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

[00:00:00] Today on Keynote

(Intro) what I'm trying to tell the team is be resilient, be curious, be adaptive, to whatever is going to happen to us, because inevitably something will, and we just want to be in a position where we can pivot.

My name is Bill Russell. I'm a former CIO for a 16 hospital system and creator of This Week Health, where we are dedicated to transforming healthcare one connection at a time. Our keynote show is designed to share conference level value with you every week.

Now, let's jump right into the episode.

(Main) All right. It's keynote. And today we're joined by Dr. Alistair Erskine, the enterprise chief information and digital officer at University in Emory.

Healthcare. Alistair, welcome to the show. Thanks for having me. Yeah, I'm looking forward to the conversation. always looking forward to any time we're in the room together. like the fact that I didn't have to send you the questions ahead of time. Probably scares some people, but we're just going to have a conversation as [00:01:00] former practitioner.

I'm getting farther and farther away from it. But somebody who's sitting in the chair. going to start with Your most recent title enterprise chief information and digital officer. It's something that really caught my attention. You're bringing together the university and healthcare IT could be so bold at a time when a lot of. Others are splitting them apart, and I'm curious what made you Emery swim against the current a little bit on that one?

great question, and there's a lot to unpack there. would say as a physician, get in the technical space and ultimately getting CIO job.

That was a bit unusual. Most of my colleagues don't go into this space. Then when it came time to take on the oversight of technology at the university, it turns out it didn't work in my favor because it was more that health care was going to take over. What the university was doing. So I had to bring out the non healthcare aspects of my training MBA, philosophy major, those kinds of things to be [00:02:00] able to make the case that no, I can actually speak to all sides when it comes to academia science and research and healthcare delivery.

I think the reason that yeah. Emory is choosing to go in this direction is actually very deliberate. I think we believe that whereas it makes sense for some organizations to split off because the Kager is very different. Healthcare may be 15, 20%, year over year growth. Whereas a university may be one or 2 percent year over year growth.

So at some point, the health care side of the organization is getting very large and relatively speaking to university, and it may require a different cadence of change and, adapting to the newest, technology needs, not to say that the university doesn't have similar, but it works a little bit at a different pace and requires In my early experience, a lot more consensus.

Drive in, then you do what is perceived as being more top down on the health care side. But at Emory, we really [00:03:00] wanted to actually use that as a competitive differentiator. The fact that we have a school of business, a school of law, school of medicine, school of nursing. Even the School of Theology, all of which can participate in the delivery of care, and all those clinicians that are doing that can participate in the teaching of the health domain in those relative schools.

And that we have a health science center, which is a combination of school of medicine, nursing, and public health with some important labs like our brain health lab that does a lot of Alzheimer's work and our global diabetes lab that does a lot of, diabetic work and a primate center. Makes for a pretty interesting ecosystem that kind of goes back to really what an academic medical center was designed to be.

It's interesting. In my previous life, I was a CIO for a university and I was a CIO for a health system, and I'm curious, do you accentuate what's the same about the two? [00:04:00] Or do you accentuate what's different about the two? focused on experience. Either way, you have patients, you have students,

you have business processes around it. You have research processes. You have teaching. There's a lot of similarities, but there's also some differences. do you accentuate?

I tend to accentuate more what's similar, because the knee jerk reaction in engaging is a fact that we're all so incredibly different.

When it comes to things like technology, whether it's thinking about enterprise architecture service desk project management. DevOps device support, those kinds of things are really just not that different. People need computer or a keyboard. They need a software, they need, data and so forth.

Those parts are very similar and helping people understand that can come from a core function. Within I. T. That can be distributed to all whether it's academic health science or health care. And in fact, that's how we've organized shop. There are really four pieces of every digital. One is the [00:05:00] core and then the other three are health care, health science and university.

So the health care side manages everything. Health care, health sciences, where there's a combination of research and, schools of health. And the university is your standard, university, both professional schools, undergraduate schools and graduate schools. But the core is made up of what you would expect a chief technology officer, a core CIO a data officer, and then the office of CIDO.

So Those functions are all the functions that hydrate all those different domains. you see,

as you're looking at innovation, do you see the innovation in one touching the innovation on the other side? have ambient listening and we have computer vision. We have all these things that are emerging on the healthcare side.

curious. You talk about the different pace at which things get adopted, but I'm curious if the innovations starting to seep across.

Yeah. In fact, it's usually at the intersection between two kind of, arbitrarily chosen [00:06:00] domains that most of the innovation pops out. It is those things that we do in health care that are more routine that are new and different in the academic side and same thing on the academic side.

Things that are routine that are more unique and different on the health care side. So we can look at each other's ways of practicing and identify, opportunities and new ways of thinking about addressing an old problem. that if we think about it just even purely from a data perspective, guys.

They may be data eluded from equipment in a wet lab and DNA sequencers. And some other aspects of research that may occur on the health care side, there'll be a electronic health record data both structured and unstructured. And when you pull back. And if you're able to say, okay, this one patient had this EMR data, had these clinical notes, had that genetic test, had these various different findings, in other omics domains and digital pathology and imaging domains, and I can tie this all back to one person, I can [00:07:00] take that and put it into a GPT analytic environment like a Microsoft, data fabric, and I bet I'm going to stumble upon new things people haven't seen yet.

I bet I'm going to get to early diagnosis in some cases, and I bet I'm going to get to new diagnosis that don't exist today, just from being able to look and thread through all that data with a GPT. Despite the fact that all the data is very different one from the other, but it still ties back to one human.

think you'll appreciate this. So I took all your interviews for the last like year and a half and I put them through essentially a large language model like, I don't want to ask the same questions everybody else is asking. And it was interesting.

I'm gonna put some of this stuff in front of you. And I guess we'll find out it actually made mistakes or not. As we

great. I'm glad didn't say just passed. Don't talk to this guy. He's got absolutely nothing. He's all over the place.

I do wanna follow up on something we talked about in our last interview.

First, it's the rounds the it fascinates me that you did this and streaming rounds approach. [00:08:00] And it's cool 'cause it solves a couple of problems. Remote work teams, connecting those teams, frontline insight, all those things. What I thought I'd do a little different on this one is take me inside one of those rounds.

What is happening on one of those rounds?

actually a perfect example because we rounded in the student campus life environment. So I took the round in. And took it to the university because I wanted to do the same thing. Basically, it's my way of being able to understand what's happening in the front line.

It's my way of being able to help my Emory Digital new team that is made up of, lots of university folks and lots of healthcare folks together. To be able to understand better what the life of a student is. So the way it works eight o'clock in the morning, we show up my assistant, grace bless her, is the camera woman and she has a gimbal with an iPhone on it and she has these little remote microphones that we all clip onto ourselves and to our guests.

And, we have a bit of a sense of, okay, we're going to have a conversation for half an hour, 45 minutes here about, the general topics, [00:09:00] and then we're going to go round. And look and feel and talk to students, talk to faculty, whatever it may be. So in this particular example, we sat down in a room and talked about the student tech support that they have, the video production studios that they have, the kinds of things that students run into in terms of

when they don't have enough money for a laptop, how we can get them laptops and, get them connected to the Internet, get them going, if ever that's a need and so forth. That conversation usually is several people at the table that support that particular area with members, I would say, quote, unquote, of the business, meaning whoever's responsible for that domain, talking about how technology helps them and how they wish they could do something better and how they, Work on canvas, which is one of the tools that are used in the university to keep track of all the classes and the student life and after kind of running down almost like a bit of this is our day in the life.

And this is the things that we worry about. Then we all get up and because we have these remote microphones that we put on to each other, people can hear the conversations. [00:10:00] Caveat. Watch out for the hot mic, and then we start walking around and so like we may show people. Oh, this is where the 3D printing occurs.

This is where you can design your CAD drawing and create, something. Let's go to the video production studio. And then you can see. Like the little room that has, as you would imagine, one or two people in them doing a podcast, and then you have literally the studio with all kinds of things hanging off the wall and lights and cameras and this clever thing where you can write on it and it's transparent and the camera can turn the letters around so that when you ultimately end up showing Yeah.

You know the video. It looks like somebody's writing on the screen, but it's right in the right way around. So things that I would have never known about, we typically ask the folks when we're around who comes here? Who's asking? In the classroom environment, the teachers to actually run production.

Like, why do they need that? Is it for didactic purposes? Is it for marketing purposes? Is it for, something else? So we try to understand why these resources [00:11:00] are being used. If they have enough of what they need to be able to accomplish what they're trying to do.

You'll get things like, oh, I wish we had more space or, this equipment's getting a little bit old. There's a little bit of a pitch to try to, get some resources or some space or some money for something. And then we walked around, the libraries. Emory has some gorgeous libraries.

And, students, what do they do? They come to school, they go to class, and they spend a lot of time in the library. At least the ones that like to study. And they study there and so there's a good intersection between library services and technology services for students. So the sort of genius bar for technology is embedded in the library.

So if somebody has trouble with their laptop, they come there and they either fix it right there or they do a swap out and get a new one. So that's what the rounds in a two hour period of time. It's a conversation with the leaders. It's going around and talking to a couple of students and a couple of faculty in their environment to better understand the context.

And [00:12:00] then two hours later it's over and there's a summary that spits out from the system that's been listening the whole time and generates a summary and usually I have to correct a couple of things and then I send it back out. There's also a laundry list of follow ups that occur because as we're having conversations stumble upon things which They're like, wait, I didn't know that happened.

Like we need to fix that. And so You know there's a few things that pepper onto the roadmap, whether it's short term projects or long term projects that manifest from that.

We had 16 hospitals and I would say in about two thirds of those, the IT staff did rounding and typically the IT leader did rounding and then they would generate reports and they would come up to me.

But I would imagine that's not nearly as powerful. I'm curious, how many people are tuning in. What things have you looked at? Insights or what changes have occurred because of people actually being connected directly into the round.

Yeah, there's a few things. First of all, from a culture perspective, I'm trying to model the behavior that I really want the rest of the digital [00:13:00] team to engage in.

In other words, be with the customer, put yourself in the shoes of the customer, understand what they need and take your expertise, your knowledge. Your talent and be verbal about the fact that you can see an improvement. If we were to implement this or add that or whatever it may be, that's number one, modeling the right behavior.

Number two is what's happened almost, by itself is the hundreds of people who are listening. So I usually have, I would say about 150 and it could be mostly digital team and a few others that just. Heard about this, and they want to understand what's going on at the library as well, so they tune in listen in.

Because you could be doing something else , while you're hearing this, and you don't have to have 100 percent attention on it, but some people do. But what happens is, as we stumble upon a problem, somebody will chime in, that's listening, and say that's my domain.

Actually, let me go ahead and make this quick change. Okay, try it now. What do you think? That is a really [00:14:00] fun moment when that happens. It doesn't happen every time, but it happens occasionally. And you can tell you've got this army of people behind you, all of who are trying to just solve problems, problem solvers.

And you could, highlight a few problems and they'll go ahead and fix it in real time.

interesting. I'm handing off more. I have a really small team now. I used to have 800 people. Now I have 15 and I'm handing things off and it's the same things keep happening.

You had something off and you have all this knowledge and then people go I'm recreating this. What are you recreating that for? You could just do this. They're like. You mean I could just do this? It's yes, you could just do this. Have you seen those kinds of moments where people were like, what's going on?

That's not what was intended.

I think that, the team tries to be open to the fact that you have a plan. And then there's reality and that the plan needs to adapt over time to reality and it was, it's going to constantly move, one of the things that happens when the team is oh God, that was a really tough year.

We did a lot of stuff who can take a break now. Yeah. No. So don't wait for the [00:15:00] change to stop because you could be waiting for Godot. It's going to be something that is it. Always changing. Like we had three major hits to the system last year. We had CrowdStrike, we had change healthcare, we had heavy fluid shortage.

I can't wait to find out what are going to be the three to five major changes this year so what I'm trying to tell the team is be resilient, be curious, be adaptive, to whatever is going to happen to us, because inevitably something will, and we just want to be in a position where we can pivot.

Our center of gravity so that we don't get taken off guard and off balance.

Captain, our beloved service dog, is back for his third year at Vive and HIMSS. This week, health is partnering with Alex's Lemonade Stand Foundation to raise 200, 000 this year in 2025. It's easy to participate with this. Snap a photo with Captain, share it on social media, tag us, and for every smiling face in the picture, we'll donate 1 to [00:16:00] support these families battling pediatric cancer.

A huge thanks to our partners Rackspace, and SureTest for making it possible. Join us and let's make a difference together. Visit ThisWeekHealth. com to learn more.

I saw something in one of your interviews and it was like in the intro part that you served in the Middle East standing up a hospital. if I read this correctly, it essentially said we had unlimited budget and we made unlimited number of mistakes.

Like, it was something, I have misread that, but can you characterize that experience and what that was like?

So I'll preface it by saying that a week ago I went to India as we spun off an offshore, a group that works, 50 people that work for Emory. And we can talk about that later, but on the way to India I stopped off in Doha in Qatar.

And was able to get off the plane, spend 10 hours in Qatar and then get back on the plane, you were able to leave the [00:17:00] airport. I went straight to this hospital called Sidra Medical Research Center because I'd never seen it in the opened state. While I was there It was always going to open in six months, and it never had actually opened, and by the time I left, which was 2013, it turns out it opened in 2018, so it's probably a good thing I took off, because it would be another five years before it actually opened.

This is a unbelievable, beautiful building, like the marble is from Brazil, the teak is from Thailand, you got this just, seven story high, quarter of a mile long. hospital that looks like a boat with healing gardens inside and waterfalls and tvs that are in the ceiling so if you're laying flat you can look at the ceiling directly you know and all this stuff it's only got 400 beds it costs like 2.

5 billion to build. So if you can do the math, it's a lot per bed, right? And when I was there, they only had 172 beds occupied. So when I say the budget is really not the primary thing that they're worried about. What they were [00:18:00] really worried about is that they were tired of sending Qataris out to the West to get proper clinical care.

And when it came to women and children, they wanted a place where they could, be local and have great care locally. So my favorite story was we had to buy a physiological monitoring system for the hospital. And so they said, Alastair, go figure it out. Yeah, it's okay. Great. So I go off and I'm like, okay, this Phillips and GE and, space labs and stuff like that.

And here are the pros and cons of this one. Here's the pros and cons. And they said no, just get them all. I said hang on a second. You're going to want to pick one because otherwise you're going to have to stop worrying about the money, right? Just get them all. When it's here, we can figure out which one we want to install.

That kind of thinking, it took me six months to get used to. And and the idea is they're after the best. They don't care much in the, money's coming out of the ground to some extent. When gas is less expensive than water you make different decisions. About energy, you make different decisions about, parties of things

just out of [00:19:00] curiosity.

How would you approach that differently? Because as an I. T. person, get all three. You don't need to get all three and bring them in to figure it out. I don't

think you don't. But again, here's another. It's not an I. T. thing directly, but the building was being built And at one point I go to the emergency department and I look around I'm like where's the cat scanner?

And they're like no, it's down the hall a quarter of a mile away and they had The sedation for the kid. These are like pediatric is sedation the kids halfway there So the idea was going to be that somebody comes in with a traumatic brain injury, right of some kind they get wheeled An eighth of a mile goes to get sedation, get wheeled another eighth of a mile, get to the head CT spin, and then come all the way back to the emergency department.

So I said, yeah, no, that's like an unstable patient. You don't want to do that. It's going to be way too dangerous. They don't have a problem. They go in and bulldoze. out that part that had been built and rebuild something that can then fit a CT scanner. So instead of making the [00:20:00] changes on blueprints, it was easier to conceive of things after it was built.

And if it needed to be changed, you just bulldoze it out and then, put new things up again. We would never do it that way, because it would be, cost overruns and it would be inconvenient and so forth. But, those are some of the examples of the change. And yes, when it came to three physiologic systems, it's you really don't want to do that.

And here's why is what the experience is going to be of the patient. If they have to have the nurse connect them to three separate systems and they're going to look like, server in the data room, as opposed to a patient. And I, you get the point across.

Yeah. So how has that shaped you?

Like that's a very unique experience that very few people have.

So I would say, a lot of times I hear people say we don't have enough money to do X, Y, Z. And that's why we can't move forward. And that's the problem. It definitely made me question, a couple of things.

First of all. The lack of money isn't necessarily the real impediment, and in fact, sometimes you can start [00:21:00] making really bad decisions if you have too much money and too many funds, because the thing is, that occurs in, U. S. healthcare is a healthy prioritization that avoids a scenario where you've got this really complicated architecture with a bunch of one offs all over the place,

for efficiency's sake, for financial's sake, we have to build things on platforms. We have to build things that are experience led. We have to rationalize our applications. When you don't have that constraint, then you're free to have something which ultimately becomes really hard to support.

Because you need to have somebody who understands that. Oh, over here. They use this application over there. They use that application, even though they do the same thing, it's almost like you're building a house without an architectural diagram. So I think that's one thing. I think the other thing is there's so much waste in all of our systems today.

Emory as well. I'm sure that we still have, even though I've gone through combing through our budget, taking software products that we don't use anymore because there are no users that are actively logging [00:22:00] in and archiving those and stop paying the license and fees. You can often find a ton of money in waste.

If we believe the fact that there's a trillion dollars worth of waste in healthcare then there's a portion of that in our own systems. You can generate a lot of revenue just by addressing the waste. So I think that's where the, we don't have any money. Strikes me as, it's an opportunity missed potentially,

to answer this question briefly because I want to ask a different question.

But how many physicians are currently enrolled on your ambient listening solution.

So we have 2000 physicians, but we only have about 650 who are using it on a sort of daily basis.

650. The reason I want to bring this up is your approach. This was different. I was looking for things that you do different.

I want to highlight those things. So the decision you made with ambient listening was. Most organizations would start with a small pilot, test it carefully, expand gradually. But you made a decision, at least from here, it appeared that you made an enterprise wide decision from the beginning to make it available to [00:23:00] everybody.

Take me inside that decision that moment.

So this was very well, I should say relatively earlier on in the process was like June of 2023 before there was a big explosion and everybody was talking about AI scribes and I understood what the technology did. I had followed it, but it hadn't gotten to this level of sophistication.

DAX express came out. And understanding what the technology did, it struck me that this is something that most people are going to benefit from. Not necessarily that most people are going to go after and want, as I'll explain the difference. Knowing that, what I didn't want to do is constrain that project by me premeditating who was going to benefit from it.

I wanted there to be enough degrees of freedom where, and in fact that's what we did, we spread it to 20 different doctors, all from different specialties to start. Then we measured who was actually, in adopting more than the other, and they went to those departments and said, okay, you guys can be all in for those departments because the one or two [00:24:00] doctors, are using it a lot, therefore we know there's a potential value.

Then after going to those departments, we went through the process of embedding it into Haiku so that it was really easy to get to, you didn't have to do anything extraordinary. And once it was in Haiku, it went to everybody. And now the only barrier to entry. So now we have about essentially a third of the doctors that are on it.

I find that at least for ambient, for us, our experiences, one third are going to go on their own. The second third are going to have to go on by coaxing them to go on. And the third are not going to go on because it's really not compatible with the way they practice, like an ophthalmologist doesn't talk a lot to the patient.

There's nothing to record. They're looking in the eyeball and they're doing all kinds of tests. There's not a lot of verbal stuff that's coming out, so it's hard to capture. So I think we're now having completed the first third. We're going after the second third. We're now actually on a campaign.

We're looking at people who are staying up until one o'clock in the morning, trying to do their documentation. and proactively going out to them. We're going to start pushing it.

[00:25:00] the bet to go enterprise. So you had to license it enterprise wide.

I licensed it with tiers.

we tagged it to the attendant physician. So whether you're a resident, a nurse, whatever. All that didn't count, only if you were an attendant physician, and it was like a thousand, two thousand, and above two thousand. Got it. And until you got to those levels, you were paying whatever tier that you were paying.

That makes sense. And at the time, we didn't really know how much this thing cost, right? So it was too early to have really found its point. And here's the other thing that was really important. We paid a license fee, but what we really wanted to do with our particular vendor is we, Went all in with the co development.

Like we were the test bed. We, they would come with their engineers. They would watch what we're doing. They go back and code some more. They come back. We would provide feedback on nine and neurosurgeon really doesn't say things that way. They say things this other way. This is the language they use. They would tweak tweak, tweak.

We wanted to be. Part of the team that helped give very [00:26:00] concrete and frequent daily weekly feedback so that the product got better because it only just helped us and everybody else who was using that product.

Do you look for those co development relationships? Or just every

single time every time.

In fact, the approach is to say, what can I do? And this is not the typical CIO because I've seen. Sometimes CIO is going to beat up the vendor. To me, it sounds totally antithetical. If you want to get a good deal, of course. But at the same time, it's like, what can I do to make the vendor better?

Because if whether it's investing in them so they can, and through arbitrage by the fact that, a large academic medical center with a great name is choosing them is going to make the valuation go up. And then, As other people use them, the valuation is going to go up.

So it'll be good for them. And it'll be good for the organization that invests in them. But also if I choose a product, I want it to be as great as possible. What motivation do I have not to try to help?

Yeah, no, absolutely. But I think I could make the case that for ambient [00:27:00] listening is 2025 for computer vision in the room. But how do you approach it differently than maybe you did ambient listening?

The thing about ambient listening is, people tended to have an iPhone, we already had Epic, you had to integrate it into Epic, you had to do some practice transformation, but it didn't require a lot of installation of stuff, right?

It was just use your iPhone. Here's the app. Now you can go talk and it, this, your note wind up in Epic. With AI cameras, you're talking about a whole different ball of wax because now you're taking that room out of commission. Oh, by the way, it's the winter, not a great time to take rooms out of commission when the hospital's full and the patient's in the emergency department.

So that you can install not only the AI camera, But the power feed the cabled ethernet cables, everything else that you're going to have to support these AI cameras at scale. And then you have to go through a process of [00:28:00] teaching, changing your admission consent to treat changing the way that the nurse introduces the patient to the room by letting them know this thing is on 24 seven.

Don't worry. The LiDAR sensor can't tell who you are. This is what it sees, right? So it's private because LiDAR doesn't find any identifiable things. But if we want to turn on the camera, you'll get 45 seconds to say yes or no. If it's an emergency, if you're having a heart attack, don't worry we'll blast through it.

But there's a lot of nursing and patient expectations that, what do you mean you got a camera in my room? Are are you monitoring me and yes, you're sick, you're in the hospital. We want to make sure that we bring the best care to you possible and we want to make sure AI is right next to us so that if you're about to fall out of bed, it tells us before you get hurt.

And so there's a number of things, practice transformation across many different clinicians, not just one doctor and one iPhone and facilities involved in the installation. ,

It's a bigger capital lift for sure. Now that I think, it's interesting.

One of the things when we're talking capital [00:29:00] I talked to Keith Perry, they're building a new building and fascinating if you're lucky enough as a healthcare CIO, you get that opportunity to dream and build that building. And it's fun, but the thing that keeps you up at night.

Famed question when you're building a building is, oh, gosh, is it three years from now? Am I going to be having to call the maintenance and bring new things in and that kind of stuff? Computer vision in the room is one of those things where it's like has the tech shaken out yet Or is it still shaking out?

I think it's still shaken out. think, most hospitals in the US don't have ambient. So I like to think of this category is ambient technologies. So ambient audio, ambient video, ambient internet and ambient location services. Those are the four technologies that really We need to bring ourselves into the century to be able to enjoy ambient audio is just, and you're listening, but in your videos, these AI cameras that do more than a one trick pony, because I can [00:30:00] do virtual nursing, virtual sitter, virtual family visits, virtual medical interpreter, virtual, handwashing.

I can use a LIDAR for seizures. I can use a LIDAR for pressure ulcers, for falls and the chips that these things carry are on the edge with a chip now that can handle 36 teraflops the language models that are on them only used eight teraflops.

There's so much room on those edge compute devices to be able to add additional algorithms that do clever things that we haven't even thought of yet, that this idea of the network could go down and the thing still works. Because it's all being processed on edge. And so having that ambient video in the room, then think about having that ambient video in the OR for an OR black box to be able to say, did timeout occur?

With a sponge count? Correct. How many times did the room. Get interrupted and the sale feel get interrupted by somebody coming in and out of the room because they didn't have the right equipment because they got the wrong tray, because whatever that [00:31:00] may be, and when something went wrong, can we go back and see if we reconstruct what the issue was and learn from the mistake so it doesn't happen again.

We are going to be able to tell people's facial expressions, how they feel, how fearful they are. There's already technology out there, another, the algorithm that you could stick onto that edge compute chip that looks at the patient's facial impressions, to be able to get sentiment and feeling.

So you know, if you're walking into a patient's room where they're either anxious or whether they're angry or whether they're, not worried when they should be or whatever it may be.

That was the hard part of the interview. There are things you say in the 229 meetings and other places that I'm like, man, I want to make sure that gets elevated.

So if somebody listens to our conversation, I want to make sure talked about this trickle feedback and I found it really fascinating in. Very manageable and and I think it's a way for the voice of the organization to really be heard all year round and to really have an impact.

Talk a little bit about that approach and [00:32:00] how it leads to culture within your organization.

So just the end outcome of this is that you get a dashboard that you can slice and dice data. And find out how do all the nurses at this hospital feel about a system that's the general sense of we now call every engaged, but it's a trickle feedback concept.

So that's the product. Now, what's the process to get to that data? The process is every day we take. The entire corpus of users at Emory. So let's say 50, 000 users and we divide that by 365. And we take that proportion one 365th and we send a very straightforward and simple survey that has one through 10 stars.

And then an open comment box. And it's if it's an Epic, it will say, does Epic help you do your job? And then the comment boxes tell us anything else that's on your mind. And what happens is 40 percent of the time, somebody will give it a star, whether it's an eight stars, 10 stars, whatever it may be.

[00:33:00] And 20 percent of the time, they'll start entering information into the comment box. Sometimes it can be paragraphs of rant, frustration they may have. Sometimes they're very happy. Oh, I've been to another Epic client. This one's the best one I've had so far, it runs the gamut. But because they're entering text and they're providing numerical value, almost like an NPS score, you can then.

Because they're logged in, who they are, what department there are and so forth. You can start slicing the dice, the data by role by facility, location, et cetera. And to the person that's sitting down on the Epic team configuring ASAP, for the mercy department, they just look up the mercy department users and they start reading through all the comments because it's excluding everything else and only showing you the comment for.

emergency department. It could even be down to emergency department in this particular hospital. And they get a sense as to what the front line is complaining about. So the challenge I think that we have is because There are 50, [00:34:00] 000 people. We have to function with a bunch of filters and interpreters that collect that information until eventually it makes it up, to the point where we can get organized and work on the thing.

It's going to be the most effective, but in that process, we lose a little bit, the voice of the customer. So this is a way to be able to completely unvarnished voice of the customer, try to capture that. And then the clever thing is now that they've done this, we can use AI to respond. the team has figured out now is when they get a comment, they can do a sentiment analysis on the comment.

If the comment is a little bit negative or very negative, there's an automated response that's generated with a team that oversees it and then sends that out to the person who wrote the negative comment and guess what happens. Almost inevitably. I can't believe you guys actually read what I wrote. This is amazing.

And all my years I can, nobody's ever, and you just create this really good sentiment that you're actually listening to what's being said and you captioned voice to the customer so that there are a [00:35:00] number of, so trickle feedback is because you're getting input every single day. So that NPS score that those.

50 or 60 people are picking, it gives you a number like 2, whatever, and that changes day by day as things occur over time. You want that to gradually trend up, but if you put a big upgrade in and suddenly drops by two points, you've screwed something up you're going to hear about

it.

So I'm only going to get one survey a year. Okay. And

just two questions, simple, not 500 questions once a year, right? That's the way we survey people today. The idea is to make it super small and free text. So the right, what do you want to write? Not what you're wrangled and forced to answer.

In large language models just made your life a lot easier with that, which is pretty, look, I've sat on survey teams before and, we're trying to get the right wording and stuff. And, 8 to 10 weeks later, we've argued over should that be and comment and [00:36:00] or comma or whatever.

And you're saying you can come up with just 2 questions that get you the sentiment that you want.

That's how it works. And we use that to be able to provide evidence, for example, to our experience councils of this is what we're hearing from the end users. And they can look it up themselves and it's the themes are pretty evident, people will complain about login and people complain about complexity of the system, they'll complain about slowness, like thematics come up and you can address those big thematics in a clever way.

But it's that individualized response to the people who are initially unhappy that really blew me away when I started seeing the positive feedback that came from somebody who was not happy.

Alistair, this is one of those occasions where I'm sad I'm not Joe Rogan, cause we'd have another after the commercial break.

We'll be back for our next hour with Alistair and we'll just go into more details. We could talk bunch more, I'm sure. And I look forward to catching up with you again. I want to thank you for your time.

It's my pleasure. Always a pleasure to talk to you, Bill. I appreciate it.

Thanks for [00:37:00] listening to this week's keynote. If you found value, share it with a peer. It's a great chance to discuss and in some cases start a mentoring relationship. One way you can support the show is to subscribe and leave us a rating. it if you could do that. Thanks for listening. That's all for now..