Disconnected digital experiences are a thing of the past. AI technologies like machine learning, natural language processing, predictive analytics, etc., allows health systems to have an integrated consumer engagement strategy. With the abundance of AI, how do healthcare leaders make sense of how and what AI can be applied to their organization?
Listen to Brian Gresh, President at Loyal, and Chris Hemphill, podcast host of Hello Healthcare, as they discuss how AI can help drive new strategies and solve business challenges, including integrated and proactive consumer engagement.
Healthcare AI University Series
This conversation is brought to you by Actium Health in partnership with the Forum for Healthcare Strategists.
Subscribe to receive emails when new episodes are released.
Thank you for subscribing!
Please check your inbox to confirm.
Brian Gresh (00:00):
Those are the kind of things, sometimes your customers tell you things before you know them, and again, I think that’s where conversational engagement is so important, and gives you an opportunity to create a lot of new experiences.
Chris Hemphill (00:20):
Hello Healthcare, one thing that we want to address consistently, and you’ll see this in a lot of our videos and hear it in a lot of our episodes, is what should we be thinking about AI and healthcare? Every time we make a release, there’s a lot of things that change over the years, and a lot of different things that happen, so we love to revisit this subject from time to time. A lot of the content, a lot of the material that comes out regarding AI, it might be more focused on what should engineers be thinking about, or what should AI practitioners be thinking about, but what we want to really drill into is how healthcare leaders can make the appropriate decisions. To make decisions in AI and to make sure that we’re doing right by our patients, that doesn’t necessarily require us to go and learn programming languages like R and SQL and things like that. What we really need to understand are what’s going on, how to evaluate what vendors are doing, and how to ultimately make these decisions the right way.
Honestly, I couldn’t think of anybody better than Brian Gresh to bring in on this subject. What we’re not going to be focusing on is getting all the way down into the super deep technical details, but Brian brings a rich background in healthcare and healthcare strategy and works for Loyal, which is a company that’s focused on enabling conversational AI and other things that help improve patient experiences and drive patient retention. What we’re going to be straddling, what Brian’s going to help us with today, is that line between healthcare strategy, artificial intelligence, and ultimately the types of things that healthcare leaders should be focusing on to make sure that we’re able to do right by our patients. Brian, I can’t go on enough about how excited I am about this conversation. 17 years of leadership in University of Utah, two and a half years at Cleveland Clinic, and now president of Loyal Health. Could you just talk about the background and why you’ve taken the path that you’ve gone?
Brian Gresh (02:26):
Yeah, I’ll try to, because it’s a long background, but no, I got into healthcare marketing kind of by accident. I moved to Utah to really enjoy and pursue the outdoors, but obviously, you have to pay the bills, so I fell into a career at University of Utah, and was part of a marketing team that was very small. It was three of us at the time, but I joined as all of the internet and online front-door capabilities of the health system were being built, so there was a committee put together for health sciences, and it was really trying to understand and establish what would the online presence for the university be. I was lucky enough to be part of that and got to develop a digital team inside the organization, which was really exciting because there was no playbook for that.
That led into a lot of other things from a digital perspective. I use that word pretty loosely, but always thinking of it, I’ve always had a real interest and a passion for thinking about consumer engagement and thinking about how can we take things from other industries and apply them to healthcare so that you can just make not just the experience better, but the outcomes better. I think that’s really important. That led to an opportunity to join Cleveland Clinic, which was equally exciting. I got to take a lot of the things I was doing at University of Utah and build upon those initiatives and that led me to Loyal. I really always enjoyed being part of teams that build things. I was definitely an internal build kind of guy at University of Utah and Cleveland Clinic, and this was just an opportunity again to try something new and build some things that hopefully can have an impact on the healthcare space.
Chris Hemphill (04:27):
Great. Honestly, I can’t wait to get into what those things that you’re building are, and then maybe even broaden it to the types. There’s all kinds of buckets and baskets where different AI solutions fit, so it’ll be good to talk broadly about those, too.
Brian Gresh (04:42):
Chris Hemphill (04:45):
Well, maybe just to jump into the broad side of the pool, transitioning to this role, into Loyal, I’m sure that there’s things that you’ve seen that you’re now seeing that, well, the things that may have surprised you, or things that are really interesting from an AI perspective. I’m just wondering, just overall, why should healthcare leaders and healthcare marketers have an interest in AI right now?
Brian Gresh (05:16):
I think they should have an interest because I think of AI as table stakes. AI is just a tool in the stack, and so if you’re looking for a new vendor, if you’re looking for a solution, AI is probably going to be part of it in today’s technology world. But AI is a big, broad term. Are we talking about NLP? Are we talking about machine learning? What parts of AI are we talking about? I think just having a general understanding of some of those different ways that AI can be applied, but I think maybe not thinking of it as a shiny object. Everyone says “AI” now because they should, because AI is part of just, again, the technology landscape, so not getting maybe distracted by the term AI, but thinking about, “What is the problem I’m trying to solve? Is this technology going to help solve it? Then is AI part of the broader guts of that technology?” That’s how I think about it and I think sometimes it’s too much of a distraction.
Chris Hemphill (06:30):
It really can be, if it’s phrased as this general, broad concept, a lot of the conversation around it, I personally think, does a lot of disservice to what the real capabilities are.
Brian Gresh (06:43):
Chris Hemphill (06:44):
One word that I queued in on when you were describing it was thinking about the ways that it’s applied, so from this, we’re in this space with healthcare marketers and communicators. What are some of the applications of AI that we should be thinking about on this side of the house?
Brian Gresh (07:02):
Yeah, I mean, the space that some of the things that we’re doing at Loyal, NLP is definitely part of that, so obviously, marketers are very focused on search, they’re focused on engagement with customers. NLP is a huge part of that space, so understanding, what are people searching for? What are they saying? What is the intent of their visit? NLP can break down their utterances, break down what they’re saying, and then help you guide that person to the next thing that they need to do, so I think that’s how it’s being applied, or the most common, maybe, example of AI in terms of the marketing space. That’s how I think of it. Then as you get into things like predictive analytics, that’s where you start to get into the machine learning aspect, and it changes the game a little bit. But I really think those are the most common applications that we’re seeing in the marketing space, for sure.
Chris Hemphill (08:04):
Very good, succinct way to spin it up. Being on the AI side of the house myself, I’m constantly throwing out terms like NLP and machine learning. Just to help with folks who might not be familiar with those terms, NLP referring to natural…
Brian Gresh (08:22):
Natural language processing, right?
Chris Hemphill (08:23):
Yeah. I think a lot of our audience might be familiar with what’s going on with predictive analytics, but natural language processing, extremely fascinating, using that to understand what people are interested in. I guess, what are some of the opportunities to use NLP?
Brian Gresh (08:47):
I think it’s being able to engage with patients or healthcare consumers maybe before their patients, just in a deeper way. I think up until recently, a lot of the ways that we would engage a healthcare consumer might be through an outgoing campaign or something, we’re driving them to a landing page, and they fill out a form, so we get their name, email, and phone number, but with NLP, you can start to have conversations with that user. They’re actually telling you what they think, what they’re looking for, what they need. You can break down those utterances and then you can start to either take that and learn and build new programs off of it. You can start to direct them to resources or transactions. Those are ways that you can naturally guide somebody to an appointment scheduling, so I think there’s a lot of ways that it’s being applied.
I mean, it’s not new. I mean, Google’s been using NLP forever. We’ve seen NLP being used in the provider space with translating notes and things like that, so the technology’s been around a long time, but I think that it really hasn’t been applied from a consumer standpoint in the healthcare space, so I think it’s exciting to be able to understand your customers at a deeper level.
Chris Hemphill (10:15):
Yeah, and it’s also exciting there’s a whole bunch of things that people are saying. There’s a whole bunch of free text data out there that gets ignored because it’s so complex, too. Yeah, but with natural language processing enabling that deeper understanding, it makes me really excited to hear about some of the things that you’re talking about.
Brian Gresh (10:36):
I get excited, too, about, I mean, the EMR is just this huge treasure trove of data, but so much of it is unstructured, and NLP can help with that. I mean, you can start to go in and look at patient notes, patient records in a different way. I think that’s another way that we’re going to learn more about our patients, our customers, and then create new solutions, and new opportunities off of it.
Chris Hemphill (11:02):
Switching from this environment where it’s focused on healthcare marketing within established structures to now going out to, like you mentioned earlier, there’s no playbook for the kinds of things that you’re embarking on. We’re talking about new ways of understanding our patient relationships and the things that they’re discussing. What’s been surprising to you? What have been some insights that you picked up on that really shook what you were thinking?
Brian Gresh (11:33):
There’s a few things. One, just they want to engage. Customers want to engage with health systems, and so if you provide the technologies for them to do it, they’re going to take advantage of it. There is such an access problem in healthcare, whether it’s trying to schedule an appointment, or whether it’s connecting with your provider, it’s a huge problem, so if you can create ways for people to access care, access providers, and even automate some of that process, it’s a huge opportunity.
But I’m always amazed at what somebody, for instance, will type into a chatbot. They’ll ask very sincere, very deep questions, not super high-level stuff that you would just assume, but they’ll ask pretty complex things, so I think that’s exciting. I think some of it’s also not surprising, I mean, that customers want to engage with their health system. It’s a special relationship, so we need to give them more opportunities to do that, and that’s where I think some of these technologies can really help.
Chris Hemphill (12:49):
One thing that’s interesting is that I think a lot of people might not be aware of how much people want to engage and how much their customers want to engage because in a lot of situations, especially if we consider a couple of years ago at the onset of COVID, there were a lot of health systems that were afraid to proactively send out communications about COVID, like at the very beginning with all the fear around it at the time, there was a lot of health systems doing the wait-and-see approach. I hope it’s super big on the screen when you said, “Customers want to engage,” because ever since the pandemic, we started seeing a lot of spikes in email response rates, and visitation rates, and things like that, just based on people wanting to know, and be informed by their health systems.
Brian Gresh (13:43):
Yeah. It’s interesting, in the early days of COVID, we have a chatbot product that many of our customers use, and we started seeing questions about COVID. I mean, COVID, wasn’t part of our intent library, COVID wasn’t a term that we had seen out there, so it was this great opportunity. It’s like, “What is this? Do we need to start training our model to understand what COVID is?” Those are the kind of things, sometimes your customers tell you things before you know them, and again, I think that’s where conversational engagement is so important, and gives you an opportunity to create a lot of new experiences.
Chris Hemphill (14:25):
Really interesting. Just curious, do y’all do any case studies about insights that you’re getting from these conversations that happen on your platform?
Brian Gresh (14:34):
Yeah, absolutely. We certainly look at conversation data, and anytime we see a trend in a certain direction, we dig a little deeper. We approach AI or NLP, we do supervised learning. It’s healthcare. You can’t just let a bot run free in healthcare, so we have analysts that are looking at data, and identifying trends. When we see that, we share that with our customers to let them know, “Hey. People are asking these questions around this topic. That’s not something that we see in your website content.” Or, “That’s not something that we’ve seen before. How would you like to address this? Where should we direct this person?” Or, “How should we answer these questions?” Yeah, that happens quite frequently.
Chris Hemphill (15:21):
Good. Well, I love the opportunity to identify, like when we talk about data, what we’re really talking about are incidents when people are trying to better their care, or get answers on questions about their healthcare, better their life status, so you’re able to look at these insights and then make suggestions from that. That’s really exciting to me. But there’s a lot of people out there that are claiming to be able to, like we’re all talking about predictive analytics, data-driven insights, and things like that, and one thing that I wanted to get from you to hear about was what healthcare leaders should be doing to… There’s all these vendors out there, like we said earlier, everybody claims AI, and that they use it as a distraction. How should healthcare leaders be making sense of this new MarTech environment?
Brian Gresh (16:10):
Yeah, I mean, that’s good question, and I don’t have the perfect answer for that. First off, I think, question everything. Again, I view AI as table stakes, but how it’s being applied differs vendor to vendor. There’s a big difference in deep learning, deep machine learning kind of stuff, and just scratching the surface. I mean, chatbot’s a great example, too. You can build a chatbot that’s very decision-tree-driven. It works. It’s not really AI. Then you take it a step further and you’re actually applying NLP. It’s the same way when you get into analytics. You can build a propensity model that’s pretty basic, but then as you start to really get into it, and start to apply machine learning.
It’s also about data. We haven’t talked a lot about data yet. We jumped right into the AI piece. But when I think about what healthcare marketers should be doing, the first thing you need to do is clean up your data. You’ve got to have a data foundation before you apply any of these things to it because it’s garbage in, garbage out. Whether it’s a chatbot, whether it’s predictive analytics, you have to have some structure around that data, you have to have some management of that data before you can just start letting this stuff run wild. I mean, that’s probably your experience, too. I mean, it’s got to be a foundation, right?
Chris Hemphill (17:41):
Yeah, AI is so fun and sexy to talk about, and cleaning up data is not, but that’s the foundation. Everything that we’re doing from an AI perspective, all the predictive analytics that we enable, imagine if we were doing that from a bunch of disparate systems, and not doing any kind of cleanup or normal… Yeah.
Brian Gresh (18:00):
Yeah, I mean, you nailed it, data’s not sexy. Nobody wants to talk about, “Hey, you should focus on data management.” That doesn’t get people excited. As soon as you start talking about AI, it makes a much better presentation, but it’s really important. Once you have that foundation of data and you are managing it well, then it’s skies the limit, you can start applying these things on top of it, and it’s just going to keep getting better. I mean, things just in the last five years have changed so much, and accelerated so much. It’s really cool to watch. The more data you feed into your models and stuff, the better they get. I’m excited, I mean, where this is all going. I think in another five years, we’re going to see a completely different landscape, so it’s good stuff.
Chris Hemphill (18:53):
I agree on that. I know I’m going to dig a little bit too deep on this data question, but I think that maybe it’s an opportunity to discuss data in a way that doesn’t sound completely scary to people. Do you have some ideas on …? If we look at our health system data, for example, a health system could have multiple EMRs, and each of those EMRs could have 50,000 different tables represented, all these elements, and everything like that, so for us to say, “Hey, we need to make sure the data’s clean,” that’s one thing, but is there a path that we can talk about for healthcare leaders to follow so that it’s less intimidating for like, “Oh, I have to be the sole person responsible for this gigantic cleanup”?
Brian Gresh (19:40):
Yeah. Well, one, I don’t think you can be the sole person. I think even taking another step back before the data, you’ve got to build relationships within a healthcare organization because there are multiple owners of data. If you are not working with the CIO, or the transformation officer, or the access folks, you have to all be working together, and create a good governance structure around data. Who owns it? Where does it come from? How are all those pieces fitting together? I think that’s a really important piece, too. It’s hard work. I mean, a lot of people don’t either want to do it, or they don’t have the resources to do it. But it’s so important because, I mean, when I was still on the health system side, I see it now being on the vendor side, there’s a lot of solutions out there that can be really helpful, but if you’re applying them just without thinking about that other stuff, oftentimes they’re not going to be successful, and so you want to make sure that you’re laying that groundwork to before you apply some of these things on top of it.
Chris Hemphill (20:48):
I want to add on another thought onto that, too, which is with so much data available, I mean, we can call it a treasure trove, but it’s a matter of finding that treasure trove, and ultimately, that depends on the use case. There’s one aspect of being able to go in, and look at all this data, we can admire the data, and run all kinds of calculations on it forever, but if we are starting from the actual use case, or starting from enabling different portions of our health system strategy, then that gives us a much better idea on what data we need to bring together to solve for this particular issue.
Brian Gresh (21:29):
Yeah. Chris, I mean, what you were saying about choosing the use case, I think, is spot-on. I think oftentimes we try to boil the ocean, especially with any new kind of technology or solution. I think starting with a use case is important, but not trying to do it all at once, and I mean, especially in large health systems, proving the case, building upon the case, not only are you learning, and understanding how to apply the technology, but you’re also building support within the organization. When you try to do too much and it fails, then nobody wants to use it anymore, so I think it’s probably the best point as part of this conversation is don’t try to bite off more than you can chew. Really, start with a use case and build from there is a great way to take it.
Chris Hemphill (22:25):
For those who are just listening and can’t see that I was shaking my head in violent agreement, yeah, that we need to get down to a use case, and we need to make it specific, iterable, in bite-sized chunks that we can start proving value early on those things.
I had one more question, which was around one thing that’s coming up a lot more in these conversations, as we have more and more AI use cases, we also see challenges, such as various AI elements discriminating based on race, and gender, and other factors. I’m curious if there are approaches or things that y’all are looking at just in terms of addressing the data ethics issue as well.
Brian Gresh (23:11):
I think transparency is probably the biggest one, allowing things to be open, allowing people to be able to see the algorithms, the underlying technology. I think it’s really important. It’s a new space and you’ve got to treat it really, really carefully. I mean, I say “new.” Not new in terms of years, but new in terms of how it’s being applied in healthcare. I think it’s transparency. It’s being open about how you’re using the technology, who’s programming the technology, and then constantly reviewing it, so making sure that you have some governance structure, making sure that you have voices at the table that can represent all of the different use cases, and the different audiences. That’s probably my best answer to that is just transparency, transparency, transparency because there’s a lot of things, a lot of technology out there where you don’t get to peek behind the hood, and I think that’s really important.
Chris Hemphill (24:16):
Absolutely. It’s really no place for purveyors of healthcare algorithms, things that are making decisions about patient lives, which again, we talked about how data isn’t just something that’s on a spreadsheet, it’s how people are engaging, and bettering their lives, and if somebody wants to keep that a secret or keep that behind a black box, then yeah, not.
Brian Gresh (24:39):
It’s not the way to do it. I mean, at the end of the day, we’re in healthcare, this is serious stuff. Facebook doesn’t have to worry about the same things that people in healthcare have to worry about, so I think you have to take baby steps sometimes. Often, people get frustrated in healthcare because things aren’t moving fast enough, but sometimes that’s a good thing. Stepping back and looking at the problem in different ways, I think, is really important. It’s not a race, it’s trying to do it in the best way, most ethical way possible, and then ultimately, it’s all about the outcome. It’s about how is the patient doing after their experience, that’s where it should all lead, so if you have that as the ultimate goal, I think you can build the steps along the way.
Chris Hemphill (25:30):
Well, let’s think about the outcome of this conversation and where you would hope people start taking their thoughts and start thinking about after they watch it. Or is there any final thoughts that you would share with the audience?
Brian Gresh (25:44):
Yeah, again, don’t get caught up in the shiny objects. We’re at a point where the technology’s proven. There’s good stuff out there, but you can’t just throw it at the wall. You’ve got to think about how you’re going to apply this stuff, so move slowly, and just be thoughtful about the way that you’re going to apply it, and don’t just run in without a plan.
Chris Hemphill (26:16):
Fantastic. Well, Brian, got to say, I really appreciate, I hope you enjoyed the conversation a fourth as much as I enjoyed it.
Brian Gresh (26:23):
This was great. Thank you. I really appreciate you having me and I’m looking forward to the rest of the conversations after this and listening to them.
Chris Hemphill (26:31):
Brian Gresh (26:31):
Chris Hemphill (26:32):
Brian, for the folks that want to engage more and learn more about what you’re doing and get more of your insights, where’s the best place for them to find you?
Brian Gresh (26:40):
Certainly, LinkedIn. I’m on LinkedIn a lot. I’m on Twitter. It’s B-M-G-R-E-S-H at Twitter, or @bmgresh is my Twitter handle. I always forget between LinkedIn and Twitter. Then email, I’m always available, email@example.com, so feel free to reach out. I love talking about this stuff.
Chris Hemphill (27:02):
Well, thank you for being so transparent. Thank you for sharing. For folks that want to learn more about making decisions about AI and healthcare, we actually set up a little path, it’s called Healthcare AI University. It should be popping up on the screen right now. For those who are just listening and don’t see it on the screen, we have a link to Healthcare AI University in the show notes, or you can look up YouTube, Actium Health, Healthcare AI University. Again, thank you very much for tuning into Hello Healthcare, and until we see you next time, hello.
Speaker 3 (27:37):
If you want more of the latest from healthcare’s thought leaders, subscribe using the button below, or you can visit hellohealthcare.com to get updates directly in your email.
Find the Clarity You’ve Been Missing
Learn how Actium Health is driving improved quality, outcomes, and revenue for innovative health systems nationwide.