Latest / Becker’s Healthcare Podcast / Unlocking Transformation: How Accountable AI Partnerships Build Smarter Health Systems

Transcript

Unlocking Transformation: How Accountable AI Partnerships Build Smarter Health Systems · Becker’s Healthcare Podcast

0:00
Hi, everyone. This is Lucas Voss with Becker's
0:02
Healthcare. Thanks so much for tuning in to
0:03
the Becker's Healthcare podcast series. Fantastic to have
0:06
you. Today, we're talking about how accountable AI
0:09
partnerships
0:10
build smarter health systems.
0:13
And I'm very excited to be joined by
0:15
Joe Benardello,
0:16
chief growth officer and cofounder of IKS Health.
0:20
Joe, thanks so much for being here today.
0:21
It's great to have you. Well, thank you,
0:23
Lucas. I appreciate it, and I appreciate you
0:26
you taking the time and and inviting me
0:28
to to join. Very interested in exploring this
0:32
conversation together. It's going to be great. For
0:34
those that might not know you, I wanna
0:35
start off with introductions.
0:37
Could you share a little bit about yourself
0:39
and and your work in health care?
0:41
Sure. So, you know, I think,
0:43
my first,
0:45
entry point in health care was, back in
0:47
2002.
0:48
And prior to that, I had worked in
0:50
finance, technology, and transportation.
0:53
And I think I was struck at that
0:55
time when I looked at the efficiencies that
0:57
other industries had gained through technology
1:01
and then saw the state of health care.
1:03
And that thinking really launched the start of
1:06
our our company in in the 2006.
1:09
We we opened our doors in in in
1:11
2007.
1:13
And the problem that we were trying to
1:16
solve then is the problem that we're continuing
1:19
to try and solve now. And namely, how
1:21
do we enable
1:23
position enterprises
1:26
to meet the operating and demand challenges
1:29
that they're faced with today? And and I
1:31
think the interesting part is we've been solving
1:34
the same problem
1:35
since 2007,
1:38
really focused on on how are we able
1:41
to
1:43
make a differentiated
1:44
value contribution that allows for a better impact
1:48
at scale
1:49
to quality of care, availability
1:51
of care, and cost of care. But what
1:54
has been really interesting is is that mission
1:57
has remained
1:58
eternal. That identity
2:00
has has remained unquestioned.
2:02
But what we do
2:04
and how we do have continued to evolve.
2:07
And because the challenges of health care are
2:10
ever changing and reimbursement is so different,
2:13
employee relationship
2:15
to work is so different,
2:17
labor availability,
2:18
access,
2:20
legislation,
2:21
And then finally, technology
2:23
is so different.
2:24
So we see AI
2:26
is is a tremendous enabler
2:29
to allow
2:30
impact at scale
2:32
at a cost never before
2:34
attainable for systems
2:36
in order again to to to to better
2:39
serve their patient population
2:41
and doing so in a financially sustainable manner.
2:43
And I think that's what I'm I'm really
2:45
excited about. But but I feel like the
2:48
same
2:49
problem that we've been solving is is ever
2:51
present
2:52
and is gonna be needed internally.
2:55
How do we equip our health care providers
2:57
to to do more,
2:58
treat better,
3:00
and yet have quality of life and and
3:02
pursue their own joy through practice?
3:04
I was gonna say again, you've seen the
3:06
sort of the full length of this. You
3:07
mentioned 2007. Right? It's quite a bit of
3:10
time. Right now, I feel like, though, the
3:12
technology landscape in health care is
3:15
possibly as crowded as we've seen it in
3:18
a long time. It's increasingly crowded. It's It's
3:20
only going to get more in the the
3:23
next year, I'd say.
3:24
From your perspective right now, how can organizations
3:28
identify
3:29
the right partner for their needs, and what
3:31
are some of the factors that you would
3:33
highlight
3:35
that really matter most
3:37
in that decision?
3:39
Yeah. And I think that we're a bit
3:40
in the age of parlor tricks, and we're
3:43
moving into the age of value. And,
3:45
you know, if you look at so many
3:47
of these companies that have emerged,
3:51
their big risk point
3:53
is in identifying what they're going to develop.
3:56
And then their big risk point is, am
3:59
I going to get this sale done?
4:02
But these models are generated that as soon
4:05
as the sale is complete,
4:07
the IT vendor captures all of their value.
4:11
And all of the risk now has moved
4:14
on to that
4:15
health care provider system
4:17
that is is looking for relief for for
4:20
one or more problems.
4:22
And and really what they're end up being
4:24
sold are capabilities
4:26
rather than solutions,
4:27
and now they've assumed all the risk
4:31
for capturing the value of those capabilities.
4:34
So my my encouragement
4:37
is
4:38
look for
4:40
IT that that
4:41
solves problems
4:43
rather than just gives capabilities.
4:45
And and I think that sometimes problem solving
4:48
means going beyond
4:50
a simple
4:52
single solution
4:53
and really addressing more of an interconnected
4:56
workflow.
4:58
And that interconnected
4:59
workflow follows the patient journey.
5:02
And if a vendor is able to make
5:04
a value impact on an interconnected
5:08
workflow,
5:09
they can start to take financial accountability
5:13
for the realization of those gains.
5:15
So one of the primary things that that
5:18
I would encourage organizations to do is
5:21
understand the problem that you're trying to solve.
5:25
And as you're understanding that the problem that
5:27
you're trying to solve,
5:29
define what better is,
5:31
and then look for partners
5:34
that are willing to stand behind
5:37
better
5:38
rather than simply
5:40
deliver a capability that now all that responsibility
5:43
is on to that that physician organization
5:46
to implement,
5:47
integrate, and and ultimately change manage to success.
5:51
And I think that's been one of the
5:53
major ways in which health care IT,
5:57
for all the good it's done, has failed
5:59
to deliver up to its promise. And if
6:00
we look at at at the age of
6:02
health care IT starting
6:04
maybe in 2007,
6:05
that's a good place because because that's right
6:07
at the time of meaningful use. You see
6:10
that we spent billions of dollars
6:13
on on health care IT. We've created household
6:16
name companies,
6:17
and yet care delivery has not gotten any
6:20
better in terms of outcomes,
6:22
cost,
6:23
access,
6:25
or even job satisfaction for our physicians.
6:28
And that accountability
6:30
gap,
6:31
I think, is economically
6:33
unsustainable
6:35
and probably ethically and morally unsustainable.
6:39
So I would encourage
6:42
CIOs, CMOs,
6:43
CEOs, CFOs, the that that entire buying suite
6:48
to think about
6:49
how will this solution be accountable
6:53
to the care teams that are on the
6:55
front lines and ultimately the communities they serve.
6:59
Yeah. You mentioned the the journey piece of
7:02
this. Right? The patient journey and and certainly
7:04
that the clinician is connected to as to
7:06
that as well.
7:07
And and I know that you've talked quite
7:08
a bit about the the holistic piece of
7:10
this too. Right? Understanding organizations and patients, not
7:13
just from from a clinical perspective, which is
7:14
the main focus, but also behaviorally, financially, etcetera.
7:18
That's really, really key.
7:19
From the partnership perspective then, what you just
7:21
touched on, how can the right partnerships and
7:24
specifically technology then enable
7:27
that holistic view and essentially improve outcomes as
7:30
well?
7:31
So so, you know, if you think about
7:33
what is it that we want to do,
7:35
we want to
7:37
increase access.
7:38
We want to give our providers
7:41
time and meaning
7:43
and and insight
7:46
and and I should say clinicians rather than
7:48
providers.
7:49
And then in addition to that, we wanna
7:51
make sure that things are financially
7:54
sustainable and viable. And so you start to
7:56
think about those goals. And then you think
7:58
about
7:59
what is it that
8:02
technology and specifically,
8:04
agentic AI technology
8:06
are able to give us. I think one
8:08
of the things that agentic AI technology is
8:11
able to give us is
8:13
predictive
8:14
reactive
8:16
knowledge.
8:17
So if we start with knowledge of patients,
8:20
and we know them by by their social
8:22
determinants of health, their ZIP codes, their marital
8:24
status, how many kids they have, what their
8:27
job is. We understand,
8:29
what their,
8:30
insurance is. We understand them clinically.
8:34
We understand all of these things. We can
8:36
start to anticipate them behaviorally.
8:39
And I think understanding our patients
8:41
behaviorally
8:43
will start to inform
8:46
how we treat them, where we treat them.
8:48
But at the same time, we need to
8:50
understand our clinicians in the same way. We
8:52
need to understand what their preferences are, how
8:55
they treat.
8:56
In the same way, we need to understand
8:58
our financial partners, our payers in the same
9:01
way. And if you start to think about
9:03
these as as as almost interconnected gears, you
9:06
start to create a flywheel effect. So imagine
9:09
a
9:11
a patient that is in a a value
9:14
based contract that that we understand clinically and
9:18
financially
9:19
makes an appointment. And yet we see that
9:21
that appointment is three months away and at
9:24
3PM.
9:25
And we know that they are a single
9:26
parent and that they work a nine to
9:28
five job.
9:30
Our anticipation
9:32
of that person missing that appointment is gonna
9:34
be pretty high.
9:36
So we have to create interventions
9:38
to make sure that that that that patient
9:40
shows shows up
9:42
and or find another time. So anticipating a
9:45
propensity to keep keep an appointment.
9:47
At the same time, understanding patient responsible, their
9:51
propensity to pay. Are they able? Are they
9:53
willing? Do they have knowledge of what they
9:55
start to pay? So you start to think
9:57
about those types of work flows. Based on
9:59
what that appointment is, we can start to
10:01
anticipate
10:03
and do some pre charting for that physician
10:07
about why this person is coming. So now
10:09
now imagine that that clinical workflow
10:12
where the MA and the physician prior to
10:14
a visit
10:15
have a session where where they're doing a
10:17
huddle, and they're looking over the prechart and
10:19
they're understanding,
10:20
you know, why missus Smith is coming and
10:22
what are the likely things that we're going
10:24
to be seeing.
10:25
Maybe based on the way that physician treats,
10:29
we can even anticipate
10:31
what types of tests they would run.
10:33
Based on that, we can reach out to
10:35
the payer and understand the types of documentation
10:38
that it was going to to to improve.
10:40
All of these things are giving incremental time,
10:43
insight, and value. These interconnected
10:46
workflows of
10:48
anticipating,
10:49
is the visit going to happen? What's going
10:51
to happen in the visit? How is that
10:53
visit going to be valued and reimbursed?
10:55
What types of tasks are going to be
10:58
needed to create efficiencies
11:00
that doctors shouldn't be doing, that shouldn't be
11:02
happening in the exam room? And then safely
11:05
doing those beforehand
11:07
through the use of technology. And what's what's
11:09
incredibly
11:10
important to us is that that all of
11:12
our solutions have a human in the loop
11:15
because these technologies
11:17
are
11:18
they're prediction machines. They're betting machines
11:21
in in a very real way. And if
11:23
we don't put a high level, high discretion
11:27
human in the loop, what we're pushing is
11:29
all of that editing, auditing, and insight onto
11:32
the clinician, and the clinician should be focused
11:35
on the patient. And so it is a
11:38
both a a safety lever,
11:40
but it is also an efficiency lever
11:44
and a burden lever. So we're pulling
11:47
not only time off that physician, we're pulling
11:50
meaningless
11:51
task off that position. And what we're really
11:54
building is a trust
11:55
type and an extended team,
11:58
and yet that team is largely technology.
12:02
They start to think about the power of
12:05
that.
12:06
Now all of a sudden, you can have
12:09
a vendor contract relationship
12:12
that takes responsibility
12:14
for a large subsection of the chores. And
12:17
I define chores as things in health care
12:19
that that that doctors,
12:21
MAs shouldn't be doing, things that shouldn't be
12:23
happening at the front desk or in the
12:24
clinic. Imagine,
12:26
an entity that can take a large responsibility
12:29
for those chores,
12:31
and their commitment back to that organization is
12:33
we're going to move access by x percent.
12:36
We're going to to increase
12:39
position
12:40
engagement by y percent. We're going to increase
12:43
productivity
12:45
by by by z percent.
12:47
And finally, we're going to increase
12:50
organizational
12:51
profitability.
12:52
We're going to create $20,000,000
12:54
in EBITDA.
12:56
And and if we don't,
12:58
this is our financial,
13:00
component of this. It goes well beyond this
13:02
notion of of of SLAs
13:05
into a true shared partnership model
13:09
where your technology vendor, again, isn't
13:12
just giving you a tool.
13:15
They are walking the journey with you to
13:18
make sure that you get the value from
13:20
that. I think that is a sea change
13:23
in thinking
13:24
that will quickly
13:27
disrupt
13:28
the, you know, this
13:30
collection of mushrooms after a rain in terms
13:33
of all of these different technology companies. There's
13:36
a new one every day.
13:38
I end up doing the same thing that
13:40
someone else is doing,
13:42
and yet none of them, I think, are
13:44
standing up to, and this is how we
13:46
make your organization
13:48
better,
13:49
more joyful, and more profitable.
13:52
And that's the sea change thinking that I'm
13:54
waiting for. So the human in the loop
13:56
piece that you've mentioned, which is really, really
13:58
important, you highlighted some of the benefits of
13:59
actually following that approach and making sure that
14:02
there is a human in the loop and
14:03
you have the human expertise shared right and
14:06
what the benefits it brings.
14:08
How important is that approach then to scalability?
14:12
Right? Is it a necessary component for organization
14:15
to be able to support adoption at scale?
14:19
Think so. Because a few things. One is
14:23
our doctors
14:24
are tired of being blamed
14:28
for
14:29
the lack of value realization around health care
14:32
IT.
14:33
And many times,
14:35
the story that comes through, health care IT
14:38
is this story around
14:40
if only the doctors would do this, if
14:43
only physicians,
14:45
had better change management. And it's why do
14:47
we need physician change management? It's because we're
14:50
asking physicians
14:52
to drive different behavior
14:55
rather than finding technology that's meeting physicians where
14:58
they are. And in order for technology to
15:01
to meet physicians where they
15:03
are, I think the human in the loop
15:05
component is is absolutely critical.
15:08
Let's not make our clinicians
15:11
QA agents.
15:12
Let's not make the our our our
15:16
physicians editors and auditors and cut and pasters
15:19
and all of these other things that we
15:21
continue to do. What's interesting for me
15:25
is that before
15:26
I
15:27
tech optimize
15:29
any workflow
15:31
in my company, there's a few things that
15:33
we do.
15:35
One is we say, okay. How is this
15:37
going to be better for patients? And we
15:39
have to be able to describe that in
15:41
a sentence. It can't be
15:42
a a
15:44
a war and peace paragraph. I wanna know
15:46
really plainly. This is gonna be better for
15:48
patients because
15:49
this is gonna be better for the clinician
15:52
experience because what's interesting is aligning those two
15:55
things is not very hard. The third one
15:57
is the magic trick, and this is going
15:59
to create financial value for the organization
16:03
because
16:04
and if we can do those three things,
16:06
then we know this is something that needs
16:09
a technological
16:10
intervention
16:11
and and warrants it and and will create
16:14
value from it. The other aspect of it
16:16
is before we do a technological
16:19
intervention,
16:20
we want to understand the context.
16:24
I see a lot of health care IT
16:26
that has capability without context. And I think
16:29
clinical context,
16:31
technical context, and emotional context need to be
16:34
built into your work
16:36
flows in a caregiving
16:38
industry.
16:39
So if we can create that context within
16:41
the workflow,
16:43
then what's interesting is the human component that
16:46
is left
16:48
is highly discretionary
16:50
and highly valuable.
16:52
And so over time,
16:55
what I used to do with 10 people,
16:57
I do with half a person.
17:00
But the value of that half a person
17:03
is significantly
17:04
higher
17:05
because everything that was nondiscretionary,
17:08
everything that was rule based, everything that you
17:11
could reasonably predict and assume
17:13
has now been covered by the machine.
17:16
And it's only that most critical piece that
17:18
that human becomes an important,
17:21
component of going forward.
17:23
In the development process,
17:26
humans are critical for for for reinforced learning
17:29
through human feedback.
17:31
You want to to speed quality and value
17:34
in in any tech led system that reinforced
17:38
learning through human feedback becomes a a a
17:40
critical angle.
17:42
Yeah.
17:43
Joe,
17:44
such a great conversation. Thank you so much
17:46
for all of your insights. I I do
17:47
wanna know again because you have such a
17:49
such a great detailed and 10,000 foot overview
17:52
combination of this.
17:54
What are you what are you excited about
17:56
for 2026?
17:57
What are some of the shifts and trends
18:00
that you're seeing that leaders should really prioritize?
18:03
You meant you mentioned iGentic AI, which is
18:05
going to be critical,
18:06
but what are some of those things that
18:08
that you're thinking that are important to really
18:11
unlock the full value of AI and digital
18:14
in 2026
18:15
and and moving beyond that?
18:18
I would go back a little bit to
18:20
a couple of the concepts I I would
18:22
talk about. So
18:23
I think responsible
18:25
is
18:26
really critical.
18:28
Safety built into workflows.
18:32
The challenging thing about AI
18:35
is as much as we believe that we're
18:37
training the AI,
18:39
the AI is training users equally.
18:43
And as it gets better and that trust
18:46
height
18:47
increases,
18:50
the rate of errors is gonna go down.
18:54
But the more we trust it,
18:57
the actual ability of a catastrophic
18:59
error is going to go up. And so
19:02
how do we manage against the catastrophic
19:05
error
19:06
when we trust the self driving car so
19:09
much?
19:11
And I think that that that so so
19:14
so systems
19:15
that have responsible
19:16
checks
19:18
in them, I think, are gonna become increasingly
19:22
important. So so I would think about responsible
19:26
safety checks built into these workflows
19:29
that does that
19:31
without making
19:32
it a
19:34
a dampener on the economic
19:36
net benefit because
19:39
health care dramatically needs, an economic net benefit
19:43
from AI, which it hasn't gotten from so
19:45
many other technologies. And I think that AI
19:47
can do that. I think the second piece
19:50
to to to reiterate
19:51
a a bit
19:53
is organizations
19:55
need to demand accountability.
19:57
So I will say that
19:59
my company sees accountability
20:02
as a differentiator
20:04
and a market opportunity.
20:06
I believe that if I take greater accountability,
20:08
I will win more clients and and grow
20:11
my business. And I believe that I have
20:13
a moral obligation in serving the health care
20:16
industry to to to do this.
20:18
But it's really up to
20:22
the CIO, the CEO, the CFO, the CMO,
20:26
COO
20:27
to demand accountability
20:30
from their tool vendors
20:33
and and and stop being the system integrator
20:37
that
20:38
makes excuses
20:40
when technology doesn't deliver. Oh, it's because of
20:44
our implementation,
20:45
our change management. Oh, we bought this from
20:47
this vendor and this from this vendor, and
20:49
now they're arguing who's responsible.
20:52
Get beyond that
20:53
and create structures that demand participation
20:56
and accountability.
20:57
And if we focus on responsible
20:59
safety and we focus on accountability,
21:03
the result should be time, meaning, and money.
21:07
We should increase care delivery time
21:10
while reducing work burden.
21:13
By increasing care delivery time, we start to
21:15
solve access in a meaningful way.
21:18
Meaning,
21:19
a big part of burnout
21:21
is not just how much time we're asking
21:23
our physicians to do, but but also the
21:26
tasks that we're asking them to do. And,
21:28
you know, the the boogeyman for a long
21:30
time was clinical documentation.
21:33
And yet we've largely fixed clinical documentation.
21:36
Physician burnout has gotten much better. There are
21:39
other
21:40
meaningless tasks, whether it is inbox. AI is
21:44
a great tool around inbox. Prescription refill, clinical
21:47
documentation management,
21:48
pre charting,
21:49
all the order entry, all of these other
21:51
areas. So so let's take meaningless tasks. Let's
21:54
take things below the license of people.
21:57
And what's interesting is I'm a firm believer
22:00
that if you get time and meaning right,
22:02
you find ways to increase money.
22:04
And
22:05
there was a time that that doing the
22:07
right thing was an imperative.
22:09
Right now, doing the right thing in health
22:11
care seems to be a luxury unless we're
22:13
able to create financial value for these organizations.
22:16
So that's what I would think about in
22:18
in in '26 and hopefully '27, 2830.
22:22
I think these are timeless.
22:24
Yeah. I think that we'll all continue to
22:26
get better in how we deliver it, and
22:29
that's what I'm passionate about today. I want
22:31
to make
22:33
health care better at scale.
22:35
I want to make our patients feel known
22:38
as they come in the door. I wanna
22:41
make treatment specific
22:44
to that patient and that physician and yet
22:47
do it in a scalable, replicable way that
22:50
makes health care easier to get and more
22:53
affordable.
22:54
Yeah. That's certainly a goal, to keep in
22:56
mind beyond the 2026
22:58
as you pointed out. Very, very crucial. Joel,
23:01
thanks so much for your time and insights
23:02
today. What a fantastic conversation. Thanks for being
23:04
here. Thank you, Lucas. Enjoyed it. Great to
23:07
have you. We also want to thank our
23:08
podcast sponsor IKS Health. You can tune in
23:10
to more podcasts from Becker's Healthcare by visiting
23:13
our podcast page at bekkershospitalreview.com.