Can AI be empathetic? Reinventing client experiences
31:50 Share

Hello, hello. Malcolm Gladwell here. I want to tell you about a new series we're launching at Pushkin Industries on the 1936 Olympic Games. Adolf Hitler's Games. Fascism, anti-Semitism, racism, high Olympic ideals, craven self-interest, naked ambition, illusion, delusion, all collide in the long, contentious lead-up to the most controversial Olympics in history. The Germans put on a propaganda show, and America went along with all of it. Why?

This season on Revisionist History, the story of the games behind the games. Listen to this season of Revisionist History wherever you get your podcasts. If you want to hear episodes before they're released to the public, subscribe to Pushkin Plus on Apple Podcasts or at pushkin.fm slash plus.

Hello, hello, hello. This is Smart Talks with IBM, a podcast from Pushkin Industries, iHeartMedia, and IBM about what it means to look at today's most challenging problems in a new way. I'm Malcolm Gladwell. Today, I'm chatting with Anil Bhatt, the Senior Vice President and Chief Technology Officer of Anthem, one of the most prominent health insurance companies in the United States. We have been now pivoting to more around, okay,

We are building these capabilities. We are building these solutions. How are they fundamentally changing and improving the lives of our members, our communities, and really making a difference to the people we serve? Anil has been with Anthem for over 13 years and has spearheaded efforts to improve customer experience and members' needs. I'll also be chatting with Glenn Finch, managing partner of Global Business Services at IBM. How you deal with

empathy in an AI system is all based on the choice of words that you use and the verbal inflections that are present when you have a voice response. Glenn is a 25-year IBM veteran. His work focuses on the most challenging and transformative engagements at IBM. I'm excited to share my conversation with Anil and Glenn about artificial intelligence and how it's influencing customers to interact with their healthcare in a new way.

All right, guys, let's get started. Hi, everyone. Thanks, guys, for joining me today. Why don't we start with the two of you just introducing yourself. Tell me...

Tell me what you do. Awesome. Great. I'm glad to be here today. Thanks for hosting us. I basically lead the technology and practice here at Anthem as a CTO, managing all the roadmaps for technology, making sure that we are building solutions that are meeting our business needs on a day-to-day basis, making sure that we are catering to the needs of our members. So overall technology roadmap, making sure that we work with partners like IBM,

to bring new technology to the forefront. And how long have you been with Anthem? I've been with Anthem for 13 years, actually, and the company has evolved. While we are in the healthcare business, our focus has been more member-centric now. So really understanding how a big organization like Anthem can make sure that we pivot from being a normal, traditional, listed company

which definitely is meeting the expectations of the stockholders, but also catering to the need of our members and the communities that we serve in. Yeah. Glenn, why don't you introduce yourself?

I'm Glenn Finch. I look after data and AI on the services side of the IBM company. We take a lot of wicked cool technology and bring it to life at clients like Anthem. And I get the great pleasure of working with Anil on a daily basis to really fundamentally change the member experience

using artificial intelligence. So usually on the cutting edge of things and just love coming to work every day. Yeah, yeah. So you said something, the two of you have been working together for some time. When did you guys meet or first know each other?

As I said, the industry has been evolving a lot, Malcolm. So a couple of years back, we basically were kind of figuring out as the consumer experience changes, as people get so much used to Netflix and Amazon and the way they do their day-to-day shopping, the way they experience things. We were looking for a partner where we could really explore the power of AI, really use our data in a way wherein we can create these personalized experiences that

So that's where Glenn and I actually talked a little bit and we figured out that there is a possibility of us partnering. IBM bringing its...

technology. And basically that's when we kind of figured out there's, there's definite role to play and partner on this journey together. And it's been great over the last two years, we have been able to deliver on some great exceptional experiences for our members. And, and we are now moving beyond to other constituents and really making sure that we make it awesome for, for members to connect with us. Yeah. Glenn, had you worked with an, an insurance provider before?

We have a variety of clients around the world, so yes, but Anthem is special to my heart. We started thinking through this because when you work with Anthem, this concept of member and member experience, you need to show up every day with that mindset.

front and center in your mind. So there are other clients who focus on cost or technical debt or something like that, but that's not true at Anthem. You need to show up front and center every day with how are you going to radically improve the member experience first. I'm fascinated by this. The relationship between the two companies and the two of you goes back so far that I'm really curious to get a sense of how the kinds of questions you've been

and problems you've been trying to solve have evolved over that time. Tell me about 10 years ago, what were you guys talking about? So I think the 10 years back, the conversations were more around, okay, how many servers do we have in our data center? How many licensing points that we are going to be spending this year? What will be our footprint? What is our network speed? Are we able to manage the new capabilities that we are delivering? And it really was very technologically focused conversations that we used to have.

And what has happened over the years, Malcolm, is that we have been now pivoting to more around, okay, we are building these capabilities, we are building these solutions. How are they fundamentally changing and improving the lives of our members, our communities, and really making a difference to the people we serve? So as we looked at technology and engineering, we kind of pivoted from that to more a platform and a product that we are building for our constituents. And

And as that pivot happened, you know, I would say around three, four years back, the conversation then evolved to more around, okay, how are we improving the experience? How are we making sure that we are making it easier for the members? And it pivoted from being reactive and kind of what I called sick care management to more wellness-oriented conversation. How do we keep our members healthy? And that's where the overall...

pioneering of personalized experiences, predictive and proactive healthcare management kind of started. And as we had interacted with IBM, we knew that they had the technology and they had the real backbone, which could serve the needs that we wanted to kind of bring forward. Glenn, talk about that pivot. I'm curious, what's driving it? Did you go to Anil and say, look, you have an opportunity to do so much more here? Does Anil come to you and say,

I don't want to be just focused on technology. Our members are telling us X, Y, and Z. Take me back to that transformative moment when you start thinking about this project in a different way. There's been a massive shift at the IBM company in general to shift away from pure technology and move towards technology on behalf of a workflow.

When you think about artificial intelligence and you are trying to have a conversation with someone, right? You don't need just deep artificial intelligence programmers. You need to have people attached to that that know how to have a conversation with people and what sequence of words are going to elicit a response and how that experience feels to a member.

That's a very different type of program than just dropping in a chat bot and hoping it works, right? To answer the 12 questions that you get most of the time, right? And you'll mention this concept of personalization, right? Just making sure we put the right people together on the program, right?

is half the battle, right? And that's a shift that IBM has made very consciously. It started about five years ago. You know, we really, in earnest, called out intelligent workflows about two or three years ago, and that's when we started doing this together. It was tough. It was ambitious as compared to anything else that we had done out here.

And one thing which Malcolm was very, very beautiful and has been very important for us, learning on the go. When you have so much data that you're capturing, when you have a technology that really can give you in a nanosecond the response to what exactly is happening, the beauty of it is that you can pivot and kind of change on the fly. The agility that you build into our system, the agility that you build into our operations is a key element.

And that's what we have been able to do. And fortunately at Anthem, we have been really at the forefront of that, investing the right dollars and bringing the agility, bringing the way we can kind of pivot to what is more important to the concerns. That has been a great thing that has been happening here. Let's go through some very specific examples.

So I'm a member of Anthem. I am on your website. I would like to accomplish something. Tell me a specific thing that an expectation a member might have and how...

you have set about trying to satisfy that expectation. And let's get super specific. Give me a scenario, a tough scenario. Yeah, yeah. Well, I think I can give you a comparison to the past, right? So when you were enrolled as a member, we probably would send you an ID card, which was a hard piece of paper.

a very good piece of paper, which costs us a lot. Then there was nothing that we would let you know other than that, hey, if you want to register on our website, please be welcome, right? And then that's where our first interaction with you as a member used to happen. And frankly, there was nothing after that. There was a vacuum. And then you would probably try to understand your benefits. You will make sure that you know what your copay is. And then you

We'll not hear from you for a long time. And all of a sudden, someday, unfortunately, somebody is sick in your family. And then you pick up the card, go to a provider and basically have a visit there. And then you go from there. So that's the traditional experience that somebody would have had.

Right now, we have totally revamped that. So as a member, when you enroll with us, we send you a welcome kit, which send you a digital welcome kit. We send you an ID card, which is available on your phone. We send you a link to our Sydney Health app, which basically you can download. You can register in a minute. But if you have been an existing member, you will get a personalized curated news feed, which is specific to you based on your...

Your prior experience and based on your claims history and other things that we know about you. We work with IBM around the AI chatbot, which is basically a Watson enabled chatbot, which you can ask the questions from. What is my copay? You don't have to call us. You don't have to send us an email. You can really ask a question there itself. You can ask for what are the providers near me?

And we'll match a provider to you based on your past history. And that's where AI comes in. What do we think Malcolm's age group, Malcolm's prior history, tells us who should be the right provider for him to take care of things? So that interactive, more personalized, more engaging experience is what is different. Let me give you an example. I'd love for both you to weigh on this. So I'm 57 years old. It is indicated for someone of my age that I get a shingles vaccine.

I didn't notice, never occurred to me. Friend of mine got shingles. It was like the worst experience of his life. He lost three weeks. It was like so painful. And he's like, whatever you do, Malcolm, you need to get a shingles vaccine right now. So I went out and got my shingles vaccine. And then I had to get the booster. And I remember the booster and blah, blah, blah. Now, when you're talking about Sydney and about drawing on past experience, if I was a longtime Anthem subscriber,

Would you reach out to me and say, Malcolm, you got to get your shingles vaccine? Would you do, is that what you're thinking about? Exactly, exactly. Not only we will tell you that you need to take shingles vaccine, we'll tell you exactly

Which provider probably is the right one for you? And that is what the beauty is right now. The nudge to care, really the care gaps that we have. How does the data tell us that these are the care gaps in Malcolm's journey? You know, you pay a lot for your insurance company to take care of you. And how do we make sure that we take care of you? We be your advocate. We be your journey partners rather than just allowing you to come to us when you feel that you are sick. So this...

This AI system is called Sydney. First of all, who came up with Sydney? I actually love the name, but whose decision was it to call this system Sydney? Actually, it was our team. We did some research in terms of what could be a very neutral name that we can keep out there.

And Malcolm, I can tell you that I love the name so much that beginning of 2020, when we had COVID hit us, my daughter was asking for a dog for a long time and we got a dog. And actually we named the dog Sydney. So that's how much I care about the name and how much I love the name. But thank you very much for that. Just so we're clear, are you Sydney? I'm getting the AI system and not your dog. That's all I want to be clear about.

We may supply you with a picture of Sydney when you come to the app, but yeah, you're getting the AI. Yeah, so what we find is that to build trust in AI systems and to build the willingness for a member to go along a journey experience,

There's some things we have to do at the table stakes level, at the grassroots level that we have to get right inexorably. And I'm going to go back to Anil's comment about the ID card. What happens if you've lost your ID card? You don't want to wait on the phone for anybody to get a replacement ID card.

You'd like to be able to do that once and done on the web or the mobile, might have to ask a couple of questions and have it done, lights out, right? So there's this combination of doing the more routine things with absolute 100% precision, lights out, complete ease of member. And then that builds this trust, right?

to have this more longitudinal journey to answer your questions or to recommend to you about shingles vaccine, right? Or a variety of other things based on, you know, your health challenges. So it's kind of a double-edged sword of taking care of the table stakes and taking people along the journey. Your point is you start with

the very prosaic stuff and you build a trust in the system and then you can move to the more high-end stuff. Tell me about how you build an AI system like this. This is not a trivial accomplishment. What went into building Sydney?

Yeah. So I think, you know, Malcolm, traditionally, we have a lot of data over the years that we have accumulated for every member. And we have, you know, 80 million lives, multiple petabytes of data, which is sitting on our systems.

And that data basically allows us to learn. The data is data as long as you don't touch it, you don't do anything. But once you start really using technologies and when we call AI, these are mathematical models that you can run on this data to give you insights. And those insights are the key at the end of the day. And as we get those insights, we have to make sure that we have a way to use those insights to make a difference in the

in the life of any member that we have or any constituent. Actually, our sales experience for our brokers, our providers getting to know exactly what they need to know is very, very important. So we are making sure that this data and the mining of this data is constant. So when we talk about the partnership with IBM, we are talking about ability for us to mine this data on the fly at a very, very quick speed.

speed. And that is what is key. Then we are able to use AI in a different context. And I'm going to give you an example of something that really we are bringing to the forefront of what we call as a nutrition tracker. So imagine that you have your phone in front of you.

You have a plate of food that came in front of you and you can open Sydney and show the food of plate to Sydney. Sydney can tell you based on what is the plate. You take a picture of the photo and Sydney looks at the photo and says, why are you loading up on carbs? I mean, is that what you're talking about? Exactly. That's what I'm talking about. So, you know, this is a great partnership we have with one of our ecosystem partners. And actually this is in pilot with our house account, which is 80,000 members right now.

And it can tell you, you can show it a cup and it can tell you this is a coffee with no milk and it's going to be 70 calories. And it keeps track of what you're eating. And basically that's how we build the healthy habits out there. So the advancement in the field of technology and how do we make sure that we move away from that legacy information technology world to really the exponential technology that is in front of us is the key.

We needed to take all of that AI and persist a conversation with a member, right? And that's where Watson came in to help Sydney persist conversations with members, right? Because crunching through data and knowing about your claim is one thing, but being able to talk to you about that claim and understand your responses back, whether you're on a keyboard, whether you're speaking, whether you're doing whatever,

That's kind of where Watson came in to help augment Sydney. And again, designing those conversations. I don't know if you've been in a situation where you're sitting next to somebody and they're talking and you say, oh my God, I can't believe they said that. Well, you have to engineer that out of conversation.

the conversations that you have with members so that, you know, all of the members are delighted. And one of the things I'm proudest of is when by our work together, we have members that are thanking Sydney when we're working with them with artificial intelligence, responding to their questions, just as if Sydney was a fully human worker, right? And that

That's what I get delight from is when we've been able to change a member experience and work through that. All of the things members might ask. Now, are you taking real life conversations, looking at them and feeding them to Sydney and saying, okay, in the last two years, these are all the phone conversations we've had with our members. These are the kinds of things they ask. Is that where it starts?

we build what we call the ontology of the conversations. You know, how are we making sure that as we get the interactions noted down for our members or providers into our system,

whether it's a phone call, whether it's a chat, whether it's basically even they came to the website and they clicked through specific things, right? So we are noting those down. We are kind of creating what we call a graph model and a flow of when a member asks this, the next question possibly is going to be this.

If you give a yes to that answer or no to that answer, they're going to probably ask you this. So that kind of flow is... Sydney can be thinking two and three steps ahead. Exactly. So Sydney is thinking two or three steps ahead and making sure that the anticipation of what you're going to be doing and beyond Sydney, our overall system is thinking two or three steps ahead and predicting...

proactively those conversations as well as those interventions that we need to give to the members. So really using AI, you know, Watson as a backbone to this, Sydney is basically what we call the human-centered design focused interaction and engagement system that sits on top of the backbone of the AI as the data at the bottom.

So that basically is layered way how Sydney is able to answer the questions that we have irrespective of what type of question it is, because our ontology of the data, as well as the AI that we have built is very, very rock solid. And that is, and the good thing is that it's a gift that keeps giving because the more data we collect, the more the system it gets. Wait, Glenn, can you stump Sydney?

Can you ask Sydney a question he can't answer? I'm sure it's possible to. And then, you know, when we get into that situation, what we want to do is we want to bring the member to a human agent.

so that the member is satisfied seamlessly, right? So that there's no daylight at all, regardless of how the member wants to connect with a human agent. A lot of members are dealing with time challenges and they don't want to call up anymore. They just want somebody to be able to chat with. We try and respond to all that. And then if somebody needs a human agent, then we go there.

Yeah, yeah. What did your members tell you about, either explicitly or implicitly, about what they wanted? You know, we've been through this. We've just been through a year and a half of craziness. Yeah. You know, where everything's being turned upside down. I'm curious, what have you learned from them over the last stretch? Is what a member wants today very different than it was two years ago? I mean, I definitely, Malcolm...

If you look at that, you know, the terms that you use in healthcare are very, very complex. And it's very difficult for people to understand what my copay is. What is an out-of-network? What is an in-network? What does a claim that needs a pre-authorization mean to me? So if you look at the conversation that we were having before, they were really very hardcore healthcare-oriented conversations. And the transparency to healthcare

to what I'm going to pay was not there. So this was an industry wherein, you know, you are going to buy insurance and you're going to buy a product without really understanding what I'm going to get at the end of the day. What we did and basically what our customers actually demanded from us is that irrespective of the channel that they come to us,

what we call here at Anthem connected experiences. We want to build the connected experiences, whether they come to us from a phone call, whether they're chatting with us, they're having a web interaction, whether they're in the provider's office. How do we make sure that we connect the experience end-to-end? Now, once we connect the experience, we want to make sure that we are building a very human-centered design way of answering their questions. So it is as simple as making sure that we connect

provide them a nudge on probably this is what you're looking for. And that clicks with them and they say, yeah, that's what I was looking for. So that input simple interaction really helps to make sure that you make the member feel good.

Having the ability to text, having an ability to get the answers while you're cooking your dinner and you can text and say that, hey, could you please tell me what will my copay for the next visit I have with Dr. X? And you go ahead and start cooking your dinner. And when you come back, you have a text back out there which tells you exactly what it is.

And the beauty of it is that we had a very constant loop out there. You know, the technologies that we use that allowed us to have a constant feedback on those complex interactions that we were having. And that's where IBM team and we worked together and kind of figured out, OK, what will be our game plan? What did you learn from...

working with other people on the Watson platform that helped Anil and Anthem. What did you bring to them from what you've learned from others? So what we've tried to do with Watson, and when Watson first started, we thought that everybody wanted a bespoke suit. And so we'd kind of go on a journey together to make a bespoke suit. And what we found the clients really wanted was, well,

look, I want you to show up with the suit partially done to answer some of the basic things. And then I want to make it my own, right? So show up ready to go so that we can get into production, answering questions in a few months.

And then we will work together to radically customize and tailor that experience. That's been my biggest learning. Right. So whether it was in financial services or health care or telco or, you know, there's about seven or eight dominant industries.

We tried to make a series of industry specific cartridges so that Watson came kind of pre-trained. Right. So that we were ready to go quickly. And then the second learning was we needed to show up with the right people because remember, you're creating a conversational interaction with someone. Right. So you,

You've got to make sure that people are designing the words correctly and the user experience, right? Those are the two things I think that we brought that tried to help Anthem accelerate. I mean, Anil said something that I thought was fascinating. You're talking about designing a system with empathy. And I'm curious, first of all, what does empathy look like in an AI system? And B, has anyone ever, has any non-healthcare person

player ever asked you, Glenn, did it put empathy in the system? Clients outside of healthcare are less focused on empathy. They are focused more on making sure to get the information out there correctly, especially in highly regulated industries. How you deal with empathy in an AI system

It's all based on the choice of words that you use and the verbal inflections that are present when you have a voice response, right? And you and I, when we're talking right now with Malcolm, with Anil, with whomever, we can just by the words that somebody chooses, we can know whether

it matters to them about what we're talking about, right? And so we try and build a lot of those human characteristics into all the responses as compared to just getting the information right. Just telling people you're sorry, right? Those are the types of things that you have to engineer in as compared to just being flawlessly precise about the answer. Yeah. Wait, one last question for both of us. This has been such a fun conversation. We talked about

10 years ago when you guys started talking, and then this transition moment five years ago. Now let's go five years in the future. So let's imagine it's 2026 and three of us are talking again. I want to know what problems you're trying to solve then. The problems we are trying to solve at that time would definitely be much different than where we are. But what I can tell you before I get there is that

We want to make sure that in the next five years, Anthem...

is treated like a platform company, which is focused on creating these solutions with the help of our partners that really meet the need of the members in the journey that they have from a healthcare perspective. And we do want to pivot from a sick care to more proactive and predictive care and wellness for our members. So we're going to double down and keep working on that because it's something that never ends and it's going to

keep going in the years to come. I'm still processing this fantastic idea about taking a photo of your meal, of your plate of food.

and getting instant feedback and analysis on that. First of all, so Sydney gets all these pictures of my food, gets a sense of what I'm eating over the course of the given day. It's the idea that, so I'm thinking about this five year from now conversation. So five years from now, I might be taking a photo of everything. And then at the end of every day, Sydney texts me and says, Malcolm, you should be aware of the fact that

your nutritional patterns of the last few days. You need to eat a few more vegetables or you'd be useful to have some. Is that what we're talking about here? But I think this idea is fantastic because there is no, we have no way of making any nutritional sense of the stuff we, unless you spend two hours on a

On Google, before you make your dinner, how do you know whether the sum total of the things you eat in a given day is going to be, is optimum? I love this. I want this now. Can I, do I have to wait five years? Can I have? No.

Guys, it's been a really, really fun conversation. I really appreciate you taking the time. Anil, Glenn, have a wonderful day. And the future cannot come fast enough, at least for me. So bring it on. I'm waiting for it. Thank you very much for having us here, Malcolm. Oh, yeah. Awesome. Thanks, Malcolm. Bye, guys. Understanding customer needs has become even more important in the wake of COVID-19.

Companies like IBM and Anthem are learning to leverage technology to deliver a more personal experience, a crucial part of our evolving healthcare system. Thanks again to Anil Bhatt and Glenn Finch for talking with me. I learned a lot.

Smart Talks with IBM is produced by Emily Rostec with Carly Migliore, edited by Karen Shakerji, engineering by Martine Gonzalez, mixed and mastered by Jason Gambrell and Ben Tolliday, music by Gramascope. Special thanks to Molly Socia, Andy Kelly, Mia Lobel, Jacob Weisberg, Heather Fane, Eric Sandler, and Maggie Taylor, and the teams at 8 Bar and IBM.

Smart Talks with IBM is a production of Pushkin Industries and iHeartMedia. You can find more Pushkin podcasts on the iHeartRadio app, Apple Podcasts, or wherever you like to listen. I'm Malcolm Gladwell. See you next time.

Hello, hello. Malcolm Gladwell here. I want to tell you about a new series we're launching at Pushkin Industries on the 1936 Olympic Games. Adolf Hitler's Games. Fascism, anti-Semitism, racism, high Olympic ideals, craven self-interest, naked ambition, illusion, delusion, all collide in the long, contentious lead-up to the most controversial Olympics in history. The Germans put on a propaganda show, and America went along with all of it. Why?

This season on Revisionist History, the story of the games behind the games. Listen to this season of Revisionist History wherever you get your podcasts. If you want to hear episodes before they're released to the public, subscribe to Pushkin Plus on Apple Podcasts or at pushkin.fm slash plus.