Latest / Becker’s Healthcare Podcast / How AI Governance and Use Cases Will Reshape Health Systems in 2026

Transcript

How AI Governance and Use Cases Will Reshape Health Systems in 2026 · Becker’s Healthcare Podcast

• 0:00
  This is Scott Becker with the Becker's Healthcare
• 0:03
  Podcast.
• 0:04
  I am thrilled to weekly
• 0:06
  get a chance to visit with Molly Gamble.
• 0:09
  Molly is the vice chair or vice president
• 0:12
  in charge of editorial affairs at Becker's Healthcare,
• 0:15
  long term editor in chief.
• 0:17
  And we we visit with Molly
• 0:19
  regularly to talk about what stories she's walking
• 0:22
  watching closely in health care. Molly, let me
• 0:25
  turn it to you and ask you, are
• 0:26
  there a couple stories that you're watching most
• 0:28
  closely currently?
• 0:30
  Yeah. Thank you so much, Scott. It's good
• 0:32
  to be back with you. I had one
• 0:34
  story in mind that I kinda wanted to
• 0:35
  go a little bit deeper on today, and
• 0:37
  that's this predictions piece that was published early
• 0:40
  this week about health care AI in the
• 0:43
  year to come, so in 2026.
• 0:45
  Just trying to talk to some really smart
• 0:48
  and bright leaders in the field about how
• 0:50
  the conversation about AI and health care could
• 0:53
  change in the year ahead.
• 0:55
  It's certainly been a big focal point for
• 0:57
  2025.
• 0:58
  And I think as a journalist and a
• 1:00
  communicator, you know, something you always wanna look
• 1:01
  to do is make sure you're not relying
• 1:04
  on the same tropes or same jargon or
• 1:07
  mottos for too long after they're useful or
• 1:10
  effective. And I think that's something that's healthy
• 1:12
  to do at the start of a new
• 1:13
  year. So there were a lot of interesting
• 1:15
  points raised, Scott. Predictions can be tricky sometimes,
• 1:18
  you know, they come more easily depending on
• 1:20
  what the topic is, than others. But there
• 1:22
  were a few that I wanted to highlight
• 1:24
  for listeners because I thought it does signal,
• 1:26
  I think, a a healthy way to move
• 1:28
  the conversation forward about AI that that already
• 1:31
  we hear from executives all the time, you
• 1:33
  know, AI hype, AI fatigue. Those things are
• 1:36
  real. So this struck me as an opportunity
• 1:38
  to kind of pause and then look ahead
• 1:40
  and and talk again, like I said, to
• 1:42
  some some really bright leaders about what they're
• 1:44
  anticipating
• 1:45
  over the next, you know, at least couple
• 1:47
  of quarters in 2026
• 1:48
  when it comes to this form of technology.
• 1:52
  Thank you. And and tell us, what are
• 1:54
  some of the things that you're hearing?
• 1:56
  Talk to us about what you're hearing and
• 1:58
  and what people are saying.
• 2:00
  Yeah. One of the bigger ones, Scott, was
• 2:02
  that this AI governance will really move from
• 2:06
  what's kind of been a backroom exercise or
• 2:08
  backroom checkbox to much more of a strategic
• 2:11
  engine in health systems.
• 2:13
  This is a really important one. You hear
• 2:14
  a lot about governance, but when it comes
• 2:17
  down to it, the resources health systems have
• 2:19
  for governance are relatively thin.
• 2:22
  In in
• 2:23
  a couple years ago, only 16% of health
• 2:26
  systems had enterprise wide governance strategy
• 2:29
  for AI.
• 2:30
  In in 2025,
• 2:32
  the average IT budget allocation of governance is
• 2:34
  still was only about 4%.
• 2:36
  So this is something that, as one expert
• 2:39
  had told me in my interviewing,
• 2:41
  from from class, you know, he said that
• 2:43
  some systems have treated governance like
• 2:46
  infection control in the past,
• 2:48
  and there's anticipation from leaders that this will
• 2:50
  become much more vital,
• 2:52
  much more urgent, much more visible.
• 2:55
  It's gonna become a way bigger part of
• 2:57
  leadership and also, you know, technology leadership and
• 3:00
  IT leadership, but also just, like, more broadly
• 3:01
  at the CEO, CFO, COO level too.
• 3:05
  So there's anticipation that there'll be more oversight
• 3:07
  bodies from the state level accrediting agencies, also
• 3:11
  federal regulators.
• 3:12
  But rather than being reactive to those
• 3:15
  external forces, there's also anticipation of health systems
• 3:18
  will really start to beef this up on
• 3:19
  their own, and again, make it more of
• 3:21
  a strategic layer of the health system versus
• 3:23
  what could be kind of a a checks
• 3:25
  a checkbox exercise.
• 3:27
  You know, I think that's so fascinating. And
• 3:29
  I think EOS
• 3:31
  AI starts with so many different use cases
• 3:34
  and so many people throughout systems
• 3:36
  start to work on and utilize
• 3:38
  test cases,
• 3:40
  I I assume that systems feel like more
• 3:42
  and more,
• 3:43
  they need to have some guardrails
• 3:44
  around what that's gonna look like and what
• 3:47
  that's gonna mean.
• 3:48
  And so it makes a lot of sense
• 3:50
  because you you could see that where you
• 3:52
  start to see this evolution of so many
• 3:54
  different use cases
• 3:56
  and something get out out in front of
• 3:58
  itself
• 3:59
  and going differently than expected. And there's gonna
• 4:02
  be tests. There's gonna be challenges,
• 4:04
  but sort of knowing what a system has
• 4:06
  decided to test and try and where the
• 4:08
  risks are
• 4:09
  seems increasingly
• 4:10
  important,
• 4:12
  fascinating, actually.
• 4:13
  Mhmm. Mhmm. I thought so too. And it
• 4:16
  was a lot of excitement about what AI
• 4:17
  will start to do meaningfully in the revenue
• 4:20
  cycle space. So, you know, denials, these have
• 4:22
  long been one of the industry's most expensive,
• 4:24
  like, also just a demoralizing
• 4:26
  form of waste for for clinicians and for
• 4:28
  health systems and patients.
• 4:30
  There's a lot of excitement from leaders about
• 4:32
  how AI will begin
• 4:34
  to not just make coding easier or,
• 4:38
  CDI
• 4:38
  or denial management, but maybe
• 4:41
  eliminate some of those
• 4:43
  things that could trigger fights with payers before
• 4:45
  the fight has even started.
• 4:47
  So, Dan Greenleaf, he's CEO at Duly Healthcare
• 4:50
  here in the Chicagoland area. Scott, I know
• 4:52
  you know Dan. He he even told me
• 4:55
  in pretty black and white terms that he
• 4:57
  anticipates his revenue cycle management spend going down
• 5:00
  by half.
• 5:02
  I don't think that's in 2026 alone, but
• 5:04
  it'll start to move in that direction.
• 5:06
  He said that the the tools are gonna
• 5:07
  be so sophisticated. There will be more alignment
• 5:09
  that this does not need a prior authorization.
• 5:12
  This,
• 5:13
  episode of care, this treatment should not be
• 5:15
  denied. And that
• 5:17
  AI's ability to kind of eliminate some of
• 5:20
  those payer provider fights that have so for
• 5:23
  so long defined
• 5:24
  the revenue cycle space will really be,
• 5:27
  a line in the sand and and a
• 5:29
  big sea change. So we'll keep an eye
• 5:31
  on this 2026.
• 5:33
  We're seeing innovative things with AI between providers
• 5:35
  and payers.
• 5:36
  You might see denials begin to shrink, not
• 5:38
  go away completely, and not because payers and
• 5:41
  providers are you know, suddenly agree,
• 5:43
  but but just because they may have more
• 5:45
  fewer administrative reasons to disagree in the first
• 5:47
  place.
• 5:48
  Well, yeah, that's fascinating because we we do
• 5:50
  hear sort of on the other side of
• 5:53
  that, just this increase in this arms race
• 5:56
  between payers and providers over so many of
• 5:59
  these issues.
• 6:00
  So it'll really be interesting to see how
• 6:02
  it actually sorts
• 6:04
  out and whether at some point is machines
• 6:06
  talking to machines and resolving more and more
• 6:08
  of it,
• 6:09
  versus one machine saying no to another machine
• 6:12
  and not resolving more and more of it.
• 6:13
  I guess we'll we'll we'll see how that
• 6:15
  plays out over the next couple years to
• 6:17
  see if,
• 6:18
  hopefully, Dan's prediction is more right than wrong
• 6:21
  that it gets closer and closer versus creating
• 6:24
  more of a burden for
• 6:25
  systems. Because I I I feel like right
• 6:27
  now, the health system side, they feel like
• 6:29
  they're trying to keep up on the arms
• 6:30
  race to
• 6:32
  cooperate or or compete with the payer denials
• 6:36
  and authorizations.
• 6:37
  And maybe we can get to a world
• 6:39
  where they actually talk to each other in
• 6:41
  a more positive way
• 6:43
  and make things easier to resolve. It it
• 6:45
  fascinating to see if that will be the
• 6:47
  case.
• 6:48
  Mhmm. And and, Scott, I'm actually interested in
• 6:50
  how what we'll see what we'll start to
• 6:51
  see from some of the the integrated systems
• 6:54
  that have both, like, you know, the the
• 6:56
  pay viders, if you will.
• 6:58
  Those that have both health plans and then
• 7:00
  the the provider arms, I think they could
• 7:02
  be a real hotbed for for how that
• 7:03
  plays out.
• 7:05
  The other trend I wanted to point to
• 7:07
  is just
• 7:08
  how much clinical decision support will evolve over
• 7:11
  this next year
• 7:12
  from AI. So for a long time, like,
• 7:14
  CVS has been seen by what some
• 7:17
  experts described as, like, a glamorized search. It
• 7:19
  was just like a really high functioning search
• 7:21
  engine,
• 7:22
  that was part of the EHR, that was
• 7:24
  part of the patient's chart. But now I
• 7:25
  think there's a lot more anticipation
• 7:27
  for
• 7:28
  AI to almost be a partner versus the
• 7:31
  tool for the clinician. So start to identify
• 7:34
  different historical trends, identify medication errors,
• 7:37
  tee up orders, support diagnostic reasoning in a
• 7:40
  much more meaningful and robust way. So there's,
• 7:43
  seems to be a lot of excitement about
• 7:44
  this, Scott. I think, the leap from, you
• 7:46
  know, being a really good search engine to
• 7:48
  then now be more of, like, a reasoning
• 7:50
  partner. It could be one of the the
• 7:52
  bigger shifts that
• 7:54
  end users and and the clinical world start
• 7:56
  to see from AI in in the year
• 7:58
  ahead. And it seems to be there's great
• 8:00
  appetite for it too.
• 8:02
  You know, I think that's right on. And
• 8:03
  we hear so much about that, this movement
• 8:05
  towards, you know,
• 8:08
  we've heard for years somebody the use cases
• 8:10
  in the
• 8:11
  revenue cycle, the administrative side. And then, obviously,
• 8:14
  this past year or two,
• 8:16
  we've had such an evolution of the use
• 8:18
  of it in the ambient listening stage to
• 8:21
  try and reduce the amount of documentation that
• 8:24
  doctors and nurses and clinicians have to do.
• 8:26
  And now you're seeing this next evolution where
• 8:29
  you're starting to see real use cases on
• 8:31
  the diagnostic side and, as you say, sort
• 8:33
  of clinical decision support systems.
• 8:36
  And and I I think that's exactly right,
• 8:38
  and the hope is that, again,
• 8:41
  these are able to handle a lot of
• 8:44
  sort of,
• 8:45
  to to almost triage things. So physicians and
• 8:47
  clinicians could focus on the more
• 8:49
  complicated and the more challenging and and and
• 8:51
  the bigger problems
• 8:53
  and and let the systems triage,
• 8:56
  some of the the the the issues that
• 8:58
  they can.
• 8:59
  Mhmm. Mhmm.
• 9:01
  And finally, just the last one I would
• 9:03
  point to, and this is what it probably
• 9:04
  excites me the the most, in in conversations
• 9:07
  with leaders is, you know, there's all these
• 9:09
  internal ways AI will improve operations and and
• 9:12
  could improve finances, retention,
• 9:15
  employee and patient experience. I think the patient
• 9:17
  experience piece is something that probably doesn't come
• 9:19
  up enough. And then and not the patients
• 9:21
  will be doing homework about how systems are
• 9:23
  using AI or what solutions or when and
• 9:26
  how their pilots went. You know, that's unrealistic
• 9:28
  to anticipate that level of engagement from patients.
• 9:31
  But
• 9:32
  will there be more noticeable
• 9:34
  improvements to their experience
• 9:36
  that start to really add up and inform
• 9:38
  patient choice? So if you've got a patient
• 9:40
  who's in a market and they've got a
• 9:42
  couple different options for where to receive care
• 9:44
  or a procedure or or where to go,
• 9:46
  will AI start to work as intended
• 9:50
  and give them time back, you know, fewer
• 9:52
  headaches, less complications in scheduling, all the things
• 9:55
  that for so many years we've gotten really
• 9:58
  accustomed to being complicated, slow moving, and tedious.
• 10:01
  We'll start to make a better experience for
• 10:03
  them overall where it'll begin to build loyalty,
• 10:06
  or they'll start to see systems
• 10:08
  maybe in different ways. Those that are investing
• 10:10
  in a better experience for them, thanks to
• 10:13
  AI, or those that maybe are just continuing
• 10:15
  on as business as usual.
• 10:17
  So, Scott, I think that's something that it
• 10:18
  it raises a lot of questions for me.
• 10:20
  I think this is something that for years
• 10:22
  and years, we've been talking about patient experience.
• 10:24
  And it seems like if AI can start
• 10:25
  to really tip the scales towards some meaningful
• 10:28
  improvements for patients,
• 10:30
  it could be, exciting to see how that
• 10:31
  turns into more of a a movement versus
• 10:33
  just like a, you
• 10:35
  know, helpful side effect of this form of
• 10:37
  technology.
• 10:39
  Yeah. That's that's simply,
• 10:41
  amazing, literally. And what's so interesting is even
• 10:43
  about even with the patient experience, both systems,
• 10:45
  what they'll do, but also the chat g
• 10:47
  p t. As you go to chat g
• 10:49
  p t and say, which system should I
• 10:51
  go to for this illness or this challenge
• 10:53
  or this problem? Who who's the who's who's
• 10:55
  the best for it? It's fascinating. Literally fascinating.
• 10:59
  Mhmm. Yeah. It's it's I mean, do you
• 11:01
  do that, by the way? I I haven't
• 11:03
  done that. Have you have you tried that,
• 11:04
  asking you that? I've not done that, but
• 11:07
  I'm using more and more of it for
• 11:09
  just all kinds of things. And I'm finding
• 11:11
  it to be crazily helpful in so many
• 11:13
  ways. You're just kinda constantly using it in
• 11:15
  different ways almost daily now. Yeah. And Yeah.
• 11:17
  And I could see where people start to
• 11:19
  use it for, if not diagnosis,
• 11:21
  or to to decide, okay, who should I
• 11:23
  go to for this or that? Like, you
• 11:25
  used to go to Google for it. Right.
• 11:27
  No. I think it's a great it's it
• 11:29
  you you just help me appreciate it. You
• 11:31
  go further it goes further upstream than just
• 11:33
  what they experience in scheduling. Right? It's like
• 11:35
  even just how it'll be used to perform
• 11:37
  where they try to schedule. So I think
• 11:38
  it's a really great point, Scott.
• 11:40
  No. A 100%.
• 11:42
  Molly, I wanna thank you as always for
• 11:44
  joining us today on the Becker's Healthcare podcast.
• 11:46
  Always fantastic.
• 11:48
  Thank you so much for joining us. Thank
• 11:49
  you so much for having me.