Creating an AI Workforce for Healthcare with Olive CEO Sean Lane
Episode 1444th November 2019 • This Week Health: Conference • This Week Health
00:00:00 00:13:51

Share Episode


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

 Welcome to this Week in Health IT events where we amplify great ideas with interviews from the floor. My name is Bill Russell. We're covering healthcare, CIO, and creator of this week in Health. It a set of podcasts and videos dedicated to developing the next generation of health IT leaders. We wanna thank our founding channel sponsors who make this content possible, health lyrics and VMware.

If you wanna be a part of our mission to develop health leaders, go to the homepage this week,, and click on sponsorship. Information. This week we're at the health conference in Las Vegas and I was looking for use cases in blockchain and ai. In this conversation, I caught up with Sean Lane, the CEO of Olive, which is standing up an AI workforce for healthcare to drive greater efficiency and productivity.

the Health Conference, health:

So you've, uh, you said some pretty provocative things on the panel in terms of, uh, where AI is gonna go in terms of . Uh, underlying within healthcare. Why don't you give us, well, let's start with this. What's your background? So, oh, my background, I actually started in intelligence. I was an intelligence officer, uh, spent most of my career at NSA National Security Agency and, and really worked in technology there.

r as an entrepreneur in about:

And I wanted to look at healthcare and see if there's a way to solve some of those problems from a technical perspective. And that's what kind of propelled me into healthcare and, and led eventually to Olive. So your, your company's focused on healthcare. That's right. A hundred percent. A hundred percent, yep.

So RPA is a foundation. It is a absolute, uh, core component of technology that we use. And then you overlay AI on top of, on top of that. Yeah. So although RPA is a subset of ai, right? So yeah. RPA robotic process automation is a core component of it, and we really add a little extra juice to the computer vision side of that RPA piece.

And then we can do machine learning and deep learning kind of side side saed with the, with the RPA to give. All of a little bit more decisioning capabilities. So let's start with some of the things you said. So, um, how are we gonna see this sort of evolve within healthcare? Well, I think, you know, and one of the things that I, I like to say, just to get people really kind of shooken around, shaken around the idea of AI is that

Uh, within five years we expect, you know, 30% of a person's administrative workforce to be the, an AI workforce instead of a human workforce. Uh, so we think that is a pretty, a big shift that people will see. And within 12 months, every health system out there will have some kind of AI workforce. Per worker on their org chart.

Yeah. And, and it's not hard to see, I mean, when if, so I, I used to be ACIO for a health system, and you'd walk into certain areas and you just see this sea of cubicles. That's right. And all these people like looking something up over here and then typing something in, looking something up. But that's where RPA really thrives.

As a foundation. Yeah, absolutely. So what happened is healthcare has no interoperability layer and things don't talk to each other. Well, we are the interoperability. Exactly. So the humans are the routers, right? So when you look at that sea of cubicles, you see routers. So we all we're saying is, let's take the human router, let's make it an intelligent system, uh, like an AI worker or a software robot.

And then let's not only give it the ability to access systems and move data around, but to make decisions and to do things intelligently. That's, um, . It's fascinating because it worked again, it works. 7 24. That's right. 365. And it's not de dehumanizing type work. Exactly right. Uh, so we're taking that, we're moving that to computers where we, where we seen it be implemented.

At this point, and I'm gonna come back to computer vision ai, we're gonna. Sure we're gonna talk through those. 'cause my, the audience is predominantly health IT staff and individuals, and they're, they're gonna want me to parse that. But yeah, so whenever, um, whenever we're coming up with a solution, a lot of what we do is we see where everybody's headed and we try to go the opposite direction.

So we saw a lot of AI headed towards clinical use cases and people thinking that's where we should start. And we from that, took a cue to go into hospital operations first. Uh, we see a lot of low hanging fruit there. . A huge amount of resources that can be freed up and those resources will make a huge, a massive impact in clinical operations and clinical care and new therapies and new drugs and patient treatment.

Uh, but we are starting with the operational side, so revenue cycle it, finance and accounting, hr, supply chain. Um, a lot of those areas where we can take, you know, massive amounts of costs. Out of the system. Um, so that's where we are today in, you know, phase two will be in the clinical side. So what does it look like to, to implement?

So, um, I, I bring in your technology. I'm looking at those workflows and those processes and I'm going, oh gosh, I'm going outside my, I don't control some of these IT systems others I do control. I gotta get some of this information back into the EHR. Mm-Hmm. . Um, how do you, how do you layer in. So we have a, we have a very, uh, specific process that we have created and we have, uh, shaped to be very, very, um, responsive to a health system.

And every health system is different. So the, the system that we have created is a kind of A to Z step. We call it from alpha to Omega from beginning. To the kind of eternity of the, of the digital, uh, worker. Um, a set of processes. And the processes start with understanding the work, identifying what, what work we should be automating, and then un we, we do a scoping process to understand the economic impact of that work.

And then we start to learn how to do the job itself and really understand and map out. The processes. Um, but the cool thing about a an AI worker is that it doesn't stop. It doesn't get put, you know, it's not software that gets built and put on a shelf. It learns just like a human does always. It's, it's constantly learning.

It's constantly growing and evolving. And that's where the machine learning and AI comes in. That's right. So if day two and beyond, if they're starting to . Submit bills and starting to get errors, they learn from that. Exactly. And they go, yep, we've gotta do something a little different. Yeah. They learn, they adapt, they change.

his is actually in, uh, about:

And we were able to create alerts before they actually completed them. But this is like the next step, right? This is like, this was alerting a person to say, Hey, do this a little different. And then they go, oh, okay, well I gotta figure how to do this a little different. What you're saying is, this is completely automated.

The computer is now . That's exactly right. So it's a very important piece that you just, that you just mentioned, is we don't want to put a brain in a jar. We don't want a great idea that is derived from a machine learning or algorithm or a deep learning model to basically be an alert for human. We want to take action.

So we want the, we want the AI worker to, to act upon. . That data, those, those, you know, things that they're finding out about claims and why they're denied. We want the AI worker to actually change it, to make the, to make the change in the system themselves. Themselves. So if I'm with a health system that's looking to reduce costs, and we heard, uh, large health, we heard Amy Compton Phillips from Providence talk about, you know, two, 2% business and those kind of things.

Um, is this an area where they can actually look for cost reductions? Oh, uh, yeah, every single time. To what magnitude? Yeah. Um, I would imagine that, you know, 20% very, very easily without too much, um, effort that goes like beyond 20% of a certain function. Yeah. Reducing that, reducing the cost of that by at least 20% initially like that, that's a very easy kind of back of the napkin and what you can think about of any organization.

But that, but that's just the beginning and. I mean, who knows what the, what this will bring as far as cost reduction inside these functions. But I think the sky's the limit. I mean, there's so much of that can that can be automated. You can imagine that, you know, software robot or a digital employee can do most of that work.

So you, you mentioned, uh, implementing within health it, what, where would you, where would you utilize it within health? it, so, uh, account creation, um, determining whether or not certain parts of. You know, um, accounts are, are not working appropriately or correctly, like looking for error messages, for example, and responding to those error messages, looking at error log, helping on help desk.

Um, you know, there's lots of humans that are working help desks right now that can be, a lot of, most of that can be automated. Um, so it's anything that really a human's doing that is, so you can take a ticket, a ticketing system gets generated within whatever, ServiceNow, whatever, and you can take those things and be the front end.

Yeah. And then escalate based on certain criteria. Sure. If, if that particular process, um, you know, created enough impact to the organization and we could do it at the right amount of speed, that would certainly be a candidate. Gosh. And account creation. I don't, I mean, do you have an idea of how complicated that is within the hot health system?

Oh, absolutely. . I mean, we do it, so, yeah, absolutely. Well, it is, it was unbelievable to me how Yeah. Um, you know, when I first got there it was like, well, you know, it takes us about a month. Well, the problem is the person started three weeks ago. Right. And you're still sort of working on it. But then it got more complicated 'cause we would have, uh, some of our hospitals were, uh, union and so we'd have strikes and we'd have to bring in like.

Just this whole massive team, and you'd have to crank that stuff up so I, I could see where the automation would really, uh, would really fit. Yeah. Uh, really well. Um, AI computer vision. Where's computer computer vision fit into this? Well, it's a, it's a enormously important part of. The automation itself. So the AI worker needs to be able to look at a screen, look at a software user interface, and is if it changes, if a button moves or something changes on the screen, it, it can't break the system.

It has to be robust. So this is beyond screen scraping. Oh, absolute. Which, which was very fragile. Yes, exactly. So computer vision is, it sees the same way I do. That's right. So computer vision gives . RPA or this automation, a layer of robustness that it makes it not fragile and it makes it enterprise grade.

So now it can go in and do things at scale and you know it's not gonna break just because a button moves or something changes. So it really is looking at a user interface and understanding what's on it. You know, how to enter information, where to enter the information, what that, what those buttons do and what they mean.

So you, it could adapt as the software changes you're go to market payers, providers, just, just providers. Just providers, yes. Okay. Yep. It's interesting. It's, it's, it's only interesting 'cause almost everyone I'm talking to now is trying to hedge their betts and be in all Yeah, we're all in on the, on the providers right now, on the provider side and, and do that.

Uh, any specific use cases you want to highlight? 'cause I've, our listeners are always asking me before, you know, where's it. Where's it happening? I, I think that, that, you know, we're, we're in about 500 hospitals. Um, olive is working every single day. She shows up for work every day, does her job extremely well, and gets smarter over time.

That's one of the things we say. Um, and she's doing work in mostly hospital operations, but the things we start with almost every single time are around claims. So claim statuses and denial management. Um, prior authorizations and benefits verification, so insurance eligibility with that, those three, uh, functions,

We understand what a health system is doing every procedure that it's, that it's doing for a patient, and if it's getting paid for it in a timely manner, and if we know that we can kind of branch out from there and do many, many things. That's exciting. Yeah. Thanks. Well, thank you very much for your time.

Absolutely. Hope you have a good conference. Yeah, you, you too. Thanks. I hope you've enjoyed the conversation. If you would like to recommend a guest or someone to be on the show, you can do that from our homepage. Uh, recommend a guest is about three quarters of the way. . Down on the homepage. Please check that out.

And don't forget to please come back every Friday for more great interviews with influencers. And don't forget, every Tuesday we take a look at the news, which is impacting health. It. This show is a production of this week in Health It for more great content, you can check out our website this week,, or the YouTube channel, which you can get to from our homepage as well.

Uh, if you get a chance, take a look at our newly redesigned guest page. I think you'll find it. Uh, fantastic. I love the way you can, uh, navigate through the content. Special thanks again to our sponsors, VMware and health lyrics for choosing to invest in developing the next generation of health leaders.

Thanks for listening. That's all for now.



More from YouTube