. And it's because there's this lack of communication.</p> <p>There's can you get it from the report to a worklist? Can you get it from the worklist to the right physician? Can you get it to the patient? Can you track all of this? And right now we're just trying to use humans for all of that. So you can imagine how that's going. >> Not firstly. So sorry to hear about that experience. Um that's really difficult to deal with. I I personally have not had to deal with that myself.</p> <p>But one thing that I found in speaking with founders and and executives like yourself is some of the most successful ones who go on to form companies have this background where their friend or their family member died or faced some severe repercussions due to this inadequacy. And what you're make what you're saying makes makes sense, right? We're putting AI into radiology. We're putting AI into diagnosis into the chart so that we can track things better.</p> <p>You know, I think back to my time working on stuff with with, you know, Colag Guard and the fit kits. We really tried to get patients to do it. We worked really hard on that to make sure that um they completed the kit and it got sent in. But did we ever really work on the follow-up piece? Uh and the answer is no a lot of times, right? And same thing with the radiology. We can detect all this stuff with AI.</p> <p>We can enhance the diagnostics, but if that isn't translating into care, uh that's doing more harm than good in some ways, right? And so my question to you is um how did you come up with the solution for that? Why hasn't it been solved yet?</p> <p>>> And you just touched on one of my greatest pet peeves is I feel like as AI vendors, technology companies, leaders, it's our job to not just highlight another problem and dump more data in the hospital's lap, but it is our job to automate that solution. And it irritates me. you just mentioned radiology.</p> <p>Um, I was sitting at a at a health system leadership table the other day and they basically said like we feel like Inflow Health is the lynch pin for everything that we need and we can't turn on any of this detection AI detection software until we get the workflow piece in place because all we do by turning this on is increase our own liability, increase the patient safety risk, increase the staff half burden.</p> <p>And so they were like, "We need inflow first and then we need to go back and turn on all of this stuff." Whereas a lot of companies are out there turning all of this on and the data is just sitting there in piles with nobody to work on it. And you're right, it's just creating more work and more noise and it's not helping us give better care. >> Yeah. >> Okay. >> How did you No, no, no. That's that was exactly what I was saying though, right?</p> <p>is that's and there's all kinds of examples you can give. You can give it even in simple examples like you know turning on co-pilot without educating your um staff on how to use co-pilot results in things like email chains that are just co-pilot talking to each other going around and that happened at the last organization I was with. Um >> that's funny. >> Yeah, it's hilarious. Right.</p> <p>And and even now, you know, my my role that I'm in now at Penn um was made because, you know, they bought the software, they implemented it, but there was only one analyst doing the work. And so they said, you know, we need to create another analyst and get this team really rolling. And that's how I came in. So these these things are happening and it seems to go against kind of what we we talk about with high reliability, right? And in what it means to be a high reliability organization.</p> <p>How did you figure that out though? Because I feel like this would have been, you know, in the age that we're in, this should have been solved, it feels like years ago. And I know part of it's a necessity thing, but how did you figure it out? How did you after this awful diagnosis and situation had unraveled? What's the time between that and then inflow kind of coming about? What did that timeline look like? >> Um that was two years.</p> <p>Um the if you look back like there was already people that were trying to solve this but they thought that the value was going to be in the AI piece. They were like, "If we can just get it from the radiology report to a worklist, that's going to solve the problem." >> And if you're part of an HRO, >> you're like me. I obsess about the problem. I obsess about every failure point in the problem.</p> <p>So when we were coming to market, I was like, "Oh, no, no, no, no, no." Like that is a cool thing and everybody wants to sell AI and say that they do AI. And so everybody thought that that's where the value was.</p> <p>But in actuality, that was just a tool or a mechanism to get to the real valuable part, which was we needed to get it from that worklist, automate it to the right provider, make it click to order, track it to scheduling, track it to a follow-up, make sure that the patient was aware, educate them via text, guide them through the health system, make sure they knew the next step.</p> <p>And when we really looked out there on a market analysis, like we weren't the first to market, it's it's usually better not to be. we were able to do our market research to say, "Oh goodness, everybody's kind of missed the point here. >> The point is not getting it to a worklist." So then humans can sit there and dial for dollars and figure out how in the world am I going to contact all of these providers and patients.</p> <p>The idea was, could you solve most of this and collapse and close most of these follow-ups within automation? And the answer was yes. like 80 to 90% of these can be captured within automation. And then what are you doing? This is the best part. You're promoting the human to the top of the pyramid. So it's like AI at the bottom, automation second.</p> <p>And now we're using our clinicians and their knowledge and their wisdom and all of their like nuanced information in their heads where they really need to be, which is at the top as the orchestrators of the AI and the automation. so that it brings joy back into their day and they're not just sitting there dealing with a spreadsheet and a phone all day. >> Has that do you think made it a lot easier?</p> <p>And this is kind of a side question, but do you think that's made it a lot easier when marketing or having these meetings with clinicians coming from that perspective of we've made your life as a clinician actually better and we're going to give you more enjoyable work to do instead of just more work? Like has that been easier in those meetings having that perspective? >> Oh gosh.</p> <p>I mean, we have clinicians and care navigators that get tears in their eyes like because they're like, "Wow, th this just takes hours of work that I do every week and makes it into a click and like them getting back to what they truly wanted to do in the beginning, which was interacting with patients that need them and solving care gaps and doing the more complex things in medicine. It allows them to do that.</p> <p>I don't need you fishing or transferring information from one spreadsheet to another or fishing for information all day. Like that part, let me solve. What I need is you to pick up the phone to call this cancer journey patient and talk them through a very difficult diagnosis and make sure that they know that the health system is there for them, what their next step is, that they're there to support them. That's what they need. They don't I don't need them sitting on in a spreadsheet all day.</p> <p>And that's that's kind of what this current situation is, right? I mean, there are some AI tools in Epic and Athena and Cerner um that do help kind of uh surface insights. Uh there's also dashboards that have been developed so that practice managers can understand what um you know outlying or things that haven't been resolved yet are right.</p> <p>So there are some tools that exist, but you're kind of what you're saying is it's very rudimentary still and the AI tools kind of help surface some of the more important things, but they don't really do anything else. What does Inflow do? How does that differ from the existing Epic and Cerner and Athena kind of equipment that's that's off the shelf already? >> Yeah, I think a lot of people really focused, as you can imagine, on the things that were already regulated.</p> <p>So like breast and mammo are the most regulated for the longest period of time. they had national guidelines, things like that. So that they were doing an okay job with. And then the next tools came around lung when it was shown that if we could catch lung nodules much sooner, we could reduce the mortality rate. And lung nodules typically have no symptomology in the beginning. And so it was you have to get screened for them. So the lung cancer screen program started to develop.</p> <p>But if you just take breast and lung out of all follow-ups, you're literally only talking 15 to 20% of follow-ups. So what are you doing with the other 80%. And that's when we came to market, we were like, cool that like Epic and Athena and Cerner are like putting these keywords from a worklist or keywords from a report on a worklist. But nine times out of 10, like it's called a worklist for a reason, right? it's creating this work list that nobody has time to do.</p> <p>And we were like, man, it's not the worklist that's the point. And also, you're missing 80% of your follow-ups over here by just focusing on these two programs. And so what we realized is the reason that they were doing that is they had no staff. They only had like one lung navigator or one breast navigator. We were like, "Okay, so how do we help them do this without staff?</p> <p>and how does the how do we automate the majority of this so that they don't have to spend all of these FTEEs supporting then this new AI detection tool. So we built our own language model um that just reads and interprets radiology reports >> and we did a top- down approach where we were looking for all follow-ups and then we were able to filter and subcategorize those by different program types.</p> <p>So, we became the first platform to market that was really doing all follow-ups in one platform across all modalities in radiology. And then you could start to sort them into all of your different programs based on type. So, lung over here, breast over here, actionable findings, incidental, explicit, all of the different programs.</p> <p>And then you add the automation piece and it's like, oh, well, I don't actually have to have a lot of I just have to have some quarterbacks or or supervisors to the automation. Um, and then I have to have that one that pyramid at the top, the human in the loop, which is the cases that okay, they're not closing within the automation. How can we help this? Like, what can we do to help this along?</p> <p>And when we figured all of that out, it really like I think we've won six out of six RFPs because nobody really is out there doing what we're doing. >> Why Why choose to build your own LLM? That's that's challenging. You know, we've investigated it a little bit at the journal. we've um you know come up with some alternatives for now that work okay for the current size of our content. Um this is a very different case of course but a lot of organizations choose not to do that.</p> <p>What what made that decision make sense for inflow? >> Yeah and we test it every year trust me. First of all um LLMs can do a fine job at reading a report and telling you if there's a follow-up. In fact their efficiency there is is really quite good. Um, but that's not really what the end goal is.</p> <p>And so when you start to get to the end goal, you really need it to be like, okay, find the followup and then tell me if it's a lung nodule and then tell me if it's a certain size and then tell me if there's a smoking history and then and it's like and then that's where the LLM start to just fall apart and hallucinate.</p> <p>So, we tried obviously to use things that were pre pre-made or generic, but what we noticed is not only financially was it not, you're talking about health systems that have two or three million in study imaging volume. We would have to ping an API, you know, like that's a very expensive. You're running two to three million reports and processing and that's an expensive >> um endeavor. So, what we found and this was I can't take credit for this.</p> <p>This was my tech and data scientist, um, Nate. He was like, if I build all of these like mini models and I build our own language structure, it's not going to be considered a large language model, but it will be considered a radiology specific language model.</p> <p>He was like, "This is we can move really fast and these mini models take up such low consumption space and it really it allows us to not be replicatable as quickly by other companies like with freaking five coding or whatever you want to whatever you want to toss out the new term of the week.</p> <p>Well, and you have so there's also of course it adds value intellectual property that you have and it you know it adds a lot of value when you're coming to to an organization to run their system for them because they know it's something that you really understand how it works which a lot of these companies because they're using a third party API they don't actually know how their own software works. So I think that's a big bonus.</p> <p>My question on that though is there is a kind of like a 100% success rate right or follow-up rate. um how does how does that work? How does the AI get so successful, right, without without running into challenges or issues? How have you avoided those hallucinations even with the small model? And I know like you you came from Duke, so Duke's a high reliability organization. Uh Penn is a high ability organization.</p> <p>So I know we kind of both speak the same language when it comes to that, but how did you figure that out? I mean, it really comes down to, if you talk to anybody that's really built out something that's successfully being used in a healthc care use case, it comes down to medical labeling and these ontologies. And that's really where the majority of our intellectual property is.</p> <p>Like somebody could come to market tomorrow with an LLM that could read through a report and tell you if there's a follow-up, but the problem is is they won't be able to solve all of these like split-off scenarios that have to happen to make clinical decision- making or support. >> And so that really is the majority of hours and hours and hours of clinician time where we had two different radiologists reviewing constantly. Okay, we said this was positive. We said it was positive for long.</p> <p>We said it was positive for a long incidental like we're rewrite like you just constantly have to go through this iteration over and over again millions and millions of radiology reports and then you get to the point where you're like okay I trust it but then you take a new health system live their data looks slightly different and you have to go through this iteration process again right and so we during our implementations go through an iterative process we go through with the team we have them do a manual pull of all lung nodules tools.</p> <p>We do a automated poll and then we do a comparison to make sure that we're working at the same performance at every single health system that we were on our training set. >> Wow. So this is not a on your end this is not like you're just a software as a service vendor where you're like you know subscribe to our stuff see you never again. You guys are like really working with these organizations as well. >> Oh my gosh. Yes.</p> <p>In fact, we developed a partnership with the American College of Radiology because they had this program called Empower and we raised our hand and we were like, "Hey, I know we're a vendor, but this is the world that we live in." And the ACR program basically helps health systems solve this very dilemma of follow-ups, follow-ups around incidentals, follow-ups around lung nodules. and they were so generous to be like, "Okay, you can come to the program.</p> <p>Bring one of your health systems." So, we did. So, we were one of the first vendors to sit at the table. And it was so incredible to watch. I think there was like six different health systems in this cohort. They were all struggling with solving this problem. They all had the same failure points and the ACR was guiding them through this A3 methodology.</p> <p>And we got to sit there at the table and every week listen to health systems just complain about everything that was wrong or broken within the follow-up management system. And we were like, "Oh, we can fix that. We can fix that. We can fix that." But also, what it made us realize, Cole, is that technology alone is never going to be able to fix everything in a health system. It has to be like people, process, and then it's all layered on tech that can do the automation.</p> <p>And so through that program now we pay for that program for every single client. We really deal with their people, process and tech during our implementation because in order to to create these high reliability management programs. We really do we can't just be like oh here's your easy button like right >> that's not going to work. But we really do help them structure their entire team. Bring everybody that needs to be at the table to the table for decision- making.</p> <p>Figure out every single failure point in their process today workflow and roadmap everything that we can fix as a technology. And then we put everything over here in the column of hey we're not going to be able to fix your access to a CT scan. So that's on you, right? What we are going to be tell telling you is how much volume are we going to be pushing towards CT so that you can solve that in advance of us pushing all that volume to CT and you can figure out how to get the access.</p> <p>So it's this constant balance between like what do we solve and what where do we sit and what do you solve and it also teaches them to iterate. So like in the future they might need the technology to do something different. How do they come to us work with us figure out how to iterate? That is really what that program teaches. and we absolutely love it. Our clients have had such phenomenal success doing our implementation in combination with the ACR program.</p> <p>Uh yeah, that that so I just had an episode with uh Seammens and then we had one with J as well and we're kind of discussing these things and it one of the things that stuck out to me that they're really both kind of competing on is maximizing that relationship so that when problems come up or when changes come up in the future, you go to kind of your vendor, right? Instead of going to a consultant.</p> <p>And this sounds almost like a kind of impossible version of that, you know, for somebody like GE, such a huge company, but Inflow is not as big, of course. And so, are you able personally as the CEO, are you the one who's involved in these things, meeting the clients, and you are doing this? >> Um, >> how does that work? >> I actually love that part, but I just I don't have a lot of time to do that part.</p> <p>Um, I do take place still in our executive like year-over-year reviews with the clients to talk. Oh, >> you have annual as well. There's an annual review. >> Yeah. So, in the beginning it's actually weekly and then it moves to quarterly and then it moves to yearly and so our clinical teams are involved with the clinical teams at the hospital like from now until forever the contract runs. >> And then what is a clinical team?</p> <p>just, you know, as we're kind of closing the conversation, what is a clinical team? What's the definition of that for inflow? >> So, it's made up of all different sorts of clinicians. We have physicians assistants, we have nurses, we have a physical therapist, we have radiologists. So, we have kind of like from every specialty coming together and creating. We also have project management experts and people who are high reliability experts and they're all on our clinical implementation team.</p> <p>Wow. So you have a kind of an organization which has these clinical teams that are doing this you know weekly monthly frequent review and then on the even on the executive level you have this great clinical background and you are participating in these reviews on this this annual basis and so it feels like things are very well maintained. It feels like if there's a variance or a you know a change you are ready to jump on it and that kind of introduces me to my final question. What is next?</p> <p>You are getting feedback from the clinical teams. You're getting feedback from your partners. Of course you are a experienced clinician. What are you seeing on the horizon as you're getting all this information? What's next for kind of healthcare AI and followup? Yeah, I mean I think you probably already know the answer. Aenic AI is the future as far as being able to reduce the amount of FTEEs that support a certain workflow. So we will be adding that.</p> <p>I think we have our first beta site going live late this summer. Um and they'll be testing all of our workflow and use cases. It'll both be birectional texting chat with bots and also um calls with bots and agents. And so training that using the same style that we've learned and making sure that we're giving it all of the clinical relevance to do that workflow. There's lots of people that have agents but they haven't necessarily been trained for the health care setting.</p> <p>And so making sure that we're ready to support that. Um, with the big beautiful bill passing, hospitals are going to be looking for ways to reduce spend, reduce FTE support, but still maintain creative revenue sources as they see some of their losses coming into the future into 2028.</p> <p>So we can be part of that solution where we're not just driving utilization towards radiology, but we're doing it in a really fiscally responsible way through automation agents and not having to add a ton of support staff to their team to really solve a huge patient safety care gap. >> Um, easier since you've already built the foundation of having a a really good AI system.</p> <p>So when you add the agents to it, is it easier than some of of what other things have faced or is it still very challenging? >> Nothing's ever as easy as it sounds, Cole. But, um, I love when I say that to my technical team, like, "Oh, this should be easy." And they're like, "Sure, Angela, it's going to be easy." [laughter] Um, it's it's it's not that it's easy.</p> <p>It all takes again being obsessed with the problem, being obsessed with each failure point, being obsessed with each workflow, and understanding that like your bot has to be able to have really sensitive conversations about a PET scan that has maybe the opportunity that the patient has cancer. Like teaching a bot to have that level of conversation, you can see where that could be a little nuanced, like what is a PET scan? I'm scared of that. Why do I have to have that?</p> <p>and having that conversation while you're talking to an AI agent. Like we want that to be as humanistic as possible. >> So that's where it's not easy to have a if I'm just calling you to be like, "Hey, you're ready for to pick up your prescription." That's easy, right? But to have a nuanced conversation about an imaging that you had imaging that has a followup and then the patient's worried like, "Do I have cancer?" That's a little more nuanced than that.</p> <p>I appreciate that answer though because it shows that you are very in-depthly aware of what the challenges your technical team is facing which I do not think is always the case with a lot of executives. No offense to executives, they're very busy and I appreciate how busy they are but it's kind of refreshing to hear that in a way. So I appreciate that very much.</p> <p>>> Uh that we we will be successful on the backs of them and so I have to uh definitely understand what they're going through every day. >> Yeah. And then the last part is we now that we've solved for radiology, our clients and our prospects are asking us to expand into other service lines. So lab, pathology, cardiology. We just went live with cardiology at one of our health systems.</p> <p>So that one's live and operational and we'll start to expand into some of these other service lines with the same process that we've built for radiology. >> Impressive. So we will be definitely having you back on in the future then to discuss. [laughter] >> Congratulations on your success. Uh, thank you for your time and we do hope to have you back on again. >> Thank you so much, Cole.</p>
Want to reach healthcare executives and decision-makers? Join industry leaders like HealthMap Solutions on our podcast.
Become a Guest