Latest / CareTalk: Healthcare. Unfiltered. / Boosting Healthcare Efficiency with Conversational AI w/ Authenticx CEO Amy Brown

Transcript

Boosting Healthcare Efficiency with Conversational AI w/ Authenticx CEO Amy Brown · CareTalk: Healthcare. Unfiltered.

• 0:00
  Healthcare executives can learn a lot by monitoring individual customer service calls, but it's not a scalable or systematic approach.
• 0:07
  The holy grail is to analyze and activate customer interaction data at scale.
• 0:12
  How can today's AI -based technologies make that goal attainable?
• 0:21
  As kids, we're always learning and growing, but as adults, we can sometimes lose that sense of curiosity that makes life exciting.
• 0:29
  It's easy to fall into a routine.
• 0:32
  Therapy can help you break out of it.
• 0:34
  If you've been considering starting therapy, now is the perfect time to give it a try.
• 0:40
  BetterHelp offers entirely online therapy that's designed to be convenient, flexible, and tailored to fit your schedule.
• 0:48
  Whether you're dealing with stress, anxiety, or seeking personal growth, BetterHelp connects you with licensed therapists who can support you on your journey to rediscovering your curiosity and becoming your best self.
• 1:01
  Don't miss out on therapy.
• 1:03
  Visit betterhelp .com slash care talk to get 10 % off your first month and take the first step toward enhancing your mental wellbeing.
• 1:12
  That's betterhelp, H -E -L -P dot com slash care talk.
• 1:17
  Welcome to Care Talk, America's home for incisive debate about healthcare business and policy.
• 1:22
  I'm David Williams, president of Health Business Group.
• 1:24
  And I'm John Driscoll, senior advisor at Walgreens.
• 1:27
  Today's guest, Amy Brown, is founder and CEO at Authentics, which is on a mission to help humans understand humans.
• 1:34
  Join the vibrant Care Talk community on LinkedIn, where you can dig deep into healthcare business and policy topics, access Care Talk content, and interact with the hosts and our guests.
• 1:45
  And please be sure to leave us a rating on Apple or Spotify while you're at it.
• 1:49
  Amy Brown, welcome to Care Talk.
• 1:51
  Thank you.
• 1:52
  It's so great to be here.
• 1:53
  So Amy, how did you...
• 1:55
  decide to take on the wonderful patient experience that US Health Care represents today.
• 2:01
  How did you come to this particular journey?
• 2:03
  Yeah, it was just something that I could no longer ignore because I was working for a couple of decades, actually, on the business side of health care.
• 2:13
  I spent most of my career at pre -authentics in the payer space, but dabbled in provider-owned health health plans, as well as a little bit of pharmaceutical manufacturing.
• 2:27
  And through all of that, my role was to manage the teams that interfaced with customers all day long.
• 2:34
  And customers might be healthcare consumers, patients, and they might be other members of the healthcare ecosystem, like providers, healthcare providers.
• 2:43
  And in managing those teams, I just saw firsthand the administrative complexity of our system.
• 2:53
  And I also could hear from our customers through our contact center just how confusing and complex it was.
• 3:02
  you know, we were, most companies are recording their conversations, but they're not really using them at scale to help improve the system itself.
• 3:14
  And so I just got to a...
• 3:16
  stage in my career and a stage in my life where I could no longer kind of sit by and watch that happen without trying to make an impact.
• 3:27
  And so in 2018, I left my corporate career and took the leap to build a company to help solve the problem of making sure everyone in the healthcare ecosystem has access to the literal voice of their customer and to learn from that in a way that can can drive systemic change.
• 3:48
  So Amy, was an interesting time, 2018.
• 3:51
  So you gave yourself a couple of years to get going before the pandemic hit.
• 3:56
  And now we're a couple of years beyond that.
• 3:58
  How has the patient experience itself evolved?
• 4:02
  Are things now better?
• 4:03
  And if you were in 2024, you'd say, OK, I wouldn't have left what I was doing because everything's been solved.
• 4:09
  Is it worse?
• 4:09
  Is it different?
• 4:10
  I have my own answer how patient experience is.
• 4:14
  But what's the general state?
• 4:16
  I wouldn't say it's better.
• 4:18
  I'd say that there's been obviously a lot of innovation in the last six years of companies and startups trying to help improve access to customers.
• 4:30
  So you see a lot of digital front door type options being spun up and deployed.
• 4:38
  We see health systems trying to rebound from COVID massive cost pressure there.
• 4:47
  We see major competition amongst specialty brands and pharmaceuticals, a lot of competitive products out there, everybody trying to capture market share.
• 5:01
  On the payer side, a lot of competition on the Medicare Advantage side.
• 5:05
  And so everybody's trying to figure out how to deliver a positive customer experience.
• 5:14
  at in the cheapest way possible.
• 5:16
  And sometimes those two things conflict with each other.
• 5:20
  And there's a lot of hope that's been put on the on the back of AI to solve the world's problems.
• 5:29
  And you know, we are an AI company, but I see a lot of turning and paralysis in the market right now around the topic of of adoption and deploying AI to solve problems.
• 5:42
  And there's a bit of a mismatch right now between what executives think AI can do for customer experience and the bottom line versus kind of the reality of deploying AI in a way that actually drives to those changes.
• 5:57
  So it's an interesting time to be in this space.
• 6:00
  But Amy, to be fair to the health care system, we've got a Wealth of data now on the patient experience.
• 6:07
  There's all the press gainy stuff, which I was involved in for a few years that takes information on patient experience that hospitals pay a lot of attention to.
• 6:16
  There are rankings of hospitals.
• 6:18
  There are the healthcare, the Medicare, HCAP scores that are now much more focused on what was the customer experience that feeds into STARS, which is how people get paid in Medicare Advantage.
• 6:33
  Shouldn't we be in the golden age of patient experience, notwithstanding David's constant complaining about things not working for him?
• 6:42
  John, I forgot you were going to get sensitive about the fact that you're on the Press Ganey board and it was supposed to have solved all problems.
• 6:48
  Yeah, was.
• 6:49
  No, no, we just call him as we see him.
• 6:51
  We're responsible for reporting.
• 6:54
  You're responsible for helping make change, David.
• 6:57
  Well, we're certainly not at a loss of of data, we're almost drowning in data.
• 7:04
  And it's easy to kind of get overwhelmed by the amount of data that the health system has access to.
• 7:13
  Our view is that historically there's been a really significant missing piece of the puzzle.
• 7:23
  You know, through the CX customer experience evolution, There are many solutions out there aimed at understanding, know, customer's feedback.
• 7:38
  And not only their feedback on satisfaction, but also, you know, we survey them to tell us what their social determinants of health are.
• 7:45
  The problem with all of that, and why I believe it's incomplete, is that it's a skewed data set.
• 7:53
  When you survey someone, you are dependent on...
• 7:58
  first of all, a subset of humans to respond to surveys.
• 8:02
  You're also designing the questions that you, the company want to know, which biases the input that you get.
• 8:13
  The reality is that there are almost a billion customer conversations that happen in this country alone every year that are recorded where customers are unsolicitedly telling you in the moment, what their frustrations are, how they're getting stuck, why they have a gap in care, why they can't get access or start their therapy, why they haven't started their mental health care plan, why they're not going to follow their discharge plan.
• 8:44
  I mean, they're telling us all day, every day.
• 8:47
  our Amy, you're kind of saying that patients are talking and providers in the system aren't listening.
• 8:54
  They're not.
• 8:54
  They're listening to survey data, which is incomplete.
• 8:58
  And I think there is a belief that, these call centers, they're administrative, they're transactional.
• 9:06
  There's nothing to glean here from these conversations that can help the care of patients.
• 9:11
  And that's just wrong.
• 9:12
  We know that from listening to the data.
• 9:14
  One of the interesting things that you described is that, you know, with all these billion conversations, all these things that are happening and all the different complexity, you got to boil it down.
• 9:26
  to something you call the Eddie effect and described as the quote from yours, you say the most reliable metric to identify customer friction points for healthcare leaders.
• 9:36
  What is this Eddie effect and is it like the grand unifying theory to explain the customer problem?
• 9:42
  Yes.
• 9:43
  Well, I'll first tell you where the Eddie effect idea came from.
• 9:47
  When we were just getting started in analyzing some size of critical massive data, as a company, I went home one day and shared with my husband, hey, we're seeing this trend in the data where customers are just having to work really hard to get from point A to point B.
• 10:11
  And we see that through callbacks.
• 10:13
  We see that through going from chat to call.
• 10:16
  We see people saying, hey, you told me I would get this thing.
• 10:20
  I'm not getting this thing, now I have to call back again.
• 10:23
  And it's just the circular experience that we're seeing.
• 10:27
  And my husband, who is a outdoorsman and a river guide, said, well, that sounds like a river, which is, you know, when you think about the current and a river flowing downward, when there's an object that falls in the river, like a big boulder, it actually creates a counter circular effect on the current and things get stuck there.
• 10:49
  And that made so much sense to me because, you know, we think of our customer journeys, we design a map, right?
• 10:55
  And we think that they're beautiful.
• 10:57
  And we assume that customers are flowing through the customer journey without obstacles.
• 11:02
  And the reality from listening to the actual customer journey is people are getting stuck all the time.
• 11:06
  So we've designed and built a proprietary machine learning model that knows how to detect customers who are getting stuck in their journey.
• 11:17
  It's called the Eddy effect.
• 11:19
  And not only do we have a model that can listen for that, we actually can quantify it.
• 11:26
  We can tell you what the root causes of the eddies are.
• 11:29
  And we can measure over time whether systems companies are making a positive impact on reducing the eddy effect rate, which by the way, has an impact not only on customer satisfaction and retention, but it also has an impact on profitability for the business because By definition, any effect interactions are preventable if you're just listening and solving for those issues.
• 11:56
  I just think it's so wonderful that you can so systematically break down what happens in rivers, because David never gets out of the house.
• 12:05
  But that whole recursive function of customer service, everyone has experienced, whether you're calling your United or Delta where you're trying to get through your auto claims, what do you think's different about health care?
• 12:23
  mean, it sounds like your approach would create solutions throughout the customer service, the barren desert of customer solutions.
• 12:33
  What makes it particularly important or relevant in health care?
• 12:38
  Yes.
• 12:39
  Well, I've been asked many times by investors, you know, why not go more horizontal with your solution?
• 12:45
  And the answer is two things.
• 12:47
  First, in order to have really effective AI to help listen for things, you need to, the AI models need to be fed data that is specific to the context of the industry.
• 13:02
  And so we've been committed to building high quality AI using healthcare specific conversations.
• 13:08
  And we think that that's what makes us competitive in the healthcare space.
• 13:14
  But the other reason is because the implications for eddies in healthcare are so much more severe than in other industries.
• 13:22
  I mean, we can measure patients who literally are dropping off therapies that are sustaining their life because of administrative burden.
• 13:36
  That obviously has implications for the patient's lives and it also has implications for the business.
• 13:43
  And so it just makes sense that we should be focused here because the stakes are so high.
• 13:49
  Amy, let me ask you a little bit about that one about therapies because a common business process for a healthcare provider is prescription renewals and they devote a lot of effort to it and there's a lot of calls that are coming in about that.
• 14:01
  Can you talk about how, like what is that issue there?
• 14:04
  where are the eddies and how does an organization learn from this sort of insight at scale?
• 14:12
  What kind of insight did they get from that other than, my prescription renewal line doesn't work too well?
• 14:19
  Yeah, so we actually just completed a pretty significant research effort where we took a million conversations that were patients stuck in eddies.
• 14:34
  This is from all of our client data and we serve payers, providers and pharma, right?
• 14:38
  So we took all that.
• 14:40
  We found a million patients who were stuck in eddies from a six month period of time.
• 14:46
  And we examined what was going on there.
• 14:51
  And the number one issue was the prior authorization process.
• 14:57
  And...
• 14:57
  And in terms of where in the journey is that getting stuck, every single part of the healthcare ecosystem is touching the prior authorization process.
• 15:09
  They are.
• 15:10
  So the provider is the one that makes the request for care or therapy.
• 15:20
  But there are forms, there are appeals, there are...
• 15:25
  financial implications, like just because you get a prior authorization doesn't mean that a patient can actually afford that.
• 15:31
  So then they call their pharmacy manufacturer to get copay assistance.
• 15:35
  And then there's step therapy.
• 15:38
  There's all of these steps that have been put in place that we think are there to protect the patient and ensure that the provider is following, you know, guidelines, but the reality is we're wasting billions of dollars just managing that process as an entire healthcare ecosystem.
• 16:04
  And the burden is mostly borne on the providers because they're the ones having to sign and to appeal and to follow up.
• 16:15
  And it's just, it's a mess.
• 16:17
  I mean, do you think I think it's the bigger burden is on the, the administrative burden is on the provider, but the bigger burdens on the patient.
• 16:26
  So most of us who've been looking at healthcare feel that as a problem.
• 16:31
  How do you start to structure a solution?
• 16:35
  You know, if you're in a river, you've sort of spin around in the eddy, but then you spin back out into the current.
• 16:41
  Currently, you know, for patients, it's more like a roach motel than an eddy in many cases, which is why they're falling off of therapy.
• 16:48
  They're stuck.
• 16:49
  and they're falling and they can't get up.
• 16:51
  I mean, how do we structure something versus just defining it that solves it?
• 16:57
  It feels overwhelming, doesn't it, John?
• 17:00
  Well, to people like me and David, yes, but probably to clever people like you, no.
• 17:04
  it's overwhelming for sure.
• 17:06
  And for us, know, our world is listening to health care consumers.
• 17:13
  And when you hear the actual voices telling their stories, over and over and over again, that weight of responsibility that I feel and that our company feels just is even even heavier because it's one thing to have the data point.
• 17:31
  It's another to hear it, right?
• 17:32
  And that's how we help.
• 17:34
  we our solution is designed to help leaders who empowered to make change.
• 17:44
  not only see data at the macro level, but be able to hear the patient friction.
• 17:50
  And we see something very, very powerful when those two data points are combined, when the qualitative voice of customer and the trends combined.
• 18:01
  Suddenly people are not only triggered in their brains, but they're also triggered in their heart and their gut.
• 18:07
  And we're seeing people drive change.
• 18:10
  So within organizations, we're seeing more cross -functional effort between companies.
• 18:14
  compliance, legal, marketing, operations, strategy, start to remove complexity from their own journey.
• 18:23
  But then there's a problem of cross -functional, But before you go there, I mean, I think you're on something super important.
• 18:30
  When I took over as CEO of CareCentrics and I was running into a lot of customer service patient challenges, one of the things I started every management meeting with was one very detailed failure for a patient and one very detailed success.
• 18:45
  And if you can connect the statistics to people's stories, it was incredible for me how it just lifted up and then people self -organize around solving problems.
• 18:56
  Because I think most people in healthcare actually want to solve those problems and they identify with the challenges that people have and can come up with solutions.
• 19:04
  But it needs to be prevented in a human context.
• 19:07
  I think that that the power of those specific stories, the voices, coloring in the details on people's lives really is, once you illuminate that, you illuminate the problem, you can also get people motivated, I think, to solve the challenges.
• 19:23
  It's a very powerful point you're making, and I've seen it happen in organizations I've led.
• 19:28
  Yeah, thank you, John.
• 19:29
  We actually have an entire practice within Authentic called Data Back Storytelling, and we offer a certification program.
• 19:37
  for our clients to become data back storytellers because it is, you're right.
• 19:42
  It's not just about getting the data.
• 19:45
  It's about using the data in an impactful way to inspire and enable human beings to solve problems.
• 19:54
  And storytelling is such a crucial component of that and nowhere needs it.
• 20:00
  There is no industry that needs it more than corporate healthcare.
• 20:04
  So David, there's an opportunity for you to get certified.
• 20:07
  John, I'm feeling so bad.
• 20:08
  Not only am I not on a river, you got me in some sort of a desert or whatever.
• 20:12
  So I'd have to see if I could recover from that.
• 20:15
  But I want to talk about technology for a minute, Amy, because when you're dealing with all this amount, one of the reasons it's hard to deal with this data, there's a lot of it and it's unstructured.
• 20:24
  It's in a format that's hard to analyze other than, I listened to something and I heard it.
• 20:30
  It's pretty obvious that natural language processing is a good fit.
• 20:34
  And I assume, you know, in 2018, that's definitely a focus.
• 20:37
  There's been quite a lot of movement in AI since then.
• 20:40
  And I'm curious about how you deploy technology, whether it's AI or other technology, and, you know, where that's been successful, and then, you know, where some of the challenges lie.
• 20:53
  Yeah, at a very early stage in our business, I brought on my CTO, my chief technology officer, who pretty quickly said, Amy, we need to build an AI company in order to really effectively analyze conversations at scale.
• 21:13
  But we're going to do so in a way that makes us both proud.
• 21:16
  Because I was skeptical about the power of AI.
• 21:23
  So we decided that we wanted to first ensure that whoever was going to start adding structure to these unstructured data sources needed to be credible in our eyes.
• 21:35
  And so we employed our own nurses, social workers, and people who have worked inside of healthcare to listen and to start to tag and label the themes and the topics that we wanted to glean out of conversations.
• 21:50
  This is where the Eddie effect, for example, was invented.
• 21:54
  our team said, look, this pattern is happening over and over again.
• 21:59
  And it's.
• 22:03
  It's so powerful that we're going to start kind of training a model.
• 22:09
  And so that's what we did.
• 22:10
  And so we have the way we think about AI is it is a think, think of AI as an umbrella term that is a set of tools.
• 22:21
  The oldest of those tools available is NLP.
• 22:24
  I think about NLP in a very simple way.
• 22:27
  It's like word search and find.
• 22:30
  You can program it to find very specific words and it can do that at scale, which is really important when you're wanting to find something very specific.
• 22:40
  But sometimes you don't know what to listen for.
• 22:45
  You don't know what the words are and you need AI to help you by telling you here's the trends.
• 22:50
  Here are the things that are happening.
• 22:52
  And that's where machine learning and deep learning comes in because they're pretty sophisticated algorithms that can identify patterns and themes that are happening inside of these conversations.
• 23:07
  And so we've got about 15 machine learning models that have been used to identify very specific things in the healthcare space.
• 23:19
  Eddy effect is one of them, but we also have an adverse event machine learning model that identifies patients that are of side effects.
• 23:27
  I'm intrigued by those frameworks and it's really, you know, I love the fact that you've got contextual experts to kind of make sure that the semantics are dialed in in the right way and the definitions and all that stuff.
• 23:41
  Why were you a skeptic of machine learning?
• 23:43
  Everybody else, I David and I are blind believers in technology.
• 23:47
  Well, because I used to be a purchaser of speech analytics and telephony platforms, and I bought into a lot of the marketing around it and was disappointed, right?
• 24:01
  And I also believed, I mean, this was 2018 before chat GBT and the real boom of AI.
• 24:09
  I believed that we needed to go deeper.
• 24:13
  We needed to understand things at a deeper level in order to drive real true change.
• 24:20
  It's not enough to have a data point.
• 24:22
  You have to understand the context.
• 24:24
  You have to understand the why behind a problem.
• 24:27
  And I really, really wanted to use conversation data to start getting at the why behind data.
• 24:33
  And conversations are so rich with context about why.
• 24:36
  I want to ask a little bit about how you might ask call centers to do things differently given a tool like yours being available.
• 24:45
  So you mentioned that you can analyze all this unsolicited feedback and get a lot of information out of that.
• 24:51
  If we know that the data is going to be analyzed in a sophisticated and powerful way, is there something else that would be done?
• 24:57
  So for example, I have plenty of opinions when John doesn't squelch them.
• 25:02
  But on a customer service call, I usually hold back because I'm just trying to do the minimum.
• 25:07
  I don't expect that the rep can actually do anything.
• 25:09
  I'm worried it's going to confuse it.
• 25:11
  Whereas sometimes you'll see at the end of the survey something like anything else.
• 25:16
  And if you could just say, just give you just stay in your own words or we know where there be value to just say, Hey, just give me some other feedback and we'll record it.
• 25:23
  We've got a way to act on it.
• 25:25
  Would that, would that make a difference or can you already deduce everything just from the nature of the conversations?
• 25:30
  There's a lot we can deduce.
• 25:32
  Honestly, we can, we can understand when a customer is experiencing something that they shouldn't have to, even if they aren't explicitly stating that.
• 25:41
  That said, post call surveys and anything else comes into our system as well.
• 25:46
  So we can analyze that unstructured data.
• 25:49
  And then some of our clients use what they've learned from listening to our conversations to restructure their conversations, right?
• 25:56
  To say things differently, to simplify, to ask questions.
• 26:02
  And they do that because they know they're gonna get to hear the impact of that at scale using our platform.
• 26:09
  But I wanna make one point.
• 26:11
  because you mentioned like, you know, contact centers using this data to change contact center interactions.
• 26:19
  And our vision certainly is to help the contact center, but it's bigger than that.
• 26:24
  Our vision is to use the insights that are in conversations to actually fuel the entire enterprise.
• 26:34
  Because there are things for marketers, there are things for UX teams, there are teams for...
• 26:41
  There are insights for product development teams, strategy, C -suite, and these conversation data.
• 26:50
  An example, market events happen all the time.
• 26:53
  There was change healthcare a few months ago.
• 26:56
  The IRA coming out, right?
• 26:59
  These are things that the ground truth of what's happening can be heard and seen in your contact center.
• 27:07
  And we now have...
• 27:09
  shareholders and boards of pretty big companies using Authentics to understand the ground truth of some of these market events to know how they need to react, how they need to respond.
• 27:22
  so our vision is to make the voice of the customer accessible, digestible, and usable by everybody in the enterprise working together.
• 27:32
  Great.
• 27:33
  John, last question to you.
• 27:35
  I guess if you're thinking about the power of what you're doing enhanced by context, data, and advanced tech, what's the line you have to draw between what people have today when they talk about the voice of the customer and the ground truth that you believe you can get at?
• 27:56
  What's your advice to leaders in health care as to how to get at that ground truth?
• 28:01
  I mean, our advice is to listen to the evidence that is existing with their actual customers by the millions on a daily basis.
• 28:15
  mean, perception is everything, right?
• 28:18
  And the perception of your customers is crucial to understanding what their behaviors are going to be, what their intentions are.
• 28:29
  They're telling you indicators of whether or not they're leaving you.
• 28:34
  as a business and going somewhere else, they're telling you what the competition is doing.
• 28:38
  They're telling you what their customer experience expectations are.
• 28:43
  And if you're not tapping into that, then you're missing out on being able to do something proactively rather than just waiting for the lagging indicators to happen of your growth or your profitability.
• 28:56
  Well, that's it for another episode of Care Talk.
• 28:59
  We've been speaking today with Authentics founder and CEO, Amy Brown about analyzing and acting on customer interaction data at scale.
• 29:07
  I'm David Williams, president of Health Business Group.
• 29:10
  And I'm John Driscoll, senior advisor at Walgreens.
• 29:12
  If you like what you heard or you didn't, we'd love you to subscribe on your favorite service.
• 29:17
  And thank you, Amy, for joining.
• 29:19
  Thank you.
• 29:19
  It's been great.