Latest / Becker's Healthcare Behavioral Health / AI and the Future of Behavioral Health, and How to Enhance Care Without Replacing Clinicians

Transcript

AI and the Future of Behavioral Health, and How to Enhance Care Without Replacing Clinicians · Becker's Healthcare Behavioral Health

• 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.
• 0:06
  - It's great to have you. Today, we're talking
• 0:07
  - about AI and the future of behavioral health
• 0:10
  - and how to enhance care
• 0:12
  - without replacing
• 0:14
  - clinicians. And joining me for today's discussion, so
• 0:16
  - excited to have him, Javier Ferreira, VP market
• 0:18
  - segment leader at NextGen. Javier, it's so great
• 0:21
  - to have you again. Welcome. Yeah. It's good
• 0:23
  - to be here. Thank you for having me.
• 0:25
  - I wanna kick us off with a statistic
• 0:29
  - while I was doing research for this podcast,
• 0:31
  - and it was pretty staggering, and that's why
• 0:33
  - I wanted to include it.
• 0:35
  - As of December
• 0:36
  - 2025,
• 0:38
  - forty
• 0:38
  - percent. Again, I'll say that again. Forty percent
• 0:42
  - of The US population lives in a federally
• 0:45
  - designated mental health professional shortage area,
• 0:48
  - and we also know from that research that
• 0:50
  - gap is projected to widen
• 0:53
  - through 2038.
• 0:54
  - So this is a decade ahead. We know
• 0:56
  - this is happening.
• 0:58
  - And we also know that these organizations
• 1:00
  - that are trying to close this gap
• 1:02
  - are experiencing tremendous
• 1:05
  - stress from workforce shortages to,
• 1:08
  - again,
• 1:08
  - increasing consumer expectations
• 1:11
  - for personalized care. There's so much happening.
• 1:14
  - From your perspective,
• 1:16
  - where is AI making the biggest difference today,
• 1:20
  - and why
• 1:22
  - is 2027
• 1:23
  - why is this moment now so important for
• 1:26
  - this?
• 1:28
  - Yeah. It's a the scary stat that you
• 1:30
  - just gave in in if you look at
• 1:31
  - the trajectory there, right, even
• 1:34
  - a larger gap projected to widen through through
• 1:38
  - that 2038
• 1:39
  - time period in that study.
• 1:41
  - I think the biggest difference AI can make
• 1:43
  - today
• 1:45
  - is helping
• 1:46
  - behavioral health organizations
• 1:48
  - expand capacity
• 1:50
  - without an already stretched workforce to simply do
• 1:53
  - more.
• 1:55
  - We have an increasing demand, no doubt, in
• 1:57
  - behavioral health and with the stat you just
• 1:59
  - shared and my experience working with behavioral health
• 2:03
  - organizations,
• 2:04
  - a real workforce shortage.
• 2:07
  - And so clinicians who are carrying a significant
• 2:11
  - amount of already administrative burden and compliance, right,
• 2:14
  - these are these are some scary stats to
• 2:16
  - think through.
• 2:17
  - When we think about augmented intelligence or AI,
• 2:20
  - you know, people use the artificial intelligence term
• 2:22
  - like I like to say augmented intelligence, and
• 2:25
  - I think that's a really important distinction.
• 2:27
  - It's important to note that AI doesn't solve
• 2:30
  - the workforce shortage by itself,
• 2:33
  - but it can help us use
• 2:35
  - the technology.
• 2:36
  - It can ultimately have an impact on the
• 2:38
  - work workforce
• 2:40
  - where it can be leveraged most effectively.
• 2:43
  - And so the the most immediate opportunity is
• 2:46
  - reducing the friction surrounding care.
• 2:49
  - That includes things like supporting documentation,
• 2:52
  - bringing relevant information forward before
• 2:56
  - and during an encounter.
• 2:58
  - And it ultimately can help clinicians. Right? I
• 3:01
  - think that's an important word, help clinicians
• 3:04
  - identify gaps that require attention.
• 3:06
  - It can help organizations
• 3:08
  - understand where patients may be falling
• 3:11
  - short or falling out of care. I think
• 3:13
  - what makes this moment particularly
• 3:16
  - important in in AI
• 3:18
  - is that we're moving from this theoretical
• 3:21
  - conversation to a more practical one.
• 3:23
  - So organizations are are are no longer asking
• 3:27
  - only what could AI eventually do.
• 3:30
  - They're
• 3:31
  - asking
• 3:33
  - where can it create measurable value today? And
• 3:35
  - I think that's an important question to be
• 3:37
  - asking.
• 3:38
  - And for me,
• 3:40
  - the right measure is not how much a
• 3:42
  - output an a an organization
• 3:44
  - can deploy AI.
• 3:46
  - I think it's whether that technology gives clinicians
• 3:49
  - more time to care.
• 3:51
  - It can improve access, how it can reduce
• 3:55
  - avoidable burden,
• 3:56
  - and then help an organization achieve better outcomes,
• 3:59
  - which I think should be the target and
• 4:01
  - goal.
• 4:02
  - And so
• 4:03
  - behavioral health or behavioral health organizations and in
• 4:06
  - behavioral health in particular,
• 4:09
  - we are especially dependent on context, patient context.
• 4:13
  - And a person's
• 4:15
  - needs cannot be understood from
• 4:17
  - one diagnosis,
• 4:19
  - one note, or one encounter.
• 4:21
  - And I think the opportunity
• 4:23
  - is to use augmented intelligence or AI to
• 4:26
  - connect information
• 4:28
  - across the care journey, right, across an episode
• 4:31
  - of care
• 4:32
  - and help the clinicians see more of more
• 4:34
  - complete picture
• 4:35
  - while keeping the clinician firmly in control.
• 4:38
  - And so for me, the value of AI
• 4:41
  - is not that it replaces capacity.
• 4:44
  - It is that it helps us recover capacity
• 4:46
  - for care, And I think that's an important
• 4:48
  - distinction.
• 4:50
  - I wanna double click on that a little
• 4:51
  - bit, and I wanna just touch on the
• 4:53
  - augmenting piece too because that's very important. And
• 4:55
  - you just touched on it as well. We
• 4:57
  - know that, again,
• 4:58
  - not that clinical care is in in itself
• 5:01
  - a very personal matter. Behavioral health today
• 5:04
  - is one of those pieces that's just deeply
• 5:07
  - personal to a lot of folks, and AI
• 5:10
  - just doesn't for a lot of people, it
• 5:12
  - doesn't quite fit into that conversation. Right? Because
• 5:15
  - they're Mhmm. They feel like it creates this
• 5:17
  - distance. Right? There is a computer. There's often
• 5:19
  - this misunderstanding, like, this computer, this robot that's
• 5:21
  - talking to me instead of the clinician.
• 5:23
  - Yep. Where do you see AI truly augmenting
• 5:26
  - all of this? You've touched on it a
• 5:28
  - little bit, but I want you to to
• 5:29
  - focus on that human piece, the human connection.
• 5:32
  - How can AI help
• 5:34
  - there and truly create
• 5:36
  - that center of care from the human piece?
• 5:40
  - Yeah. I think in behavioral health, the human
• 5:42
  - connection I I shouldn't say I think. I
• 5:44
  - know.
• 5:45
  - Right? The human connection
• 5:47
  - is not an accessory to care. It's central
• 5:50
  - to care. And so trust, listening, clinical judgment,
• 5:53
  - and the relationship
• 5:54
  - between the clinician
• 5:56
  - and the the person that they're taking care
• 5:59
  - of, those certain things
• 6:01
  - should not and cannot be augmented.
• 6:04
  - I think the appropriate role for augmented intelligence
• 6:07
  - or AI is to support that relationship,
• 6:10
  - not to insert itself between the clinician and
• 6:13
  - the patient.
• 6:15
  - An example, straightforward example, is documentation.
• 6:18
  - So the clinician can spend less time looking
• 6:20
  - at the screen, recalling information
• 6:23
  - after the visit or completing things that are
• 6:25
  - repetitive or repetitive administrative steps.
• 6:29
  - That creates more opportunity to be fully present
• 6:31
  - in that conversation.
• 6:33
  - And so AI can help organize information
• 6:36
  - that may spread across the record. It can
• 6:38
  - surface relevant history, changes over time, potential gaps,
• 6:42
  - or even issues that deserve
• 6:44
  - a a much closer look. But searching
• 6:47
  - servicing information is really different than making the
• 6:50
  - clinical decision.
• 6:52
  - The clinician still has to apply judgment, understand
• 6:55
  - the context of the patient,
• 6:57
  - and ultimately determine what is appropriate for that
• 6:59
  - person.
• 7:01
  - I'd like to describe
• 7:02
  - and this is really important. The model is
• 7:05
  - human led and AI supported.
• 7:08
  - The technology can listen,
• 7:10
  - organize,
• 7:12
  - and then
• 7:13
  - ultimately prompt the clinician.
• 7:16
  - The clinician still has a responsibility
• 7:18
  - to interpret,
• 7:19
  - validate,
• 7:20
  - decide, and connect the dots.
• 7:23
  - If you kind of look at the way
• 7:24
  - that,
• 7:26
  - you know, kind of the old way of
• 7:27
  - an electronic account record
• 7:29
  - versus a a a more system of context
• 7:33
  - AI enabled platform within an electronic
• 7:35
  - account record, a legacy system is like a
• 7:37
  - digital filing cabinet. Right? Clinicians have to dig
• 7:40
  - through the records to surface information.
• 7:43
  - Right? They have to read it, interpret it.
• 7:46
  - Whereas
• 7:47
  - a new EHR that's AI enabled, right, helping
• 7:50
  - the clinician,
• 7:52
  - it's surfacing
• 7:54
  - relevant information
• 7:55
  - at the right time at the point of
• 7:57
  - care, and that's an important distinction.
• 8:01
  - And that distinction is important because behavioral health
• 8:04
  - care is really highly personal.
• 8:07
  - So two individuals with similar if you take
• 8:09
  - a person, for example, two individuals with similar
• 8:12
  - clinical information,
• 8:13
  - they may have different circumstances.
• 8:16
  - They might have different treatment goals,
• 8:18
  - support system,
• 8:20
  - and you think about social determinants of health
• 8:22
  - and barriers.
• 8:24
  - AI
• 8:25
  - augmented intelligence can provide another layer of insight,
• 8:28
  - but ultimately cannot replace the clinician's understanding of
• 8:31
  - the person that's sitting right in front of
• 8:34
  - And so if AI is implemented well, and
• 8:36
  - I do lots of presentations on this,
• 8:38
  - patients should ultimately experience more, not less
• 8:42
  - human intention. And I think that's really important
• 8:45
  - in this conversation.
• 8:47
  - And so AI can bring information forward,
• 8:50
  - but the clinician brings the meaning to that
• 8:52
  - information and ultimately better serve the the patient
• 8:55
  - in front of them.
• 8:57
  - One of the important steps in this that
• 8:59
  - you started the conversation is the fact that
• 9:02
  - we are moving from pilots to
• 9:05
  - integrated systems. Right? A lot of this is
• 9:07
  - happening across health care.
• 9:09
  - We're using it. It's part of the job
• 9:12
  - in in 2026.
• 9:15
  - You've mentioned the the insights piece to being
• 9:17
  - able to surface more insights, surface information.
• 9:21
  - That's what AI does really well, and that's
• 9:23
  - why it's being used so widely. When we're
• 9:25
  - looking at behavioral health, though,
• 9:28
  - what are you seeing there in terms of
• 9:30
  - how AI helps
• 9:33
  - understand
• 9:34
  - certain risks earlier, for example? And and also
• 9:38
  - then deliver more some some more of those
• 9:40
  - personalized interventions
• 9:42
  - that you've mentioned. How does AI do that?
• 9:45
  - Yeah. I I think this is a huge
• 9:47
  - opportunity. So one of the greatest opportunities
• 9:50
  - in behavioral health with augmented intelligence is moving
• 9:53
  - from
• 9:54
  - reactive care
• 9:56
  - towards earlier more informed action, and I think
• 9:59
  - that's what you're talking about here.
• 10:01
  - Behavioral health organizations hold a tremendous amount of
• 10:04
  - information,
• 10:05
  - but it often we it often exists between
• 10:09
  - multiple encounters, multiple nodes, multiple assessments,
• 10:13
  - treatment plans over time,
• 10:15
  - and then operational workflow.
• 10:18
  - And the challenge is turning that information into
• 10:20
  - something actionable
• 10:22
  - at the right time.
• 10:24
  - And what AI can do is AI can
• 10:26
  - ultimately help identify changes or patterns
• 10:30
  - that may deserve more attention by the clinician.
• 10:33
  - This might include changes to patient engagement strategy,
• 10:37
  - missed appointments,
• 10:38
  - gaps in follow-up,
• 10:39
  - shifts in documentation needs, or other indicators that
• 10:42
  - a person's care journey may be moving off
• 10:44
  - track.
• 10:45
  - And so the important point here is that
• 10:47
  - these insights should guide outreach and clinical consideration,
• 10:51
  - but they should not label the person or
• 10:53
  - make
• 10:55
  - automatic determination of care.
• 10:57
  - Personalization
• 10:58
  - also means understanding the whole person, which is
• 11:01
  - a a huge topic in behavioral health. Right?
• 11:03
  - We've made a shift towards integrated care and
• 11:06
  - full person care.
• 11:08
  - Behavioral health outcomes can be influenced by physical
• 11:11
  - health, social needs, access barriers, family circumstances,
• 11:15
  - and then ultimately the person's individual phone calls
• 11:18
  - that are laid out in a treatment plan.
• 11:21
  - So when technology helps bring that broader context
• 11:24
  - together,
• 11:25
  - care teams can actually have a more informed
• 11:28
  - conversation and intervene
• 11:30
  - in a way that
• 11:32
  - relevant that's really relevant to the person that's
• 11:34
  - sitting in front of you. And if you
• 11:36
  - take that kind of a level up at
• 11:38
  - an organizational
• 11:39
  - level,
• 11:40
  - those insights
• 11:41
  - can help leaders understand where populations are experiencing
• 11:45
  - gaps,
• 11:46
  - who's not engaging, where transitions are breaking down,
• 11:49
  - and where and which interventions.
• 11:52
  - So it seems like med management or medication,
• 11:55
  - you know, administration,
• 11:57
  - where those interventions
• 11:59
  - appear to be having
• 12:00
  - the most significant impact.
• 12:02
  - So, ultimately, the the value is not the
• 12:05
  - prediction itself.
• 12:07
  - The value is in the action that follows.
• 12:10
  - So whether that is an earlier outreach, a
• 12:13
  - change in the treatment plan,
• 12:15
  - or additional capacity or resources to that person.
• 12:19
  - Also or simply just being more the more
• 12:22
  - informed conversation.
• 12:23
  - And so an insight only has value if
• 12:26
  - it reaches the right person at the right
• 12:28
  - time
• 12:29
  - and leads to the right action. And I
• 12:31
  - think that's what we're really focused on in
• 12:34
  - some of these AI use cases is
• 12:37
  - is exactly just that. It's making sure we're
• 12:39
  - surfacing information and insights
• 12:42
  - at the right time for the right person
• 12:44
  - and then ultimately leading to the right action.
• 12:47
  - This is all
• 12:48
  - interconnected.
• 12:49
  - You've described it perfectly just now in your
• 12:52
  - answer. It it's all it all has it
• 12:54
  - it all connects together. We haven't touched on
• 12:57
  - the burnout piece in this at all, and
• 13:00
  - I do wanna do that because, again, especially
• 13:02
  - in behavioral health,
• 13:03
  - it's it's large. It looms large. It's all
• 13:06
  - of health care, but especially in behavioral health.
• 13:08
  - When we're thinking about burnout, when we're thinking
• 13:10
  - about clinicians, where do you see the greatest
• 13:12
  - opportunities there for AI to help? Right? We
• 13:16
  - think about documentation,
• 13:17
  - but, also, you've highlighted other operational tasks there.
• 13:20
  - And, again, how does that then when we're
• 13:23
  - helping the clinician ultimately translate into better experiences
• 13:28
  - for patients but also for providers?
• 13:31
  - Well, documentation,
• 13:32
  - ambiance,
• 13:33
  - you know, ambiance documentation is obviously a starting
• 13:36
  - point because it's been one of the most
• 13:38
  - visible sources of burden for the clinician.
• 13:41
  - But the opportunity is so much broader than
• 13:43
  - simply producing a note.
• 13:46
  - Augmented intelligence
• 13:47
  - AI, you hear me say that word over
• 13:49
  - and over,
• 13:50
  - can ultimately help capture and organize relevant information
• 13:54
  - from the encounter, create an initial documentation
• 13:58
  - draft.
• 13:59
  - But I think reducing some of the repetitive
• 14:02
  - administrative
• 14:04
  - some of the tasks
• 14:05
  - that occurs before and after the visit are
• 14:08
  - also really great use cases in in AI.
• 14:12
  - The clinician still ultimately I'm gonna circle back
• 14:15
  - to the scoop human in the loop. The
• 14:16
  - clinician still must review and validate everything
• 14:19
  - and take responsibility
• 14:21
  - for the ultimate clinical record, but the starting
• 14:24
  - point becomes more efficient. And I think that's
• 14:26
  - where we should be having the conversation.
• 14:29
  - The larger opportunity is supporting the workflow around
• 14:32
  - the encounter. That includes helping prepare the clinician
• 14:35
  - with relevant context about the visit,
• 14:38
  - supporting documentation or ambient documentation during the visit,
• 14:42
  - and then identifying
• 14:43
  - follow-up
• 14:44
  - activities that need attention. And then what in
• 14:46
  - those
• 14:47
  - follow-up activities can we automate? Right? Is there
• 14:50
  - a referral that needs to go out the
• 14:51
  - door?
• 14:53
  - If you're a psychiatrist or psych nurse practitioner,
• 14:55
  - can we automate the processing of the medication,
• 14:59
  - the labs, some of those very cumbersome tasks
• 15:02
  - that, ultimately can be,
• 15:04
  - you know,
• 15:05
  - still have the human in the loop in
• 15:07
  - that whole process, but but automate some of
• 15:09
  - those,
• 15:11
  - administrative tasks.
• 15:12
  - And this matters because administrative burden has accumulated
• 15:16
  - cumulative
• 15:17
  - effect. So you take five minutes, say, from
• 15:19
  - one workflow,
• 15:21
  - that doesn't sound
• 15:22
  - transformational
• 15:23
  - by itself.
• 15:24
  - But when you apply that across every encounter,
• 15:27
  - every provider,
• 15:28
  - every day, it ultimately can return meaningful capacity
• 15:32
  - to the organization.
• 15:33
  - And for providers, that means
• 15:35
  - probably the best use case or the best
• 15:37
  - proof points that we've seen is after hours
• 15:39
  - documentation.
• 15:41
  - Some of our use cases we've seen, on
• 15:43
  - an average, providers in behavioral health and our
• 15:46
  - federally qualified health centers and tribal organizations, they're
• 15:48
  - spending about two hours of after hour documentation.
• 15:52
  - And in in leveraging
• 15:55
  - AI, an AI enabled platform, not just documentation,
• 15:58
  - we've seen that, you know, over a twelve
• 16:01
  - month case study, that reduced down to six
• 16:03
  - minutes, right, after hours. That's a significant
• 16:06
  - impact to providers.
• 16:08
  - I want also wanna be real careful about
• 16:10
  - the capacity
• 16:11
  - in producing more conversation, which seems to be
• 16:14
  - like a CFO favorite conversation. And I can
• 16:16
  - say that because I was a former behavioral
• 16:18
  - home CFO.
• 16:20
  - I think the conversation should start with more
• 16:22
  - happy providers,
• 16:24
  - retention strategies, because that ultimately down the road
• 16:27
  - means more capacity, means better productivity,
• 16:30
  - but that focus should really be centered around
• 16:33
  - happier providers and and better patient satisfaction.
• 16:37
  - And then if you dig into patient satisfaction,
• 16:39
  - right, so patients, it means the clinician
• 16:41
  - is more present. They're more informed.
• 16:44
  - They're better able to spend time with that
• 16:46
  - person listening rather than navigating, which we see
• 16:49
  - in the traditional
• 16:50
  - EHR. They're navigating the EHR with the person
• 16:53
  - sitting in front of them. And so we
• 16:55
  - get more to that what clinicians really love
• 16:58
  - to do, and that's being working with the
• 17:00
  - patients and looking face to face with them
• 17:02
  - and having a really good understanding to to
• 17:05
  - better enable, you know, to enable better care.
• 17:09
  - When you look at operations, operational efficiency and
• 17:12
  - patient experience
• 17:13
  - are not competing objectives here. When we remove
• 17:17
  - that unnecessary
• 17:18
  - friction from the provider,
• 17:20
  - we frequently improve the experience for the patient
• 17:23
  - as well. Right? And I think that's super
• 17:25
  - important.
• 17:26
  - The best technology, in my opinion, becomes less
• 17:29
  - visible in the encounter so the clinician can
• 17:31
  - be more present and visible to the patient.
• 17:35
  - Javier, I know you work with a lot
• 17:38
  - of organizations
• 17:39
  - across the country. I know that you've probably
• 17:42
  - seen your fair share of fail failed pilots
• 17:45
  - Oh, yeah. And your your share of failed,
• 17:48
  - initiatives here.
• 17:50
  - From, again, from your perspective, implementation,
• 17:52
  - etcetera,
• 17:54
  - what separates technologies
• 17:56
  - that really deliver value for organizations
• 18:00
  - beyond just adding another technology to the stack?
• 18:03
  - Right?
• 18:04
  - And where do you see the greatest opportunity
• 18:06
  - here in the next couple of years? Yeah.
• 18:08
  - This is probably one of my favorite
• 18:11
  - parts of the conversation.
• 18:12
  - And in fact, I'm doing this week, I'm
• 18:14
  - doing a presentation at the Florida Behavioral Health
• 18:17
  - Conference
• 18:18
  - specifically on this topic.
• 18:20
  - What we've learned and MIT put out a
• 18:22
  - study,
• 18:23
  - you know, 95%
• 18:25
  - of
• 18:26
  - of these pilots have failed.
• 18:28
  - These AI pilots have failed and and and
• 18:31
  - most of the cases of zero return on
• 18:34
  - investment.
• 18:35
  - And what we've learned is that the difference
• 18:37
  - ultimately comes down to intentional adoption.
• 18:41
  - Technology creates value when it begins
• 18:43
  - with a defined problem,
• 18:46
  - and then fits naturally into the workflow and
• 18:49
  - produces an outcome that an organization can ultimately
• 18:52
  - measure.
• 18:54
  - It adds complexity when organizations
• 18:56
  - deploy it because it's relatively new
• 18:59
  - without being clear about who will use it,
• 19:02
  - what decision it will improve,
• 19:04
  - how success will be evaluated.
• 19:07
  - So I would look at five things. First,
• 19:09
  - does it solve a meaningful clinical or operational
• 19:12
  - problem?
• 19:13
  - Second,
• 19:14
  - does it fit into the clinician's existing workflow,
• 19:17
  - or does it create another destination
• 19:20
  - or another task?
• 19:22
  - Third, is the information transparent
• 19:25
  - and reviewable so clinicians can validate it
• 19:28
  - rather than simply trust an output?
• 19:31
  - Fourth,
• 19:32
  - is that if an organization must have clear
• 19:35
  - governance
• 19:36
  - around privacy, security,
• 19:38
  - quality. Is is the technology
• 19:40
  - they're using unbiased?
• 19:43
  - Is it trustworthy?
• 19:44
  - Is there,
• 19:45
  - the right policies around accountability and most importantly,
• 19:48
  - appropriate views?
• 19:50
  - And then fifth, probably one of the most
• 19:52
  - important, can the organization
• 19:54
  - measure whether the technology
• 19:56
  - is actually improving capacity,
• 19:59
  - improving the experience, improving patient access, quality, and
• 20:02
  - outcomes? Probably the most important conversation.
• 20:05
  - So looking ahead, I believe the greatest opportunity
• 20:08
  - is connecting these capabilities across the full character.
• 20:12
  - A lot of times what we see, they're
• 20:13
  - they're very isolated
• 20:15
  - use cases or isolated
• 20:18
  - AI features, almost like widget.
• 20:21
  - But an intelligent workflow
• 20:23
  - helps
• 20:24
  - the care team understand the whole person
• 20:26
  - and then ultimately reduce the administrative
• 20:29
  - burden that we see in behavioral health, especially
• 20:31
  - within the compliance and regulations that we deal
• 20:33
  - with and recognize
• 20:35
  - in the patient's care where attention is needed
• 20:38
  - and coordinate the best action.
• 20:40
  - The winners that we see, right, the best
• 20:43
  - use cases
• 20:44
  - and and and my probably recommendation
• 20:47
  - to anybody listening, the winners will not necessarily
• 20:50
  - be organizations
• 20:51
  - that have the most AI technology or who
• 20:54
  - adopt it the most.
• 20:56
  - They will be the organizations that adopt it
• 20:58
  - the most intentionally. Right? Where's the highest friction?
• 21:02
  - Where is the best value that they can
• 21:04
  - bring to to the journey?
• 21:06
  - And then, ultimately, with clinicians involved, patients protected,
• 21:10
  - outcomes defined,
• 21:12
  - and the human relationship
• 21:14
  - is kept at the center, that's where we
• 21:16
  - see the most success. And so the goal
• 21:18
  - shouldn't be more AI. The goal should be
• 21:21
  - better care
• 21:22
  - enabled by the right technology in AI.
• 21:25
  - Avi, so great to have you. Thanks for
• 21:26
  - taking some time for us.
• 21:28
  - Yes. Thank you for having me, and, look
• 21:31
  - forward to the next one.
• 21:32
  - Absolutely. And we also want to thank our
• 21:34
  - podcast sponsor, NextGen Healthcare. You can tune in
• 21:37
  - to more podcasts from Becker's Healthcare by visiting
• 21:39
  - our podcast page at beckershospitalreview.com.