This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.

Newsday: Epic's Code Red - Why AI Agents Are Healthcare's Next Power Struggle with Jacob Hansen

[00:00:00] We want to thank our partner, Avasure. Over 1, 100 hospitals are using Avasure's virtual care platform to engage with patients, optimize staffing, and seamlessly blend Remote and in person treatment at scale, drive measurable outcomes and augment clinicians with an AI powered solution that deeply integrates into your clinical workflows.

Aversure offers virtual care solutions supported by a secure, scalable infrastructure that helps you to lead your organization into a future where cutting edge technology is at your fingertips. and Compassionate Care Converge. For more information, check them out at thisweekhealth. com slash Avisure.

I'm Bill Russell, creator of this week Health, where our mission is to transform healthcare one connection at a time. Welcome to Newsday, breaking Down the Health it headlines that matter most. Let's jump into the news.

Bill Russell: Hey, it's Newsday, and today we are joined by. our normal panel of, Suspects, [00:01:00] I don't know if that's the right word. Sarah Richardson. Drex to Ford. And, our special guest today, Jacob Hanson, uh, chief product.

And, what'd you say? Technology officer.

Jacob Hansen: That's the one.

Bill Russell: Wow. pretty good with Avasure and, we love, love talking to you. We get to talk about where, this whole AI movement's going, where the whole, product movement's going. And, you sent us some stories today and I've been reading a couple of different things on this very topic, so I'm looking forward to this.

first story, we're gonna look at Brandon, or Brendan, Keeler Health, API guy, published just today. How's that for breaking news? Of course, by the time this airs, it'll be a week old. By the way, did you guys check out the artwork on this article?

Drex DeFord | This Week Health: Yeah.

Bill Russell: Very nice. I mean, Judy, Judy looks like she's conquering the new world in that, ship going over.

And, she's got, and her and shiv with like arrows and stuff going at each other. That's pretty interesting as [00:02:00] well.

Jacob Hansen: it was entertaining for sure. He used some AI to develop some fantastic artwork.

Bill Russell: Oh, I think he was sitting there with his pens last night doing a little ink drawing and whatnot. Uh, really what this is about is the, EPIC AI charting the scribe thing that is going on, and, I learned a couple things from this article. you know, I mean the, the first thing is you had, epic CMO, Jackie Gerhardt talking about, you know, they're not viewing it as a scribe, just a scribe.

They see it as an active clinical co-pilot, which I think is interesting. And then the second thing, and then I'll let let you guys get going is, it's not. I thought it was Dragon Dax on the backend, and I've heard more than one person say it's Dragon Dax on the backend. This article, uh, comes, it make, gives us some clarity and it says, EPIC is using Azure to access multiple L LLMs for different specialties in [00:03:00] different tasks.

And I don't know if that's a, token thing or if, it probably is partially a token thing, but it's also, using the right model for the right tool, which is. Pretty smart architecture. So, Jacob, you sent this one over, so I'll, I'll give you, uh, I'd, I'd love for you to, to talk a little bit about what you know, what, what jumped out at you in this story.

Jacob Hansen: The first is anything involving ambient clinical documentation is particularly interesting to us as a virtual care platform, especially with an open API because anybody that wants access to, streamed audio out of a virtual consult or that wants access to. Ambi ambient audio as delivered out of the room. When a patient and a and a a care team member are talking, we have the opportunity to deliver that audio to those third parties. So we're always very interested in what's happening in that [00:04:00] space. I thought it was very interesting.

There was a particular graphic in there. that talked about, how the, I think it said the arena is heating up is the phrase that was used it was basically referencing all these different, uh, existing AI use cases. Right. Rev cycle's been around for a long time for back office, interactions with the patient.

There's an opportunity for AI and a copilot to support patients. Then you've got work with the EHR scheduling. And all of these things are, um, have an opportunity to create value. That means that, endpoints like ours have to be sensitive to just more and more use cases. And the speed at which they are proliferating only gonna just go faster and faster. So how do we make sure our tools are ready for that? And then the second piece of the puzzle, there was a particular quote that I grabbed, It says In a world where the marginal cost of writing [00:05:00] software is approaching zero differentiation, no longer comes from building a feature. It comes from assembling all the functions.

A business depends on that dynamic, pushes vendors to expand laterally. It's now a race to build out all the functions a business might need, and then later said, the model makers will be the arms dealers. And I thought that was particularly interesting because we are hearing about vendor consolidation on the daily from health systems, right?

They're all saying it, oh, I have 1500 clinical applications. I need to shrink that to 500, and I gotta do it in, you know, a year and a half. Um, and oh, by the way, if these other big, big vendors do it, let them do it. Don't touch it. Make sure you stay focused over here. This is just the perfect example of, how these various use cases come together.

I think it's gonna be really interesting to watch.

Bill Russell: although he does make the point that he doesn't think that [00:06:00] these third party, you know, one of the questions he strives to answer is. What about these third party ambient listening tools? And, to your point, he's like, they are going to expand, across laterally as well.

And the thing is, I think as Epic was sort of looking at this entire thing, they said, Hey, we gotta protect, we gotta protect ourselves here. Right? Like they're starting to work at the. first of all, they're, the interface to our client. And second of all, they're starting to work across this whole layer.

they're gonna create the workflow. They're gonna create, things that generate value.

Sarah Richardson: Jacob,

I consider your product becoming an aggregator, almost like middleware, for all the right reasons. So all those conversations you can capture, you can distribute, you become one of those 500. Because of the ability to integrate in so many different aspects of the patient room, your technology's already set up and in many cases already doing some of that.

So you become one of those spaces that people say, I'm partnering with [00:07:00] you to be able to leverage all these other capabilities as it comes down the pipeline. You have the rapport, you have the reputation, you have the quality, you have the check marks built in. Like what a perfect place to be right now.

Jacob Hansen: it's also going to really pull hard at the appetite for taking on cost with infrastructure, with cloud compute. I mean, you think about it. as the endpoint as the aggregator, right? We could play that role. what if the health system says I have, and let's just pretend in the future it's some big number 15 vision AI models that I want to interrogate the video coming out of that room. if you take that video and you start streaming that out to the cloud in 15 different places, what does that mean in terms of, broadband and, and um, capacity? What do we have to be able to support? It just could get out of control relatively quickly. if you [00:08:00] don't grant everybody access at the same time, what do you do to arbitrate it? You gotta have rules. You gotta have somebody working hard about clinically. What matters most? How do we empower health systems with that kind of choice to say, here's the different tools we want to use. Here's the priority order by which they get access. Becomes an interesting thing that we really have only started to scratch the surface on.

Drex DeFord | This Week Health: I think the architecture and the governance part of this winds up being really important. When you think back to the early days of meaningful use, why a lot of health systems bought Epic. I think in the beginning a lot of it wasn't necessarily about, the electronic health record itself, although it was an electronic health record, a lot of it was about the kind of mandate that they came with about You use the tool like this, you do the process like this, and then the tool will work really well for you. And so. Use it this way. And I feel like there's probably a little bit of that's now happening with ambient in that conversation. [00:09:00] at the same time, having been around for 30 something years, I remember the good old, bad old days when, it was best of breed and there were lots of different solutions that you could choose from.

Uh, and you know, that. Made end user physicians and nurses and Lab and RAD and everyone else happy because they had their thing that worked for them. And so I think we're in this really interesting point where we're going through an evolution right now of are we going to be on one platform? Are we going to figure out how to use a lot of different solutions that really satisfy our needs? All in the context of do we have the resources to manage all of those? Bananas. Uh, as Sarah would say, as we look at lots and lots and lots of different solutions, when we already have 500, you know, applications inside most of our organizations, how are we gonna do that and do it well?

Bill Russell: Epic is looking at this as a Code red environment. If you're not familiar with Code Red, this is [00:10:00] like the email that Bill Gates sent to the organization when he realized the internet was a thing or when, Google realized that their search business was going to get cannibalized by these large language models.

these organizations create a. Code red environment where they're like, if we don't do something, we're going to,

Drex DeFord | This Week Health: alive.

Bill Russell: we're gonna be, get, get eaten alive. And this was a code red moment for, epic. And it's at, at one point they had a very strong partnership with a bridge. As the rumor mill goes, they had an investment in a bridge.

They ended up selling off that investment, middle of last year and announcing their own, direction here. 'cause they realized that whole aspect of, Hey, wait a minute, they're gonna be orchestrating workflow on top of our platform. Like, that's where value is generated for the clinician, for the health system.

their response to that, obviously they made that announcement that, hey, we're coming out with this stuff, which is what they usually do. Hey, we're doing these things. the [00:11:00] other thing they did is they came out with, uh, lemme see if I remember all these Art Emmy Penny.

Emmy Penny. They came out with those three, which I've not heard great things about yet. they're there and they're developing and everyone gives Epic the benefit of the doubt. 'cause generally their first release of anything is okay. It, it, it is almost like a pilot. they get to see what it directionally is going to do.

they have enough trust and faith in Epic because they've done this over and over again where they, they announced something to slow things down. They come out with, and then three years later they're sitting there going, okay, this is functional. It's working. But the thing I think that flies under the radar is their agent factory.

So they started talking about their agent factory back at UGM, and they are now. Very much pushing the agent factory like we're hearing from health systems that the agent factory is their response to, Hey, do you realize Claude Coat can do this? Do you realize [00:12:00] that, Palantir's helping this organization to do this?

Like they're taking this data and they're doing these really cool things and they're enhancing workflow. And what Epic is saying is, look, we have this. agent factory where they're focused now on, the ability to build out these multi-step, functions within the EHR in their factory setting, which will have the guardrails and whatnot.

They're essentially positioning themselves as, Hey, we're gonna be able to augment your workforce, augment your workflows, take things to the next level they want to protect that the value that gets generated from the system and not just be known as the system of record, if you will, and get, pushed into that.

That's what I'm seeing. That's what I'm hearing. That's what I'm seeing in these articles is that, the dawn of agentic ai, this digital agentic workforce is upon us. It's not, it's healthcare, right? So it's not gonna be. As rapid as it [00:13:00] is in other places, but it's here and it is gonna be really interesting.

I don't know if you saw this, but, Claude code, they released, this legal framework for doing like NDA reviews and all this other, so Claude code, I'm sorry, not cloud, cloud cowork can do multi-step ag agent AI stuff. And then they release this framework from within that. So lit.

Literally you could take your legal document, drop it in there and say, do your legal review on this, and it comes back with whatever. and essentially the argument's being made that this is trimmed like a hundred billion dollars in market cap off of companies that were sort of positioning themselves to do this kind of stuff.

that's the kind of cannibalization we're looking at This age agentic world that we're entering in, people are, if nothing else, they don't know where it's gonna lead. And they're going, oh my gosh, this is gonna change how software's developed. This is gonna change where the value creation is done.

Jacob Hansen: I think that's some of what, the author of this first article Brendan was arguing is, [00:14:00] there, there will be. There will be consolidation. That some of them are gonna start attacking adjacencies to justify both your existence and, present the opportunity to raise your own prices.

Like where's the growth gonna come from? It's gonna become, uh, the fact of life that if you're in a driving position, you're gonna have to start looking what is happening, in these workflows next to yours. And there's likely gonna be somebody already doing it. But can you bolt it on top of yours, bundle and do it better? uh, it's gonna be a, a, the notion of innovating there, I think it'll be a funny word to use, this idea of innovation because of how fast things are coming out and changing. it's really who can do it and do it well in a way that covers enough different workflows to make it even feasible.

Bill Russell: So Jacob as a product guy though, as a product guy In that ecosystem, do you [00:15:00] say, you know what we are, we're going to fine tune the relationship between the end user and the system of record, but we're gonna allow that agen layer to really determine where the value gets created even from your data streams.

Or are you gonna say, no, no, no, we've gotta keep value on, on the data streams at our layer, like we're gonna have our own layer that creates value.

Jacob Hansen: one, there's a real question of core versus context, right? Being really, really clear and honest with ourselves about what we can be the best in the world at. If we look backward at our history and we, we have to acknowledge what, what, what has made Asure, a pioneer in this category. It's an organization that understands how to deploy and manage clinical grade endpoints in a way that nobody else has experience doing. an organization that has lots of experience with driving clinical change management practice on top of [00:16:00] smart technology. so for us core is managing a fleet of devices that need to be remotely monitored, remotely controlled, remotely, um, that allow our applications and AI plus others to access it. then ultimately our core also includes a commitment to computer vision because video is the, is the hard part in continuously monitoring rooms. And we've been doing that for a long, long time. video out of rooms for long periods, days, or even weeks in supportive risk at risk patients.

So computer vision and the platform itself are huge focuses for Avior. There's no question though. in the world we live in now, we cannot have any pride of ownership or authorship over different utilities that you use. Just like what we saw in that article about Epic, if somebody else has something that they've created that we can [00:17:00] use inside of our stack, we're gonna do it. We don't need to build everything. We will be too slow if we do it that way. It's just the facts.

Bill Russell: what would constitute a code? Red? So Sarah and Drex, you're a CIO. What would constitute a code red that you're looking at in terms of this world that we're entering into that you're just looking at going, Hey, this is either this is unsustainable or this is coming faster. there's something in this article as well about this of, Hey, one of the things that could change is everyone's trying to take this layer.

Outside of Epic, completely like, like this whole caring for people and what they're trying to take it outside. It is like, you know, if, if that ever happens, not only does the value of Epic go down, but the value of the health system goes down.

Drex DeFord | This Week Health: some of the code red conversation, and in my head, I immediately go to Mountain Dew Code Red because of those stacks of cans that, exist in our data. or offices when things are [00:18:00] going wrong. But, thing that I look, I think about is I kind of go back to the Epic thing again.

We deployed EHRs, but what we really wanted from Epic was tell us how to do the workflows better, use the tool to do the workflow flows. And that's really why we paid a premium for it. I feel like we're getting to this point now where with the agent conversation, The code red concern that I have is, are we making a decision?

Are some health systems ultimately gonna make a decision that we're not really gonna do innovation anymore? The innovation that we do is going to be. the folks at Epic saying, leave the innovation to us. We'll tell you what agents to use. We'll tell you how one workflow connects to the other. you know, we are gonna give you all the advice and guidance that you need on innovation.

So there's gonna be the. Epic innovation folks. And then there's gonna be the folks who want to color outside that box and they're going to use [00:19:00] agents from other places and figure out their own integrations or work with Epic to do integrations. I think, you know, Jacob is, as much as you all can stay Switzerland in all of this, and like help the companies who are gonna be doing a lot of these things on their own, but also. Continue to stay, integrated with Epic and Oracle and others. probably the, you know, maybe the better off you'll be, but it's, I worry about us deciding at some point. We're just gonna outsource innovation. We're gonna leave it to Epic.

Sarah Richardson: Oh, see, I'm gonna totally take the counter on

Drex DeFord | This Week Health: Let's go,

Sarah Richardson: and you know that

Bill Russell: All right.

Drex DeFord | This Week Health: let's go.

Sarah Richardson: because you all know exactly what I would be doing right now. I do this Jacob, and then I call Jacob and I say, my code red is what I'm gonna call reverse all the G. And now you get to build my roadmap with me. Best of breed yet with your foundation in place. And I'm gonna pull in my four or five top partners and I'm gonna create the universe that I've always wanted. And I'm always gonna put [00:20:00] consumerism and patient experience first so that the clinical. Workflow makes sense to the doctor and the reimbursement models are all there.

I want to disrupt where I'm being held hostage, which is what Bill has already talked about. Bill's been really polite in his fiction series about taking a scenario that's actually there. We haven't quite dipped into the fact you're also held hostage by your EMRs right now the amount of money you're spending to get almost good enough and wait for it, now you get to actually pull those elements in.

And I've always had like a ninja crew. Who did stuff they weren't supposed to do to disrupt and have better outcomes in my facilities. This now almost makes me wanna jump back into a hospital and be like, let's code Reddit from our own

Bill Russell: Yeah,

Sarah Richardson: of my phone calls because you have everything you already talked about in place and you know how to do it.

And we could just ideate and think, where do I become an innovation test bed for you? That benefits everything happening in my health system.

Jacob Hansen: Gonna say.

Drex DeFord | This Week Health: Sarah, we have the same issue, right? I'm afraid that we're just gonna say we're [00:21:00] gonna, we're going to just our innovation capabilities to Epic and just say, we'll just take whatever Epic brings. And you're saying that's a bad thing. And I think it may be a bad thing too, but.

Sarah Richardson: But we

Jacob Hansen: I.

Sarah Richardson: it right. We also don't wanna be the ones who went out there first and all of a sudden those canaries are laying dead in front of your mind.

Jacob Hansen: Yep. That's the, that's the, I think that's the uh, um. interesting piece of the, the puzzle with this is, everybody's concerned about being first.

Drex DeFord | This Week Health: Mm-hmm.

Jacob Hansen: have definitely been a number of health systems that have been a bit burned by, um, by being, early. out the gates with something new. But we are definitely getting a lot of phone calls that are saying, Hey, we have this interesting problem we wanna solve. want to co-develop. We wanna understand what the commercial model looks like. This is gonna require incredible flexibility legally and commercially going forward because health systems want more control over. [00:22:00] Their own destiny for sure, and they want to know that they're gonna be able to focus on the problems that matter to their system based on the cohorts, the patients that they take care of and the problems those patients face. And that doesn't necessarily mean that, it's a one size fits all.

I think that's gonna create some interesting friction as well.

Bill Russell: So, just to close this one out, all these stories sort of weave together really nice. Aaron, Einstein, who is the CMO at Notable has been on the show before. love talking to Aaron. so we talked about technology, we talked about the scribe, we talked about agents, and the platform.

the question that I think should be keeping CIOs up right now is who on your team is gonna build all of this so they're so epic. Has this builder program, right? They have physician builders and all this other stuff. And generally what they've done is they've translated these requirements and they've built out features or put a [00:23:00] button in or whatever, and that's what they've done.

When you look at these age agent workflows, they're a lot more fluid, a lot more. rapid in terms of the development. And so there's two things I think about here. One is the, staff. The staff to build this out is really gonna have to understand the entire workflow. It's no longer gonna be just listening to somebody.

In fact, I would say some of the nurses don't understand the entire workflow. Some of the doctors don't understand the entire workflow. And so your, your team, your clinical informaticists are gonna become. And, and this was the point of his, his story are gonna become critical and they're gonna become more than just, builders.

They're gonna be AI builders, they're gonna be people that see this workflow and they're gonna be able to create the on the, uh, uh, agent factory. They're gonna be able to create these multi-step, workflows that are going to change how your system operates. And so, you know, it shifts from order taker.

And builder to sort of a co-owner [00:24:00] of the operational, process or the operational performance. And that's not just a tool change. I mean, that's a career defining org chart changing identity level shift for health it. So that's a big deal of itself. And then the second is Epic likes to do their updates every quarter, every six months.

Alright, so when we start layering these things on top.

Drex DeFord | This Week Health: Nonstop.

Bill Russell: Are we going to get to the point where that's a form of best of breed, and then all of a sudden These upgrades are breaking more things that they didn't anticipate, or the guardrails gonna be such that the agent factory takes all that into account and allows that, that lower level to be able to continue to change.

I think they're smart enough to figure that out and make sure that that doesn't happen, but that would be a concern of mine. The other is. Do I have the staff for this new world that we are walking into? Because if I'm a builder, even if I'm a, Hey, we're gonna [00:25:00] go do our own thing, or even if I'm a, I'm gonna rely on Verona to be my innovation arm, my team skills, in both cases dramatically shift.

That's for me, that's the red carpet.

Drex DeFord | This Week Health: red.

Bill Russell: Code red. Thank you. Red carpet. Red bull code. Red. Uh, but it could be a Red Bull moment, man. We need to start studying. We need to start, it's gonna be a, a, an interesting shift,

Drex DeFord | This Week Health: shaking his head no, though.

Bill Russell: please.

Sarah Richardson: able to release code daily and not have it be disruptive.

Bill Russell: That's a different world.

Jacob Hansen: that's where I was going as a product guy is, is how many of, in what way will these various, AI vendors who take advantage of the infrastructure that. Epic or somebody else is providing. In what way are they all supporting continuous integration and deployment principles? How, how is their architecture going to support that?

And not only that, [00:26:00] but I'm sorry, just, just everybody's got a different, a different attitude about speed to market and depth of testing and quality. There's no way, there's no way. The guardrails cannot be that deep. That you're gonna, that you're gonna have that level of control. And then the minute they start interacting with each other. You get two or three different AI solutions each with their own unique view on how much testing is enough testing. Was it smoke? Was it integration, was it penetration, was it full test? And then they rush out the door and now they're talking to each other. The speed at which we're gonna expose defects and bugs and confusion. I guess my confidence is not as, deep as yours bill relative to how well that's going to happen. And the other part of the puzzle here is clinical efficacy and monitoring actual clinical value. who's gonna drive the studies? And this is something [00:27:00] so important to our business. I'm so grateful to work with our chief clinical officer, Lizbeth, uh, Viva.

She, she brings such rigor to. If we're gonna use this, what does this mean to, measurable patient or staff outcomes? Why do we believe it? Did we run a study that actually demonstrates it? Otherwise, it's just noise.

Drex DeFord | This Week Health: It'll be interesting too, bill, that this idea of, know, you were sort of talking about the, the change that really happens now is that the CIO and the CIO's team through these agents,

Most of those knobs were in the hands of managers and clinical department leaders and rev cycle leaders, and now the CIO's team. May very well have their hands on a lot of those knobs. So those outcomes you're talking about, Jacob, how do they get measured? What is the efficacy? How much money do we save our patients healthier? there's also gonna be a grab for all those [00:28:00] knobs on the machine to who's,

Bill Russell: Well.

Drex DeFord | This Week Health: what,

Bill Russell: but Drex, I'm not sure it's knobs anymore. When you think about where this agent layer sits, it's as if it's a person sitting at a keyboard with a mouse and a keyboard going,

yeah.

Jacob Hansen: Giving away. It's giving away control in a way we've never seen before. And that's where I find some of the other phrases, the things that I'd seen in some of the articles on your site, augmented empathy. There's just certain things we're not going to offload or that we should not offload.

Empathy, I sure hope is one of them. Like, the capacity to care for a person. Yes. let's get the actionable things and, and look, maybe I'm too conservative in this way, but, I I don't want to take away the meaningful connection between human beings. When somebody's ill, what I do wanna do is create more opportunities for that authentic, empathy driven interaction to happen by removing all the other stuff that gets in the way [00:29:00] of it. Um. But I, I like your point Bill saying it's not just a knob. You're right. This is not just one step, right? It's NLP listening to a conversation, structuring that data, sending it to a note, generating an order. that order leads to action, that action's automatically coded, which leads to rev cycle and, all of this stuff is gonna get chained together. And then where are the checkpoints of. Confidence. This is really gonna test our trust in these, in, in these

agents

for sure.

Bill Russell: it is, one of the first use cases I saw and I was reading about was the, uh, uh, problem lists.

Jacob Hansen: Mm-hmm.

Bill Russell: Or, or, or med rec. You know, this is the kind of thing that you could have an agent call a person before they come in. 'cause the agents can talk, right? So the agent calls ahead and says, Hey, just wanna verify some things.

And it's annoying to us as a patient every time we walk in and they ask us the same set of questions. [00:30:00] But if we've already answered those things and it can clean up the med rec, it can do med rec and it can clean up the problem list before I'm actually in the room with the doc. Well, actually it all gets cleaned up before the doc shows up anyway.

'cause it's usually an assistant that you're talking to. But um, yeah, and think about that is that, that is not like database level access. that's access at a, layer almost above workflow.

Jacob Hansen: you just reminded me of something else earlier this week. I was in a conversation. Adam McMullen, our CEO and I, were meeting with a partner and we were chatting about some of the benefits that have come through. Digitization over the last, you know, call it 15, uh, plus years, where at times there was too much variability in care being provided, not enough, commonality and structure.

And so treatment of disease was unique in one state versus another state. And it probably shouldn't have been. But now are we [00:31:00] gonna actually cross that chasm and go way to the other end? We're now, we're oversimplifying treatment because an agent only knows what it knows based on what it's seen in the past.

And it's gonna look for sort of the least common denominator. In what way do we actually leave behind too much variability and head to, a place where, you know, hammer sees a nail and smacks it, but it really wasn't the problem. Like I've sat with nurses so many times watching them work where they're sitting there and they're saying to themselves. I've seen this before.

Bill Russell: Jacob, I love the argument you're making. It's just not an argument anyone's made to here. Like, we're not, we're not arguing with you. We're not trying to intermediate, we're not trying to get in between the. Technology is not trying to get in between the clinician and the family member for empathy.

They're not trying to get in between, the diagnosis and that kinda stuff. That's not the argument we're making here. What the argument we're making here is that there's gonna be builders now who are looking at [00:32:00] workflow that the nurses are doing and saying, look, you do those three steps every time a patient is comes in, you do those three steps.

I can now build that, like I can go into Agent Factory and go step one, step two, step three, you know, bring in this, and do this

Drex DeFord | This Week Health: it

Bill Russell: thing.

Drex DeFord | This Week Health: for you.

Bill Russell: Yeah, I don't think, in fact, I know that Epic would never allow their agent factory to diagnose or get in between. they'll never do that.

Now, third parties. I mean, if we're talking open AI and we're talking, uh, well, especially open ai, I mean, they're flat out saying, Hey, we want to, all those things, your arguments you're making, they absolutely are positioning themselves to be an empathy machine, to be a diagnostic machine, to be all those things.

That's that. That's scary.

Jacob Hansen: Here's the thing. Officer is an AI company. we build computer vision ai. I'm in favor of automating manual workflow and augmenting [00:33:00] what human care team members can do. I believe in the value of doing that. I'm not opposed to it. I just also believe. in being methodical in the way that we

Bill Russell: Oh.

Jacob Hansen: I know we're just about out of time, but I just wanted to call out this last, thing I sent you guys about the Rent to Human site. Um,

Drex DeFord | This Week Health: Yes.

Jacob Hansen: I point this out because I just saw this yesterday, this new website built by a human, where it opens up to,

Bill Russell: Are you sure it was built by human?

Jacob Hansen: Good question. but it's a place where AI agents come together and they can literally rent a human to do something that, that the agent couldn't have done on their own. Well, in health tech, my first thought went to, oh my gosh. if, what if there was a really weird future where AI was trying to rent a care team member, rent a clinician, right? I need this done. I can see the data. I know what I need Clinician just show up and do [00:34:00] exactly as I tell you. that for, for me, even, I'm a product guy.

I love innovation. I'm motivated by it. I love solving problems with technology. I see an article like that and that gives me serious pause. It does.

Bill Russell: Well, we just had this long conversation about Claude Bot or whatever it's called now, Open Claw is the now thing. And, so I kept digging deeper into it after our conversation yesterday, Drex and Sarah. And, essentially it's a series of heartbeats and, CR jobs and whatever, and it's ai.

It creates this AI circle on top of each other with, by the way, infinite memory. 'cause it's just storing things in markdown files and whatnot. And so it appears to be sentient in ways like it's doing these things, but the reality is if you turn off that CR job, it's done. It is like, and, and it's not making great decisions all along the way.

Drex DeFord | This Week Health: it's not making great decisions. [00:35:00] Yet, like this is the worst. It's going to be how it is right now. It only gets better at an astronomical rate of speed. So I know we, you know, I like the whole empathy, thing. I don't know that there's gonna be enough humans for those humans to talk to all the people who are involved in the crisis of, not having friends and being lonely. we don't have enough people to solve that problem today. And I think there's a lot of great potential uses. Now, also a problem. Every one of these things is a, this could be really cool and could really help. And of course my next breath is always, and this could be really bad because people will get addicted to their AI friends and they will stop interacting with the regular humans.

Bill Russell: the next step.

Drex DeFord | This Week Health: the empathy idea.

Bill Russell: Exactly the, the next episode's gonna be three AI things that agree with me, and we're just gonna do a show. It's gonna be great. Oh man. Hey, this was a, a great [00:36:00] conversation. This will be a fun one for our editor. Landon will have to, figure out how to cut this down.

'cause I think we went 40 minutes and he's, he's a 30 minute kind of guy. So, Jacob, that's a testament to you. I love the conversation. Love where we took it. Love the content that you sent over. great discussion. So thanks everybody for being here.

Jacob Hansen: Thank you.

Sarah Richardson: Always fun.

That's Newsday. Stay informed between episodes with our Daily Insights email. And remember, every healthcare leader needs a community they can lean on and learn from. Subscribe at this week, health.com/subscribe. Thanks for listening. That's all for now.