0:14

Welcome, everyone, to the AI

0:17

and Business Podcast. I'm Matthew

0:19

D'Amelio, senior editor here at

0:21

Emerge Technology Research. Today's guest

0:23

is Shane Bray, Chief Customer

0:25

Experience Officer at Blue Cross

0:27

and Blue Shield of Louisiana.

0:29

Shane joins me on the

0:31

program today to discuss how

0:33

the convergence of health care

0:35

and financial services presents a

0:37

compelling opportunity for the integration

0:39

of AI to elevate patient

0:41

experiences and customer interactions in

0:43

insurance workflows. Later, we take

0:45

a closer look at the advantages

0:48

of generative AI in addressing problems

0:50

like the interoperability of different health

0:52

care IT systems and giving caretakers

0:55

a deeper understanding of patient behaviors

0:57

and sentiment analysis. Today's episode is

0:59

sponsored by Unifor, and without further

1:02

ado, here's our conversation. Shane,

1:09

thanks so much for being with us on the program

1:11

this week. Yeah, it's great to be here, Matthew. Thanks

1:13

for the invite. This is a really great opportunity

1:15

for us. I know we've been

1:17

talking about FinServe customer experience workflows

1:19

and health care patient experience workflows

1:21

separately on the podcast up until

1:23

this point. We've got a grand

1:25

opportunity to talk about them together

1:27

to a certain extent, especially for

1:30

the health care insurance space. In

1:32

that way, I want to just

1:34

kind of set the table in

1:36

terms of maybe how to think

1:38

about the challenges too, even

1:41

in this space where they're the

1:43

same person, even if they're going to

1:45

have two different kinds of experiences across

1:47

two different workflows. What do you see

1:50

as the biggest challenges right now in

1:52

customer service workflows in health care, particularly

1:54

health care insurance? Well,

1:56

I think when you think about customer experience, both outside

1:58

of the business, I think of healthcare and within

2:01

healthcare. And then within healthcare, it becomes a

2:03

couple different things. It becomes a member experience

2:05

when they're dealing with their health, their health

2:07

insurance provider or payer. It becomes patient experience

2:09

when they're actually dealing with, with the hospitals

2:12

and the doctors. So there's a lot of

2:14

different types of experiences and outside of the

2:16

health, healthcare system, people are looking for things

2:18

that are easy. They're convenient in a lot

2:20

of ways. They're looking for something

2:22

that they really want to be consumers of.

2:25

So, so they have some expectations of what

2:27

they want to happen often in, in

2:29

customer experience, within healthcare. A lot of times you don't

2:31

want to be there and you don't want to be a

2:33

consumer of the services that you're buying. It's, it's something that,

2:36

that, that happens to you. And now you have to be

2:38

a consumer of these services. So it

2:40

is a completely different world in terms

2:42

of patients and member experience is, is

2:44

something that you don't have a lot

2:46

of control over. You, you get sick.

2:48

So one of your loved ones to

2:50

get sick, you have to, you have

2:52

to navigate a system. That's extremely difficult.

2:54

That's as unknown costs and expenses versus

2:56

the, the other customer experiences

2:59

where you know what to expect.

3:01

You're excited about the product and oftentimes based

3:03

on what you can't afford or what experiences

3:05

you're looking for, you have a lot of

3:08

control over the experience and so there's just

3:10

a lot of differences between the healthcare space

3:12

and the commercial space in general. Absolutely. And

3:14

I mean, I think you're putting a fine

3:18

point just based on the, on

3:20

the very essence of where they're

3:22

entering the workflow involuntarily. And we

3:24

know across especially insurance, you know,

3:26

oh, there's, there's emotional obstacles

3:29

to work around. There's, there's different

3:31

kinds of impacts that are going

3:33

to have on sentiment, particularly whether

3:36

or not the customer in question

3:38

chooses to be there. I know

3:40

this is particularly poignant for collections

3:42

workflows. Just in terms of streamlining

3:44

healthcare workflows, particularly when it comes

3:47

to the clinic and the pharmacy

3:49

and the payer, where are we

3:51

seeing in terms of like streamlining

3:54

Those processes where the rubber hits the

3:56

road, that there are the biggest frictions

3:58

for the patient staying in patient. Experience.

4:00

Well. When you think about the health

4:02

care experience let's say you have a

4:04

covert diagnosis right? and it's not. It's

4:07

not a severe case of toted, you

4:09

have flu like symptoms, maybe you you?

4:11

you don't go to the office neighbors

4:13

to hell of a tele health is

4:15

it.your doctor for a few minutes he

4:17

confirms your symptoms, maybe describes er packs

4:19

with it and then you recuperated home

4:21

and so it's of his even was

4:23

something like as like a not severe

4:25

over diagnosis is not so bad but

4:27

the sicker you get the more complex

4:29

those those those workflows. Get as sick

4:31

about someone who has just may be found

4:33

a lump in their breast and so they

4:36

they notice a lump in their breasts and

4:38

they go see their primary care physician. The

4:40

primary care physician takes a look at it

4:42

realizes probably something that needs and diagnostic. So

4:44

this and them for for imaging. So now

4:46

you're no longer within the context providers office.

4:48

Now it's off to imaging to do some

4:51

unknown tests and then imaging. maybe find something

4:53

so they send you back here for your

4:55

primary care provider than maybe they send you

4:57

to at the in this case they would

4:59

send you to a specialist. Specialist others are

5:01

a few more test as a few more

5:04

things in the maybe they need to send

5:06

you to a have another specialist which was

5:08

happens often so it is d as it

5:10

is that the complexity determines the complexity of

5:12

the navigation and I'm often times when it

5:14

when it something simple it's not so bad

5:16

you couldn't You know you can take a

5:19

few knox and using the to take a

5:21

few inconveniences because you're going to be better

5:23

than him three to seven days but it

5:25

will get to the point where do you

5:27

have a real illness or a chronic illness

5:29

that navigation becomes a real problem. Because

5:31

a lot of times as the the

5:33

providers themselves don't understand it, they understand

5:36

their domain and they know I know

5:38

where you need to go next. But

5:40

there's the navigation piece of patient experience

5:42

gets more and more complex the more

5:44

and more sick you get. So it's

5:46

a bit of an inverse relationship of

5:48

that. The sicker you are, the harder

5:50

it gets it. and it makes it

5:52

very difficult for all sorts of that

5:55

aspects. Financials? That physical? Mental? Yes, absolutely.

5:57

And I know from the conversations we've

5:59

had else. Where in the healthcare

6:01

space obviously healthcare systems are aspiring

6:03

to make sure that you know

6:05

it. They're completely focused on the

6:07

patient and even just given the

6:09

regulatory burden in the healthcare system,

6:11

there ends up being different priorities.

6:13

Just as an example, Eat When

6:16

the A Were when the A

6:18

Cia was passed bill Little over

6:20

ten years ago, it was written

6:22

in such a way And hipper.

6:24

I know this is written in

6:26

this way as well. A Pray

6:28

it prioritizes he Hr data. For

6:30

the ends of payment, not for the ends

6:33

of care and I know that that ends

6:35

up in suits frictions for the patient experience

6:37

A in to get in how they interact

6:39

with the health care system especially for that

6:42

dynamic you were just men in mentioning as

6:44

the more they get six the more the

6:46

last they tend to get in these systems

6:48

in terms of balancing what has to be

6:51

important to the health care system. From a

6:53

regulatory standpoint from attack debt burden standpoints vs

6:55

what's important to the pace and their own

6:58

care. How to How are we Closing. That

7:00

Gap or wouldn't need maybe even taking a

7:02

step back from their what is that Gap

7:04

look like that we need to close. I

7:06

mean that's that's a great question and I

7:08

would probably spend a lot of eye on

7:10

that one or more. So if you look

7:12

at the have been talking about the A

7:14

C A one of the Ac A do

7:16

they They they didn't started to your point

7:18

necessarily dictate the outcome they they dictated what

7:20

A T Wage Fear quality Health plan means

7:22

to cover So they said it is planning

7:25

to cover these things that our those things

7:27

great. But you know that at that point

7:29

that there's no focus. necessarily on on call

7:31

your outcomes it says saying that hey your

7:33

insurance has to pay for this then the

7:35

hospitals have different metrics you know there's they're

7:38

looking at his admissions discharges open beds or

7:40

how many people they can put in their

7:42

oh or number of patients they can move

7:44

through so the hospitals are very much looking

7:46

at it like a business which they should

7:48

do because it they they are business but

7:51

when when when that happens it is it's

7:53

really easy to to lose focus of what

7:55

business there and i think that that that

7:57

business is really the human experience business to

8:00

care of the sick, the ill. And then

8:02

when we start to look at the things

8:04

that we are measuring, let's say readmissions and

8:06

infection rates, those things are really, really important.

8:09

But what are what the

8:11

patients are thinking about is they're thinking

8:13

about, am I getting the right treatment?

8:15

Do I have the right opinion? What's

8:17

the delays in diagnostic? You know, sometimes

8:19

it's a day or two, again, in

8:21

the example of COVID, sometimes it's weeks.

8:23

What are my diagnostic delays going to

8:25

look like? What is my diagnostic accuracy

8:27

going to look like? We don't measure

8:30

those types of things today. We don't look

8:32

at things that our patients are thinking about

8:34

in terms of, am I

8:36

going to be okay? And am I going to

8:38

live through this? And am I paying for things

8:40

that I don't need? And then I think one

8:43

of the other things that is a big gap

8:45

is that hospitals and health

8:47

systems tend to treat individuals as

8:50

a group. You know, they look

8:52

at the group metrics and associate

8:54

things as big

8:56

chunks of people and big chunks of data

8:59

versus being treated really as an individual. And

9:01

yes, that's difficult. But in terms of the

9:03

things that we want to improve, we really

9:05

have to look at the individual

9:08

aspect of it because when it comes down

9:10

to it, it is very personalized. It is

9:12

very individualistic. And not every cancer case is

9:14

the same. Not every COVID case is

9:17

the same. So treating patients as

9:19

groups or as case studies are things that are

9:21

backed by significant studies while sometimes

9:23

effective really takes the individualism, I think, out

9:26

of it. But I guess to sum all

9:28

that up, we need to look at things

9:30

that are important to the patient versus

9:33

what's important to the health system

9:36

itself. Yes. I also

9:38

remember from our initial calls, just

9:40

setting up this conversation, prior authorizations

9:42

are a major pain for members

9:45

and providers as well. And

9:47

I know this is at least meant to

9:49

facilitate, at least as, you

9:51

know, a practice. It's meant to facilitate the

9:53

process. What's the pain point here? And how

9:55

is it bogging down the patient experience? Prior

9:58

authorizations, you know, there is a need. for

10:01

those processes sometimes because you

10:03

know oftentimes things are misdiagnosed,

10:05

things are mistreated, there does

10:08

need to be I believe

10:10

some oversight in terms of

10:13

the payers and

10:15

some entities looking at the appropriateness

10:18

of care. But then when it comes down to

10:20

it I think this is a fantastic application of

10:22

AI in the future is what's

10:24

the accuracy of the physician who is

10:27

requesting said treatment, how often does

10:29

it happen, how often is

10:31

it effective and really applicable.

10:33

But the issue

10:35

that I have with prior authorizations

10:37

is it not only causes major

10:39

pain for the patients, is it

10:42

causes a major pain for carers.

10:44

And now we have two of the

10:47

the major consumer groups or the two

10:49

of the major actors in the scenario

10:51

are both feeling a major pain point

10:54

from something that is imposed

10:56

typically by the payer to

10:58

ensure that costs are controlled. There is a

11:00

lot of I think there certainly

11:02

is an argument for the

11:05

necessity of prior authorizations but it requires back

11:07

office staff, it requires a lot of

11:09

administrative work, how much money are we

11:11

spending on those types of things versus just

11:13

the treatments themselves and enabling

11:15

physicians to do their jobs. So I

11:18

think a lot of the controls we put in place tend

11:21

to add a significant amount of overhead, a

11:23

significant amount of pain. Prior authorizations is one

11:25

that's just got on in my skin lately

11:27

because of the significant pain that it causes

11:29

to multiple players within the ecosystem.

11:32

Yes and we're always here for the for

11:34

the latest in terms of where the problems

11:36

are popping up in in these workflows. Just

11:39

in terms of you were mentioning AI a

11:41

second ago and I think

11:43

it's very well established. I know

11:46

we have a few episodes in

11:48

this regard talking about you know

11:50

the application particularly of new generative

11:52

AI tools to help with tech

11:54

debt to help with administrative tasks.

11:56

I think that's out there also and

11:59

I know we were talking about this in our

12:01

outlines and in

12:17

terms of improving patient experiences. How in

12:19

your view can customer experience

12:22

healthcare and insurance leaders leverage data tools

12:24

to solve these problems and what's that

12:26

looking like on the ground? Yeah,

12:29

and you know I think

12:31

the answer really to most

12:33

of those entities is really

12:35

the interoperability of the

12:37

data and transparency in that

12:40

data. When we

12:42

think about efficient drug treatments,

12:44

patient goes and sees the physician,

12:47

the physician writes a prescription, has

12:49

no idea if the patient can even afford the

12:52

prescription, he just knows that it's effective for

12:56

the condition that he's decided that

12:58

day that you have. Then there's the

13:00

role of the pharmacy and they're receiving

13:02

a prescription. How is it

13:05

being paid for? Patients wondering how they're going to pay

13:07

for it. But it's

13:09

a very I think

13:11

complex system that stands for

13:14

a lot of improvement. What if

13:16

the physician's right up front knew the cost of the medication

13:18

and asked the patient, hey can you afford this? If

13:21

they can't then let's address that right up

13:23

front. Let's not even make it to the

13:25

pharmacy before you figure out your surprise price

13:27

and know what your copayment is going to

13:29

be or what the cost of the medication

13:31

is going to be. I think that one

13:33

of the issues that we're seeing today is

13:35

not necessarily people not wanting to consume healthcare

13:37

but not being able to afford to consume

13:39

healthcare and having that transparency up front

13:42

because of the interoperability of the data

13:45

that allows them to see if that

13:47

treatment is at all sustainable for them

13:49

financially or at all feasible.

13:52

I think if we had better data

13:55

interoperability one that powers AI

13:57

because AI is just a

13:59

consumer. of data that looks at trends

14:01

and synthesizes

14:03

new concepts based on data.

14:06

Let me ask you a question right there, just

14:09

in terms of where

14:11

the affordability problem just

14:14

being able to get those concerns

14:16

upfront, maybe be a little bit

14:19

more proactive in the process. I

14:22

know also that a lot of

14:24

what we talk about in terms

14:26

of AI and leveraging data solutions

14:28

is envisioning a future where we're

14:30

being more proactive or preventative about care

14:33

rather than reactive. I think even in

14:35

the payments process, I think there are

14:37

huge opportunities there, as you're just mentioning,

14:39

in terms of knowing beforehand what patients

14:42

can afford before they walk into the

14:44

room, before they're in front of doctors

14:46

to explain what they need. Where do

14:49

you see the gap there just in

14:51

terms of how far we have to

14:53

go in order to really make

14:56

those systems intuitive and then get to the

14:58

point where we're being a lot more

15:00

proactive about patients being where they belong

15:03

in the system and receiving

15:05

care that they know that they can't afford? You

15:08

know, I've actually thought about this quite a bit in the past few weeks.

15:12

It kind of stirred up some new thoughts that

15:14

I have. I think that

15:16

there's a lot of hope around

15:18

AI and how it can improve

15:20

proactive and preventative treatments

15:23

and illness. I think that

15:25

there certainly is a great

15:27

opportunity there. However, I think that

15:29

it might be a little bit,

15:33

what's the word I'm looking for,

15:35

overoptimistic, because I

15:37

think that a lot of times the outcomes

15:40

are dependent not necessarily on

15:43

not being proactive but behavioral

15:45

issues that are

15:47

innate in patients.

15:50

In some cases, I think

15:52

that patients will respond to

15:55

proactive prompts

15:58

that say, hey, maybe it's time to go

16:00

to the hospital. get that mammogram, maybe it's

16:02

time to get that colonoscopy, have

16:04

your blood pressure checked, keep your sugar

16:07

in check. But then there's also the

16:09

behavioral issues that I think

16:12

we're seeing more and more problems

16:14

with diabetes, the escalation

16:17

in obesity. You

16:20

can't be proactive enough in

16:22

some of those cases, but it all

16:25

comes down to behavioral changes that are

16:27

necessary in the patient population for any

16:29

of that to be effective. So I

16:32

think where there's a lot of hope around

16:34

AI having this proactive approach and patients really

16:36

wanting to consume that and go about it,

16:38

yeah, some people are going to do that.

16:40

But I think the issue that we're going

16:42

to have with AI is that the optimism

16:44

will probably die down a little bit when

16:46

we find out we can't really change the

16:48

behaviors. And that's what I think what we're

16:50

going to need to do is figure out

16:52

how we modify behavior versus modify technology. Roger.

16:55

And I mean, especially here at Emerge, we

16:58

take a very discerning look at hype

17:00

cycles with artificial intelligence. And these are

17:02

funny things, especially in the world of

17:04

generative AI, where yeah, a lot of

17:06

it is very much hyped. All at

17:09

the same time, the

17:11

capabilities of AI that

17:14

are concrete, and in many cases,

17:16

very seldom known, are

17:18

not publicized enough about, or

17:20

are surprising to the point where I think

17:23

we're still contradicting things we were told

17:25

in elementary school that robots would never

17:27

make art and things like that, or

17:29

could ever be creative. I think anybody

17:32

who's even seen, you know,

17:34

two seconds of a Dali demonstration

17:36

knows that that's really not the

17:38

case going forward. All that said,

17:40

just in terms of, you know,

17:42

what we previously thought artificial systems

17:44

could never do, we know from

17:47

Collections use cases that artificial

17:49

intelligence is actually uniquely equipped

17:51

to understand sentiment and be

17:54

able to really assist other

17:56

human beings and call agents,

17:58

especially with. For and

18:00

bedside manner of especially when they're

18:03

guiding customers and patience and in

18:05

in this case through very, very

18:07

difficult circumstances. It isn't just with

18:09

respect to your point right there

18:12

about behaviors and in diabetes and

18:14

in obesity I'm wondering end in

18:16

understand just from the perspective of

18:19

you know, seen We're still waiting

18:21

to see where a I can

18:23

take us. But if we're able

18:26

to take better data on sentiment

18:28

analysis about how patients are. Talking

18:30

about those challenges? what the behaviors

18:32

are that maybe we can analyze

18:34

them in a way that's a

18:36

at least ten bring a data

18:38

approach to what actually improves patient

18:40

outcomes in terms of how to

18:43

talk to them about their behaviors

18:45

and assist caretakers and providers with

18:47

bedside manner with being more persuasive

18:49

about changing those behaviors. I do

18:51

see those those capabilities at all.

18:53

Or do you see them as

18:55

as overhyped? Know? you know? Actually,

18:57

I think in that context I

18:59

think. That. There's a significant opportunity, particularly

19:01

in the development of human relations

19:04

or at least the perception of

19:06

those human relations who the as

19:08

you mentioned sentiment to the alley

19:10

All of those things Ai is

19:12

it is innately good at because

19:14

it's look at our language patterns

19:17

over forever and it knows that

19:19

to have a sympathetic response. These

19:21

are the types of things that

19:23

a person that a sympathetic person

19:25

would say M A I does

19:27

a really good job of of

19:30

replicating. that and i think the be

19:32

interesting piece there is that when people

19:34

will not respond to data they very

19:36

often will respond to relationships and i

19:38

think that that's where it's going to

19:40

be important is not so much giving

19:42

them the data and saying hey if

19:44

you don't do this the miss will

19:46

probably happens or because we have the

19:48

data and me of the case studies

19:50

and me up the metrics and weekend

19:52

so that is if you're a wave

19:54

is added controlled and year for sentence

19:56

for are tied to those up and

19:58

then hypertension all of us things follow.

20:00

But then going back to the

20:02

replicating human responses and human emotions,

20:05

it doesn't necessarily, in my opinion,

20:07

have to be a human that

20:09

is eliciting

20:11

those things for people to have

20:14

a response. I think when AI

20:16

gets to the point, and it's

20:18

very close, I think right now

20:20

with the generative AI capabilities, that

20:22

focuses less on the

20:24

data aspects of it and

20:26

more about building those relationships

20:28

with patients, that's where I think

20:30

we're going to see behavioral change. Because people

20:32

are going to have those internal feelings of,

20:35

I feel safe, I feel like this, I

20:37

can trust this, I feel like it cares

20:39

for me. And even if it's a machine,

20:41

you know, people will

20:43

respond to building those relationships. So

20:45

I think that the

20:47

human aspects of AI can

20:49

be extremely powerful in

20:52

helping to modify behaviors that

20:54

right now care management teams

20:56

are just overstressed on.

20:59

But I think that there's a lot of application, a lot of promise

21:01

there. Yes, and no shortage

21:03

of use cases across industries for

21:05

relationship management, especially where we can

21:07

collect data and have a very,

21:09

very strong sense of where it's

21:11

on the ground changing behaviors. Really

21:14

radical stuff. Shane, thank

21:16

you so much for being on the show this

21:18

week. It's been an absolute blast. You bet. Enjoy

21:20

the conversation. Thank you. Before

21:40

we wrap up today's episode,

21:42

some talking points Shane had

21:44

discussed that I think should

21:46

leave a lasting impression on

21:48

our executive listening audience. They

21:50

include that Shane began his

21:52

episode by emphasizing that streamlining

21:54

clinic, pharmacy and payer workflows

21:56

is crucial for improving patient

21:58

experiences as these are the

22:00

areas where friction tends to

22:02

occur. He also notes that

22:04

the sicker the patient is,

22:06

the more complex the navigation

22:08

becomes, citing examples of patients

22:10

with chronic illnesses facing challenges

22:12

and accessing proper care. Concerning

22:14

the gap between healthcare system

22:16

priorities and patient-centered care, Shane

22:18

highlights the importance of prioritizing

22:20

the latter over the former,

22:22

focusing on individual experiences and

22:24

concerns over group metrics. He

22:26

also notes that prior authorizations

22:28

add overhead and administrative work,

22:30

but AI tools can help

22:32

with tech debt and administrative

22:35

tasks. Better data interoperability powers

22:37

AI, which synthesizes new concepts

22:39

based on trends, accessing affordability,

22:41

and feasibility issues for patients.

22:43

Shane also raises concerns about

22:46

AI's ability to change patient

22:48

behaviors, citing examples of diabetes

22:50

and obesity. Towards the end

22:52

of the show, we discussed

22:54

modifying behaviors through sentiment analysis

22:56

in bedside manner in healthcare.

22:59

Shane concludes by citing the

23:01

importance of relationship management in

23:03

various industries and the potential

23:05

for AI to collect data

23:07

to change behaviors. On

23:09

behalf of Daniel Fajela, our CEO and Head

23:12

of Research, as well as the rest of

23:14

the team here at Emerge Technology Research, thanks

23:16

so much for joining us today and we'll

23:18

catch you next time on the AI and

23:20

Business Podcast. Thank

23:24

you. Thank

23:54

you.