Latest / HumanifAI HealthTech / Fixing Healthcare’s Referral Black Hole w/ Trey Holterman, Co-Founder & CEO, Tennr
Transcript
- 0:00Brian: Every day in healthcare, a referral goes out and a patient disappears into the fax, into the queue, into someone's pile.
- 0:09Brian: The specialist then never calls back.
- 0:11Brian: The primary care doctor never hears what happened, and the patient is left in limbo wondering if they've been forgotten.
- 0:19Brian: Trey Holterman watched his mom, a family physician, live with this for years, so he built a company and even a proprietary AI model to fix it.
- 0:36Brian: Welcome back to Humanify Health Tech, where we explore how AI is rewriting the rules of healthcare and putting people back at the center where they belong.
- 0:45Brian: I'm Brian Chee, senior director of content and communications at Net Health, and I'll be filling in for our host, Francesca Vereb.
- 0:53Brian: I have big shoes to fill, so I'm treating this like value-based care, aiming to deliver good outcomes but with limited resources.
- 1:02Brian: Speaking of outcomes, today's guest built one of healthcare AI's fastest-growing companies around a problem everyone in and out of this industry has
- 1:12experienced: the referral that goes nowhere, or as our guest likes to put it, the patient referral black hole.
- 1:20experienced: I'm excited to be joined by Tray Holtermann, co-founder and CEO of Tennr, an AI platform that automates the referral intake process so patients don't fall through the cracks.
- 1:31experienced: At the core of it is Ray Elum, Tennr's own AI model trained specifically on healthcare documents.
- 1:38experienced: Tray is a Stanford graduate, one of Forbes top founders in AI, and has raised $162 million in under two years.
- 1:46experienced: Tennr now processes 10 million healthcare documents a month and works with hundreds of organizations across the country.
- 1:54experienced: And I have heard that you process 10% of all Americans.
- 2:00experienced: Tray, is that true?
- 2:02Trey: Yeah, that's right.
- 2:03Trey: And on the sending side, uh, we're in three hundred fourteen thousand providers now across the country and, uh, 89.7% of counties.
- 2:09Trey: So hoping to get to, uh, 90% before the weekend.
- 2:13Trey: But, uh- Got it.
- 2:14Trey: That's amazing … push the needle on that.
- 2:16Brian: All right.
- 2:17Brian: So before we get into Tennr, I wanna talk to you about you.
- 2:21Brian: Stand-up comic, captain of the Stanford crew, founder, engineer- Okay.
- 2:30Brian: Make that make sense for me.
- 2:31Brian: Talk to me about how such a diverse background has informed what you do today as a leader.
- 2:38Trey: Um, well, uh, man, I app- I appreciate it.
- 2:42Trey: I, I would say, um, it's like, um, in Finding Nemo when, uh, you know, they go to the dad Marlin, and they're like, "Oh, you're a clownfish.
- 2:49Trey: Say something funny." I… You know, when you tell somebody you did, uh, years, a couple years of standup and a couple years of improv before that, um, to be honest, it's more just a, an, an obsession that was bore out of some really harsh feedback I got from somebody that said, uh… It was actually a manager when I was an intern, and she said, "Look, if you wanna speak in front of the company again, you better take some, uh, public speaking lessons." And, uh, and so I was just researching, like, what's the best way to get really snappy and better at public speaking, and I tried to take it to its logical conclusion, was just open mics and open mic and open mic, and you start getting addicted to the rejection of that, and you start doing it a lot.
- 3:27Trey: And so, and then honestly, similarly in rowing, um, and, and making that, you know, a big part of my life, I was frankly, uh, just a little bit of a overweight kid, and my mom was the most healthy, you know, person you can imagine.
- 3:39Trey: I used to like to joke that she just didn't wanna take her work home with her.
- 3:42Trey: And, uh, so I was, uh, you know, probably 14, and I was frankly just trying to get really, really fit.
- 3:48Trey: And, um, you know, so I d- tried to take rowing to its logical conclusion and be the best that I could be at, um, at every stage.
- 3:55Trey: So I know, uh, you know, obviously those sort, those two activities seem like sort of paradigmatically opposed.
- 4:01Trey: One is, like, indicated by your ability to stare at a machine or, you know, look into the, uh, the, the vast, like, land of water for a, a long period of time, and one is, you know, kind of about trying to be quick-witted.
- 4:14Trey: But to be honest, I think both of them are just practices that are really painful that you will inevitably get good at if you spend years doing.
- 4:22Trey: Um, and I just sort of had the, uh, the frustration with myself before either of them that led me to wanting to do it.
- 4:28Trey: Um, and that's probably the same mentality going into this company.
- 4:32Trey: By the way, I don't do any standup today, and frankly, I probably don't work out more than 20 minutes a day.
- 4:36Trey: It's all focus all the time on this.
- 4:39Brian: So, so yeah, this, the discipline aspect of it, right?
- 4:41Brian: I mean, th- that's like… I, I noticed also that your co-founders are also, al- you also rowed with them.
- 4:46Brian: Is that, that discipline, sort of that seeing an area to improve, seeing something to fix, seeing a solution to, to come up with, and having the discipline to follow through, is that something that, like, sort of formed through your rowing and through your experience?
- 5:01Trey: You know, one of the things I always ask new employees is just,"What's the best team that you've ever been a part of?" Or really candidates that we're late stage with.
- 5:08Trey: Um, because I don't think if you've been on a great team before, uh, and you know what that feels like, you're, you have the-- it's, it's really difficult to learn the capacity of being a great teammate.
- 5:17Trey: And with, you know, Diego, my co-founder, who was a rower, um, I just seen the guy work so hard and struggle through so much.
- 5:23Trey: Most founders quit.
- 5:25Trey: Um, you know, you mentioned, right, we'd raised all that money in two years.
- 5:27Trey: We've been doing this for five and a half, almost six years now.
- 5:29Trey: So, um, the vast majority of this has just been like pain and struggle, trying to convince people that this was a problem that we could solve and that it was worth solving, and it was really only in the last couple of years that we've been able to show that momentum to the world.
- 5:43Trey: And so Diego is just sort of like my ride or die because, uh, we, we'd had that.
- 5:47Trey: Tyler, on the other hand, is just an absolute genius.
- 5:50Trey: Um, I was a TA in the computer science department, and there was just legend of Tyler Johnson.
- 5:55Trey: Um, and, uh, I was never, you know, quite as talented as him from an engineering perspective, but I could see his grades.
- 6:02Trey: Um, and, uh, he was, uh, and he was just somebody that I kind of attached myself to, and I watched what he could do from a project-based perspective and just how he showed up as a, as a team teammate.
- 6:12Trey: Um, he, he I think he was like the president of his fraternity and, um, and so I just knew he, he knew what it was like to kind of rally, uh, people up.
- 6:20Trey: So that was sort of the unifying thing was just three, three, um, you know, not always friends, honestly, um, but we just knew none of us would quit on each other, and we would, uh, we would still be doing this today probably even if none of the, you know, statistics were looking good.
- 6:34Brian: Awesome.
- 6:35Brian: All right.
- 6:35Brian: Let's talk about Black Holes.
- 6:37Brian: Uh, y-y-y- I said in the intro that y- your mom is a family physician, um, would complain to you, right?
- 6:43Brian: She would say, uh, she would describe or com- talk about a process that felt fundamentally broken.
- 6:48Brian: The patients go out into a referral system, and that no one knows what happens to them after that.
- 6:53Brian: They just disappear.
- 6:56Brian: What was she describing when she talked about that exactly?
- 7:00Brian: And, and at what point did that frustration become a problem you felt you could, you could solve?
- 7:06Trey: Totally.
- 7:07Trey: You know, one, um, slight, you know, detail, uh, is she was a, uh, she's a physician assistant, and so what I think, why, why I think that's actually very, like, lucky was, you know, she really has the perspective of what happens, uh, all the annoying work that often gets like, you know, dropped onto MAs or PAs or nurses that have to do some of this really gritty stuff to get a patient, you know, through the other side.
- 7:30Trey: And then she also, you know, stopped working for a long time to raise five kids.
- 7:33Trey: And so when she went back into the workforce, um, when I was in college, it was-- I think she had this just really, like, logical, had watched Silicon Valley develop, um, out in the Bay Area for 18 years, and now it was like, "Whoa, this is really how it all still works." And, um, and so she just had a tremendous perspective that, like, this can't be-- this can't make sense when I am sending patients that, who to businesses whose business depends on processing those patients effectively and, uh, you know, with great care, and they're dropping the ball.
- 8:08Trey: Um, so it just didn't click for her, and then, um, it was actually, like, through some, like, frankly, desperation on my side trying to figure out some, a, a research project to work on that, uh, you know, her and her infinite wisdom kind of just explained the whole thing, uh, to us.
- 8:24Trey: Um, of course, I didn't listen to her, and it wasn't until then Diego got hospitalized for, uh, some stomach issues, and then six weeks later gets a call from the GI, and I was like, "Oh, I know why that happens, why they're so bad at processing those patients." And that was when it sort of put it together and, um, and we really kind of dove in.
- 8:41Trey: But the key was, the, the real, like, fortunate insight from us was we learned it from the sender's perspective and how frustrating it is, but we also understood it's so hard to change the sender's behavior.
- 8:52Trey: If you think about a patient going from point A to point B, point B is the one that makes money off the patient.
- 8:57Trey: Point B is the one that has the incentive to drive that patient through to a service, a procedure, or therapy.
- 9:03Trey: And point A, you need to make their life really easy And I knew that from, you know, just looking at the hassle on the sending side, that we had to solve it for B in a way that was super convenient and nice for, for A, um, if that makes sense.
- 9:18Brian: Yeah.
- 9:18Brian: One of the things that's fascinating about it to me is that you sort of, like you f- you sort of shifted the paradigm a little bit in what you did, right?
- 9:25Brian: Because in most, I feel it's fair to say that most technologists will come into something like this and say, "Okay, well, the way this, this is happening needs to change.
- 9:34Brian: So get rid of the fax, get rid of the PDFs, get rid of the," what, what do you call it? Unstructured data, right? "And, and let's structure it.
- 9:43Brian: Let's, let's find a system." Yet that's not how people work. And it feels like what you did was you said, "Hey, you know what?
- 9:50Brian: We solve this by working with the f- the providers and how they work and find a solution to that, that meets them there." Is that right?
- 10:00Brian: Is that, is that a, is that a right read?
- 10:02Trey: It's totally spot on.
- 10:03Trey: I mean, honestly, we just, um, like the thing that we tell ourselves a lot is, uh, we look at that… I don't know if you've seen this.
- 10:10Trey: It's a very popular internet meme now where it's like on the, it's the IQ is on the X axis, and on the left side you have the incredible low IQ, you know, person.
- 10:19Trey: On the far right side you have the incredibly high IQ person, and they usually come to the same conclusion about things.
- 10:24Trey: The problem is in this middle IQ zone that the vast majority of us are all in, where you're overthinking things, you're trying to be so sophisticated, you're trying to like, you know, always skate to where the puck is going.
- 10:37Trey: And for us, we just always remind ourselves we need to go to the, to the left side, 'cause we're not, we have no hope of getting over to the, to the super high IQ side.
- 10:45Trey: So let's, let's always try to think about life from the, from the, uh, you know, the simplest terms.
- 10:50Trey: And we always just viewed it as we need to solve this problem, and we never saw any other option to solving the problem that was legitimately better.
- 10:59Trey: Almost everything, other solution that's been purported, that's been tried, people have tried to kill all these systems, tried to make them illegal even at one point, and it was always predicated on changing behavior from somebody that had no incentive to change behavior.
- 11:12Trey: And so we alway- we thought, "Huh, maybe if the person that has no incentive to drive to change their behavior, maybe we should make this work without them changing their behavior.
- 11:22Trey: And, or, and maybe long term we can, like, get them there.
- 11:25Trey: Right now we do actually have some really good products for those senders to see patients, update information, whatever.
- 11:30Trey: But even then, it's like it has to be linked into the EMR.
- 11:32Trey: Like, it's not, let's not change things, um, and let's just have this, like, really kind of ruthless, simplistic mindset to what reality actually is.
- 11:41Trey: And I feel like they're, the, uh, most of, of health tech that I've observed, um, that, that fails, it's never for a lack of intelligence, never for a lack of grit.
- 11:50Trey: It's almost always for folks trying to, um, you know, invent the future as opposed to ride like what is, what is, where is, where are things steadily going, but how do you just solve the problem today, right?
- 12:02Brian: Yeah.
- 12:02Brian: You said, so I, I heard you talk about, um, I, uh, you- I heard you talk about a, a scenario in which you saw, um, technology failing because it couldn't pass the test the providers were giving it.
- 12:18Brian: Like, so, you know, AI in this case, you know, we're talking, uh, we're talking about AI.
- 12:22Brian: You know, in, in the, in the case of Tennr, you would go to a provider, you would go to a practice, and the person who was doing the referrals was using paper, uh, a paper process that was actually faster and more efficient than what the digital version that they had was, right?
- 12:40Brian: And they, and, and I think you talked about how that was an eye-opening experience for you, that, that this person who was doing this for 20 years was more efficient and faster than, than the newfangled thing that they were, that, that they were trying to use but wasn't working.
- 12:56Brian: And, and how do you get AI, in this case your proprietary, your proprietary AI, to actually, I guess, replicate what they were doing with their, their, their more traditional approach?
- 13:13Trey: You know, there's a lesson that you learn in, like, uh, in, like, ba-- you know, the intro to product design.
- 13:20Trey: I didn't take it that far, but I re- you know, it's like the… And it's this accepted idea in, in Silicon Valley, which is you're not going to change somebody's workflow until the solution that you provide is at least 10 times better, like truly 10 times better.
- 13:32Trey: And when, uh, kind of cloud technology and really just the digitization movement came to healthte- healthcare, it did not usually 10x the experience of the users doing the work.
- 13:44Trey: It 10x the experience of interoperability for claims submission, it 10x the ability, uh, you know, to have one canonical patient object, but it almost never actually 10x the ability for somebody to do their work.
- 13:59Trey: I would argue, right, in many ways, it's actually, uh, five years ago, it was faster to just, like, jot things down in a notepad, talking to a patient, looking at them, taking little scribbles down, and then file it away than it is to do 37 clicks to get the patient's information and send orders and all that stuff.
- 14:16Trey: Of course, they, you know, everybody had to force that workflow change onto people to get the digitization, but the end user experience wasn't actually that good.
- 14:25Trey: So then you ask, "Well, why wasn't it actually that good?" Because healthcare is very similar to legal, where it's very unstructured by nature because it all starts from the non-deterministic, uh, determination of a, uh, of, you know, a very, very skilled profession.
- 14:41Trey: There's narratives.
- 14:42Trey: And then, by the way, on the payer side, just like how lawyers have these narratives and, and, you know, their own diagnoses as to a situation, there's also the law.
- 14:50Trey: You know, in medical, you have the payer criteria, which is sort of the law, right, of what you can do and what you can't do.
- 14:55Trey: And then also there's a separate kind of law, which is clinical decisioning and triage.
- 14:59Trey: And so you have unstructured determinations going up against unstructured guidelines in both legal and healthcare.
- 15:06Trey: And there was this, obviously this incredible kind of moment where, um, the context of what has happened into a pa- what has happened with a patient and the context of what do payers need to see and what is the, you know, guidelines of what our business wants to, you know, put the journey that this, that our business thinks is best to put this patient on, were both unstructured.
- 15:30Trey: And if language models are good at doing one thing, just one thing, it's taking unstructured context and applying it into an unstructured situation so that work can continue.
- 15:39Trey: And that very simple paradigm of what language models have given us is what allows us to build a 10x experience.
- 15:46Trey: Everything else, all the other stuff that we do, which is 90% of the work that we do, is structured workflow, but we're able to do it without forcing an end user to do an experience that's worse than notepads.
- 15:57Trey: It's so much more elegant.
- 15:58Trey: It's so much more streamlined, all because we're able to deal with that unstructured, um, information that way.
- 16:05Brian: So what convinced you that the general purpose, quote-unquote,"large language models" wasn't gonna cut it for what you needed to do?
- 16:15Trey: Well, to be totally blunt, uh, it could, and you can if you do like, uh, if you, um, if you know how to orchestrate them in the right moments of the workflow, you could absolutely use an Opus four-seven or a five-five and kinda do that moment.
- 16:32Trey: The reason we run and tune our own models is because we operate at a scale now where if you don't control, if you don't control it down to a very, very fine-tooth, uh, level of detail, you have no control over your costs.
- 16:47Trey: And so in many cases, doing this work with models running tokens would actually be more expensive than doing the work manually.
- 16:53Trey: And so you have to control your costs so that we can pass that on to our users.
- 16:57Trey: We had a funny estimate when we were kinda getting started that we would burn through our Series A was like eighteen million dollars, and that would've been our monthly burn rate just on model consumption at the time in '23 if we didn't run our own inference and control our model.
- 17:12Trey: And then number two is control in the context of accuracy.
- 17:16Trey: When we look at our, you know, model to be ninety-eight percent accurate on a, you know, very complex determination against a very complex payer criteria guideline with very complex medical record history, and you can see a generalizable model that does that at ninety-two, ninety-three percent accuracy, the reality is, like, you're gonna have a great 10X experience with the ninety-two, ninety-three percent accuracy.
- 17:37Trey: But when you're processing hundreds of thousands of, of patient encounters a day, five percent is the difference between, you know, a support team needing to be, you know, ten to twenty times larger.
- 17:46Trey: It's also incredibly bad experiences for patients.
- 17:50Trey: So I wanna be clear that you can actually, but it's all about control, and it's all about making sure you understand how to, um, you know, deliver those models in a, into a workflow, into a system that drives an outcome that solves the problem.
- 18:03Brian: One of the things that we've seen is, you know, o- one of the decisions that we made at Net Health was to go into a native A- to, to build a native, uh, AI into our, into our, um, systems because we found that if we didn't do that, the specialized area that we w- that we're expert in would not actually work as well.
- 18:24Brian: So, like, that came home to me w- talking to a client the other day when-- because he was saying to me essentially, "Hey, you know what?
- 18:31Brian: I used it for the first time, and I realized immediately that this is actually, this ambient documentation in this example is going to actually save me time." I-- And he literally walked me through like, "Hey, going from the computer to the, to the, um, to the, to the computer by-- I don't have to do that four times.
- 18:53Brian: I actually can just stay in one place and talk to the patient, and that actually makes a fifteen-minute difference in the evaluation that I'm doing.
- 19:02Brian: And that actually makes me more efficient, and I actually generate more revenue as a result of that." So when you're talking about the, that aspect of it, what- Where does that play into this idea where you-- where, where, where our clients, your clients have this hesitation around demos because they see the demo and it looks like it works great, but then it doesn't really work great.
- 19:32Brian: So how do, how do you cross that bridge to get to the point with that guy where he's like, "Oh my God, this actually worked right. This actually did what I needed it to do, unlike the other demo that I saw that was just a bunch of flash." You know, how do you get there from, from A to B?
- 19:48Trey: I mean, I think you have probably a very unique moment in time where there has never been more borderline fraud, slop, fake vaporware being created because it has never been easier to create the surface tip of the iceberg visual representation that you can do interesting things.
- 20:09Trey: Um, it's never been easier to get to that point.
- 20:13Trey: And yet simultaneously, and you can look at our, you know, payroll and how we have to continue to invest engineers, even though we are like a very, very AI native and forward operation, to build these products that work at scale with systems is as difficult as it has ever been.
- 20:28Trey: There's just higher expectations for-- and you have to deal with more config and more customization and more whatever.
- 20:34Trey: Now, how we-- to actually answer your question, how we solve that problem is I will not, we will not work… I won't sign the other end of a contract with a customer if they have not seen a, a actual like Tennr in their system operating and doing that work.
- 20:49Trey: Because if, you know, you're in a, in a f- where everybody can demo whatever the hell they wanna see, I'd, I'd frankly, I wouldn't wanna waste our time working with somebody that's only looked at demos bec- uh, they need to see it operating on their system and with their workflow to feel that confidence.
- 21:08Trey: And that-- and frankly, if folks can't do that, it's because, you know, they have that vaporware.
- 21:12Trey: And, um, and what does that mean ultimately?
- 21:14Trey: Does it, does it bring your cost up to, you know, service a client because you're now saying, "Hey, every client I need, I need you to see this live in action on your systems, and you should really test it"?
- 21:23Trey: Of course.
- 21:24Trey: But it's, it's, there's a, there should be a higher barrier to trust in this, in this kind of new world
- 21:30Brian: One of the funny things I think is that, that the work that you do is similar to the work that we do in that we both work with clients that are kind of skeptical about technology.
- 21:40Brian: They're actually probably fair to say they don't like it.
- 21:42Brian: They're really-- like, they look at technology and say, "That's taking me away from my work. That's taking me away from what I spent 12 years in school doing, and, and I, and I worked so hard, and I, and I, and I have a passion for." You know, for us, it's the rehab therapists, it's the wound care nurses, you know, it, it's the clinicians.
- 22:01Brian: Um, that skepticism around technology, evangelizing that to that group is a special thing.
- 22:07Brian: What have you seen in terms of what you've done that has just sort of created a light bulb moment for your clients where they've said, "Oh, wow, I can actually do what I love.
- 22:19Brian: I can actually create moments that matter to me and to my patients because of this, because of Tennr, because of what we're
- 22:27Trey: doing here"?
- 22:29Trey: It goes back to that experience you described of the person that used the ambient listener and, right, they were promised all this beautiful stuff of digitization.
- 22:38Trey: They were promised that, oh, it was gonna make life so much easier.
- 22:41Trey: And what did it do?
- 22:42Trey: They fe- spent way more time filling out more structured fields.
- 22:46Trey: When you have a system that can deal with unstructured narrative, in that case of an ambient listener, which, I mean, I'm just a hu- I'm just the, the biggest fan of these ambient listener companies.
- 22:56Trey: Um, and you are giving what the promise of digitization was for that business years ago now.
- 23:05Trey: Like, it's finally actually delivering on what these folks thought was gonna happen 10 to 15 years ago.
- 23:10Trey: Honestly, if I worked in healthcare, I would hate technology.
- 23:12Trey: I mean, goodness gracious, I've seen the, the, the tools that the health tech world has brought into healthcare, you know, and it's like, it's bad.
- 23:20Trey: It's so bad, Brian.
- 23:22Trey: And, and there's, like, a new wave of great technology, uh, that you see out there, like Abridge makes good products.
- 23:27Trey: You see Ambient's good products.
- 23:28Trey: Like, there's good products.
- 23:29Trey: Um, but the old… Like, some of the old stuff is like, really?
- 23:33Trey: That… And, and you know what it comes down to, Brian, honestly, is because there's so many, um, there were so many weird warped incentives.
- 23:41Trey: There were these strange government plans trying to drive adoption.
- 23:44Trey: You had weird, you know, kickbacks going on.
- 23:47Trey: I mean, you can look at these lawsuits.
- 23:48Trey: And so all these people are using the technology that wasn't actually built, like, properly for them to do their work.
- 23:55Trey: It's just so frustrating and, um, it was… And then you have, you know, uh, just, you have basically every bad dynamic you could possibly imagine, and then you have buyers that don't understand the workflows in many cases.
- 24:09Trey: Um, and that's, that's w- how I think we've ended up in this mess where health tech is just such a, in, in general, like, from the last 10 years, has been such an overpromise and such a squelch in actuality.
- 24:20Trey: I'm very optimistic of, of seeing that, like, the products like ours, and certainly our product, the, the, the most proud board slide I've ever presented and shown was slide one, just all executives of the companies that we've worked with talking about the outcomes that they have achieved at the highest levels, patient conversions increasing, denials dropping dramatically, OpEx being lowered because they don't have so much back and forth.
- 24:43Trey: Then slide two was the people that review the work that Tennr does and just like,"I love it." And they've said, "This has, this has really changed how I do work." They love, uh, the user experience.
- 24:54Trey: We spent so much time on that.
- 24:55Trey: And this is what I think is, like, such a bummer about folks kind of thinking of towards AI, is it's like, it's, it's, um, it really is able to, if the tooling is done right, by the way, it's not always done right, but, like, I, I've certainly seen firsthand that when it's done well with a problem in mind, it creates both incredible business outcomes, and it's, like, just such a less of a drain to use than like the o- you know, the old school, antiquated, horrifyingly slow, form-filled-out, uh, based software that folks are used to.
- 25:29Brian: A lo- a lot of times we say we're, we're actually making good on the promise of what EHRs are supposed to do or what our, what our data is supposed to do, what your data is supposed to do for you because of AI, and that's one of the things that's really inspiring personally for me, is really inspiring, is that in this time and place, um, with the technology that we have, we can actually connect those dots.
- 25:51Brian: We can actually make it happen where those promises actually, uh, make a difference.
- 25:57Brian: That's gotta be super cool for you and your team to see what you're doing on the referral side and the differences that you're making there, right?
- 26:05Trey: 100%.
- 26:06Trey: And like, you know, I piss people off by saying like, uh, you know, I, I actually don't like it when somebody is too excited about the technology, right?
- 26:15Trey: We should be super excited about the problem to be solved, and then we should make sure that we always have the best engineers, uh, you know, that, that we can possibly recruit to go deliver against it.
- 26:26Trey: And if tomorrow, for whatever reason, um, you know, the technology changed shape and we needed to adapt, right, we would adapt that way.
- 26:32Trey: The key is being very problem-obsessed, being very problem-focused on like, what is the problem with getting a patient from point A to point B?
- 26:39Trey: And by the way, just for, for a little bit of nuance, right, the typical, you know, how Tennr grew initially was all through referrals, but then internal orders has been a huge anchor of growth for us.
- 26:50Trey: So a lot of folks don't actually call that a referral, but it's patient came into a consult and they dropped off.
- 26:55Trey: They didn't come back, you know, for the addition- uh, they didn't go and do the additional imaging and then come back for, you know, a pre-op.
- 27:00Trey: Um, that's been a huge growth angle for, for Tennr as well.
- 27:05Brian: Okay.
- 27:05Brian: You know, and let, let's talk for a minute about, um, um, reality.
- 27:11Brian: Let's call it reality, right?
- 27:12Brian: Let's talk about health equity for a, for a minute, right?
- 27:15Brian: You know, we're talking about AI, we're talking about this, this new, this n- you know, new era or ongoing era that we're in right now.
- 27:23Brian: But, you know, we see a lot of clients who are in rural areas that are in underserved populations, especially on the wound care side, right?
- 27:31Brian: And, and, and that doesn't necessarily mean that they get the benefit of this, right?
- 27:36Brian: That they're-- that the, you know, ma- the, the resources that are at their local community hospital aren't there.
- 27:44Brian: Their, their, the private practices, uh, the, are not, the clinics are not there as far as the technology goes.
- 27:51My question basically is: How does automating a referral process like, such as what you do, close that health equity gap for rural patients, for pa- for patients that live in areas that maybe aren't, um, at a major health center or that sort of thing?
- 28:09My question basically is: And, and, and is that something that you actively build and train for?
- 28:13Trey: You know, one of the, like, things when, when folks think about automation, they tend to think about how do you automate how it's done today, when in reality, the best frame of mind is how should it, how, you know, how really could it be done if you had unlimited bandwidth?
- 28:25Trey: This concept of, um, if we had not 10 people doing this work, but 1,000 people doing this work, how would you do it?
- 28:31Trey: And so one of the things when, you know, with our primary care clients, um, they are really cognizant of, um, I don't want a patient driving three hours when they could drive one.
- 28:42Trey: Um, I don't want patients going, um, you know, to a really low-quality, uh, provider.
- 28:47Trey: And, you know, today, if they're, they're usually very large folks we work with, and so they might not have the ability today to train every single person that's gonna review an outbound referral and dictate Uh, this is the quality score that needs to be met.
- 29:03Trey: This is the location of the patient against the location of the clinician.
- 29:06Trey: Of course, you have some great employees that are gonna do that work intuitively, um, but ensuring it's done across every single patient in, in these, you know, rural patient demographics, like it's just one of the massive benefits of like structuring processes like these so that you can ensure that a patient is gonna know that they have options and, um, and they're not going to get lost in the mix.
- 29:29Trey: You know, another thing, and this sounds stupid, but it's, it's so true.
- 29:32Trey: Um, you know, I like to really think about how many patients are you calling by the time they get to the parking lot?
- 29:39Trey: Um, you know, especially if you get a same-day appointment You know, we, you know, there's oncology, massive, you know, um, practices that before working with us were not contacting patients for, you know, three-plus weeks, right?
- 29:53Trey: And what does it look like when you're contacting them while they're walking to the parking lot saying, "Hey, we have all your documentation. We know your insurance. This is gonna be the next step." What does that do from just a, like an assurance, especially when somebody's got to drive three hours home?
- 30:07Trey: How do they know or, you know, frankly, if they can get in same day to somewhere else?
- 30:10Trey: It's just awesome.
- 30:11Trey: So i- it's all predicated on if you structure these processes, you think about how they could be done in the best case and not in the, um, the current state, it's just way better outcomes for especially folks like that.
- 30:22Trey: Yeah.
- 30:22Brian: No, I mean, you talk about, like, you talk about patients disappearing.
- 30:25Brian: That would make me disappear if I felt like as though I got a call three weeks later, or I felt like as though I was being asked to do something that was unacceptable or just not reality for me.
- 30:38Brian: You know, I've got a family.
- 30:39Brian: I've got things I've got to do.
- 30:40Brian: I've got, I've got all sorts of responsibilities.
- 30:43Brian: That would just make me disappear.
- 30:45Brian: I would just not, I would just not follow up.
- 30:46Brian: I would just not be involved in it at all.
- 30:48Brian: I'd just move on.
- 30:50Brian: That's a, that's a problem.
- 30:51Brian: Okay- Yeah, absolutely … you touched on something earlier.
- 30:53Brian: Yeah, thanks.
- 30:54Brian: Yeah, yeah.
- 30:54Trey: We always say, we always say, "It's a real problem." Like- Yeah.
- 30:57Trey: … it's just a problem- Yeah … that needs to be solved, that in 30 years we should all look back and say, "Wow, that was a real problem. I can't believe that existed, and now it's solved." That's what it sh- we should all be able to say.
- 31:07Brian: Yeah, yeah.
- 31:08Brian: Yeah, right?
- 31:08Brian: Yeah.
- 31:09Brian: Okay, you touched on something earlier I wanna go back to, um, and connect to, um, connect a little bit differently.
- 31:16Brian: You talked about-- We talked, we were talking about skepticism around AI.
- 31:20Brian: You talked about how your, you have a certain very strict structured criteria around, um, when somebody is ready for Tennr, ex- uh, they need to see that it works.
- 31:31Brian: They, they, you need to prove your case for a client to actually buy in and, and, and, and, and for this to occur where there's a benefit.
- 31:42Brian: Uh, this, a, a lot of people with their arms raising, their arms up in the air, a lot of AI companies that maybe aren't kind of delivering what they're supposed to deliver.
- 31:51Brian: It's a real issue for people who are actually trying to make change within the structure of the, of the healthcare space itself, right?
- 32:00Brian: What-- Give me some context here.
- 32:03Brian: I'm, I'm, I wanna press in on that a little bit, right?
- 32:06Brian: What do you think separates the companies that are, the, the, those companies that are gonna go away, right?
- 32:11Brian: 'Cause the, a lot of people talking about a reckoning, you know, that these are, some of these companies are gonna go away.
- 32:16Brian: What do you think separates that, the companies that are gonna still be standing three years from now?
- 32:20Brian: And, and, and when you talk about providers, like specialty care providers for us- They can't afford to make a bad bet.
- 32:28Brian: Their margins are very thin.
- 32:30Brian: How, uh, uh, how c- they can't afford to make a bad bet on some- somebody who promises something that they can't deliver.
- 32:37Brian: What should they be asking those, their AI vendors to make sure that they're getting what they, what they need to be more efficient?
- 32:46Trey: Yeah, and I do not envy the position of these folks that have to evaluate these systems.
- 32:51Trey: Um, because look, you, you-- we would all agree that there is, you know, discrepancies of, uh, of, of doctors that are incredibly good and maybe some that aren't so good.
- 33:02Trey: Now, in, in the world of computers and engineering, um, there is a far, far more vast discrepancy between the best engineering teams and the average engineering teams.
- 33:13Trey: Like, it's just dramatic.
- 33:15Trey: And so the-- in our world, like why I-- why we have strict criteria on our side for who we take on is, you know, the people that actually do what they say they can do, it makes you a luxury good.
- 33:30Trey: Like, there are so few people that act-- their do to say, you know, ratio is, uh, is healthy, and it's really important to us that we do every single thing that we say we can do and do it incredibly well, and that allows your growth to come from people talking about you, and it allows you to, um, you know, take clients that, um, you know, you're gonna smash it out of the park with.
- 33:52Trey: Um, you know, in terms of the-- how I would look at it from the, uh, if I was on the buying side, it's like, one, I need to like see excellence in just about every touchpoint.
- 34:02Trey: Frankly, I really wouldn't… You know, in many cases, like unless I know I have like a dev team or I'm gonna really co-dev something, I would not be working with like a team that doesn't, you know, couldn't drop everything and fly four to five people out to solve a problem.
- 34:15Trey: You're, you're just increasing cybersecurity risk to a threat that's just incredibly, incredibly risky.
- 34:21Trey: Um, yeah, I wouldn't frankly be working with somebody who doesn't have like tens of millions, ideally a hundred-plus million, like, dollars in, in a bank account somewhere because again, cyber risk, this stuff is so real, and frankly, it's so expensive if something custom needs to be done to save the day.
- 34:35Trey: You know, what happens when ten thousand patients get lost, right?
- 34:38Trey: It's like there's-- you can have the most redundant system in the world, and you will mess something up, and so you better have like a really sharp team that can come in and, you know, save the day in that scenario.
- 34:47Trey: So I would be highly scrutinizing, or I would just wait in a lot of cases from folks.
- 34:51Trey: Like, that waiting is an option.
- 34:53Trey: I know that's probably not something you would expect, you know, me to-- us to say, right?
- 34:55Trey: It's just like some people should just wait.
- 34:57Trey: They shouldn't be such early adopters jumping at the bit.
- 35:00Trey: And then finally, folks really need to, um, like understand, uh, how much they are tying themselves at the hip to some of these businesses and understanding one-way doors or two-way doors.
- 35:12Trey: So a two-way door product might be trying a new eligibility and benefits tool.
- 35:16Trey: "Hey, guys, we're in every, you know, patient today, I wanna run benefits through this other tool." Cool.
- 35:22Trey: Two-way door.
- 35:23Trey: Doesn't work, eh, we'll change it tomorrow.
- 35:24Trey: No problem.
- 35:26Trey: You know, a one-way door is, "Hey, um, we are going to fundamentally change the journey of every single patient- that we take on, how we take them on, how we engage them, uh, the rule sets that we triage them against and qualify them against, and, uh, how we engage our referring providers.
- 35:46Trey: You know, that, that's a one-way door.
- 35:47Trey: Like, you're getting in, you're, you're getting in serious stuff with somebody, and so you should be highly, um, you should run a real process.
- 35:54Trey: And if they're not running a real process, you should be very skeptical.
- 35:56Trey: Um, those are all the things that I guess I would look for is it, it comes down to, like, s- kind of professionalism that you see on both sides.
- 36:03Trey: Because if you're working with, you know, a client-facing person that isn't really dialed in and buttoned up, you should assume that's gonna be what the rest of the org potentially looks like.
- 36:12Trey: And that's why we try to hold high standards to everybody on the team.
- 36:16Brian: Awesome.
- 36:16Brian: All right, last question for you about the future, of course, right?
- 36:21Brian: Um, when you imagine healthcare in 2030, what's, what's different for the patient who used to fall through the cracks?
- 36:27Brian: You know, you talked about, you talked about, uh, patients that disappear.
- 36:31Brian: Um, what would, what makes you feel like the work that you're doing is gonna be worth it when you, when you think about 2030?
- 36:40Trey: Um, you know, I think it's like, uh, it's funny because I almost think nobody will notice it when-- as it's happening, uh, right?
- 36:48Trey: And so you'll just notice when the lack of or something.
- 36:51Trey: Um, but you know, a patient will get discharged, um, you know, they were hospitalized for, you know, something, uh, you know, like a renal issue or something like that, and their primary care will know within twenty minutes and already be calling the emergency contacts or the patient to make sure they can get back in and be reassessed through a value-based framework.
- 37:10Trey: Um, you know, that the, uh, the physician that then sees that patient in the, in the, in the practice, in the fi- in the, in the primary care practice will, um, have all the context of the entire episode that just happened and will already have outlined a very clear plan of what is likely to happen, and they will have that added context, and they'll be prepped, and they'll be able to have a quick and more effective conversation with that patient about what's gonna come next.
- 37:35Trey: They will then, you know, send the patient over to a specialist, like say, you know, a nephrologist and, you know, and that nephrologist is gonna call them within, you know, twenty minutes by the time they're walking to the parking lot.
- 37:45Trey: He's gonna walk them through what steps need to be gathered, what documentation might be, you know, back stuck at the hospital for whatever reason, and, um, that's gonna be fetched.
- 37:52Trey: If their insurance changes, that's gonna be real time.
- 37:55Trey: And, uh, "Oh, wait, we need lab before we can get you in." Okay.
- 37:59Trey: The, you know, lab site is gonna be calling that patient, um, and showing them every step of the way.
- 38:04Trey: So what it'll feel like basically is like you have this sort of invis- like invisible usher that's walking you through every step of the way.
- 38:10Trey: And I have this sort of pessimistic take that, like, the only people that get good care in the US healthcare system have really good mothers.
- 38:16Trey: You basically need a hyp-hyper-competent, uh, borderline omnipotent mother, um, that is doing that work pro-proactively for you to get you to that next stage, and that's what I think it'll feel like.
- 38:28Trey: It's like, damn, you know, is there like some mom behind the scenes here that's, uh, doing this work on my behalf that's telling me when there are issues before I even know about them?
- 38:37Trey: Um, you know, we've all had, uh, you know, hopefully everybody gets to have that experience of when you have a hyper-competent person organizing and project managing you through a really difficult situation, and that's just what it'll look like.
- 38:47Trey: It'll-- all the hard work is behind the scenes.
- 38:49Trey: You're just getting calls as a patient, and, and it's like, "Oh, that's my next step. Amazing." Right?
- 38:53Trey: That's what we want to happen.
- 38:54Brian: You know, you, you, you talked to the, the-- That's so great because you, you actually had mentioned it on another interview.
- 39:00Brian: You mentioned, uh, the idea of having your primary care physician reminding you, nudging you, and your specialist reminding you and nudging you and, and sort of ushering you along the way with full transparency into pricing and the rest and, and, and the, the rest of the things that you need to know as a patient so that you're-- there's no, there's no missed gap there.
- 39:24Brian: There's no black hole in that, in that scenario because both sides of it are actually informing the patient.
- 39:31Brian: That's- It- Is, is,
- 39:33Trey: is that right?
- 39:34Trey: It's exactly right.
- 39:35Trey: If I order-- If I ordered a package and I added it to my cart and pressed checkout on Amazon, and it said, "You know, we maybe have your credit card details.
- 39:43Trey: We're not totally sure what it's gonna cost.
- 39:45Trey: This might ship in four days.
- 39:46Trey: It might ship in three weeks.
- 39:48Trey: If you call a number, we might pick up," you would not, you would not buy the paper towels from Amazon anymore.
- 39:54Trey: And healthcare has so much demand because it's not paper towels.
- 39:57Trey: It's like life-changing treatment.
- 39:59Trey: And, uh, and so people are persistent, and they work through that sort of pain.
- 40:03Trey: But that doesn't mean that that barrier should exist.
- 40:05Trey: And you know what's crazy about it, right, is Amazon sells way more paper towels and provides be- better service to those, you know, consumers because they have streamlined every step of the process with complete visibility.
- 40:17Trey: The providers that are doing this and offering this to their providers, this is not, you know, terminology you hear much in healthcare, but they're just winning.
- 40:23Trey: They're just winning.
- 40:24Trey: You know how we find a lot of our customers?
- 40:26Trey: They start losing in the market, um, from-- to one of our customers that's a competitor of theirs, and then they're like, "What the heck are they-- What's going on over there?
- 40:33Trey: Why are all of our patients getting sent over to that specialist?" And it's because they're providing such an incredible service and, you know, concierge experience to those patients, you can't compete.
- 40:42Trey: And that's, you know, what the, the, you know, the f- that's the free market in the, in its best here is driving such good experience that you would be crazy to send a patient anywhere else.
- 40:53Brian: Wow.
- 40:54Brian: All right.
- 40:54Brian: Well, Trey, thank you for this conversation.
- 40:56Brian: It's been really great.
- 40:57Brian: And, and honestly, um, man, I could talk to you for another hour and a half at least about what, what you know and what you've learned through this process and this journey that you've taken.
- 41:07Brian: I mean, I, I-- and I think, I think honestly with our-- for our listeners, what you're calling a black hole, we call a gap, a gap between the-- where, where technology-- a gap that technology can fill between, um- Between the, the, the work that needs to be done to make things happen and the actual care that needs to be provided, um, at the same time.
- 41:33Brian: And that I think is- Right … that I think is the most important thing of all.
- 41:37Brian: So thank you.
- 41:37Brian: Thank you again, um, for, for joining us, and, um, good luck.
- 41:43Trey: I appreciate it, Brian, and thank you so much for having me, honestly.
- 41:45Trey: Likewise.
- 41:45Trey: Uh, the-- I could probably fill another hour wanting to ask you the questions, um, but we don't have time , so I appreciate it, man.
- 41:52Trey: Uh, have a great rest of your, uh, week.
- 41:54Brian: Yeah, you too.
- 41:54Brian: Thank you.
- 41:59Brian: All right, recap time.
- 42:01Brian: Here are the top three things I took away from my conversation with Trey today.
- 42:05Brian: First, the referral black hole has persisted because the raw material at the center of the problem, unstructured faxes, handwritten notes, payer criteria buried in PDFs, is genuinely hard to read reliably at scale.
- 42:20Brian: General purpose AI wasn't trained on that material and wasn't designed for those edge cases.
- 42:25Brian: Tennr's wager was that fixing this required a purpose-built model trained on the actual documents the system would encounter in the wild, not a layer on top of something that was built for a different job.
- 42:37Brian: Second, automation that earns adoption lives inside the workflows people already use.
- 42:44Brian: If frontline staff have to change how they work before they see any benefit, the rollout is already in trouble, and the skepticism they bring is earned, not unreasonable.
- 42:55Brian: The organizations getting real results from AI and administrative workflows are the ones that start with the friction people are already feeling, not the ones that introduce a new system and ask people to adjust.
- 43:08Brian: Finally, tech should be facilitating a natural transition to seamless care.
- 43:13Brian: In other words, you shouldn't need a mom to coordinate your healthcare journey.
- 43:17Brian: Patients should actually get scheduled with no fuss.
- 43:21Brian: Discharges should happen the right way without a thought.
- 43:24Brian: Notes and treatment plans should follow patients home for continued care.
- 43:29Brian: That's what we're working for, and that's the promise of AI.
- 43:33Brian: And that should do it for today's episode of Humanify Health Tech.
- 43:36Brian: As always, I'd like to thank all of our listeners for tuning in.
- 43:40Brian: Francesca will be back next time, but it's been a pleasure filling in.
- 43:44Brian: If you're joining us on Spotify or YouTube, we'd love to hear your thoughts on, in the comments, and don't forget to leave us a rating while you're there.
- 43:52Brian: Once again, I'm Brian Chee, Senior Director of Content and Communications at Net Health.
- 43:57Brian: I'll see you next time.