Latest / Becker’s Healthcare Podcast / Building AI Solutions That Work for Healthcare with Nishith Khandwala, Sam Schwager, Steve Liou, & Dr. Bo Gu

Transcript

Building AI Solutions That Work for Healthcare with Nishith Khandwala, Sam Schwager, Steve Liou, & Dr. Bo Gu · Becker’s Healthcare Podcast

• 0:00
  - This is Scott Becker with a special edition
• 0:03
  - of the Becker's Healthcare and Becker Business podcast.
• 0:07
  - We're joined today by by four founders,
• 0:10
  - of health companies,
• 0:12
  - mostly digital health, but they'll tell us more
• 0:15
  - about their specific stories and what they're doing.
• 0:18
  - Just thrilled to have them with us today.
• 0:20
  - We've got,
• 0:22
  - Bo Ku, the the founder of doctor Gu,
• 0:24
  - the founder of Uify.
• 0:27
  - We have Nishkandwala,
• 0:29
  - the the founder of Bunker Hill Health,
• 0:32
  - Sam Swager,
• 0:34
  - the founder
• 0:35
  - of Superdial,
• 0:36
  - and finally with us is Steve Liu, the
• 0:38
  - founder of Clariant.
• 0:40
  - I think our audience will enjoy this tremendously,
• 0:43
  - for the brightest people that I get to
• 0:45
  - visit with. I I'm gonna ask each of
• 0:46
  - you a few different questions. We'll ask you
• 0:48
  - to tell us about your company and what
• 0:50
  - you do.
• 0:51
  - We'll also ask about where you're seeing things
• 0:54
  - work for customers
• 0:56
  - beyond sort of AI hype, what's actually working.
• 1:00
  - Then we'll focus on health system CEOs.
• 1:04
  - Where's health system CEO be focusing the resources?
• 1:07
  - Everybody's got some constraints.
• 1:10
  - What solutions and where should they be focusing
• 1:12
  - now?
• 1:13
  - And finally, we'll ask you each to give
• 1:15
  - some advice to founders. Each of you are
• 1:17
  - founders.
• 1:18
  - Each of you are are funded and and,
• 1:21
  - you know, invested in by some of
• 1:23
  - the biggest names in investing, in venture capital
• 1:27
  - investing.
• 1:27
  - And so just thrilled to have the four
• 1:29
  - of you with us today.
• 1:31
  - I'll I'll start by asking doctor Gu,
• 1:34
  - Yulify to take a second, introduce yourself,
• 1:37
  - tell us what problem you're solving, and you've
• 1:39
  - got this remarkable background
• 1:42
  - from brilliant physician to founder.
• 1:45
  - And talk about what you're doing, what you're
• 1:47
  - most focused on. And and, Beau, let me
• 1:48
  - ask you to kick us off. Yeah. Absolutely,
• 1:51
  - Scott. And, thanks for having me again. It's
• 1:53
  - always good to to be on Becker's.
• 1:56
  - So my background is unique. I'm a former
• 1:58
  - cardiothoracic
• 1:59
  - surgeon and emergency physician and a former hospital
• 2:03
  - executive.
• 2:04
  - You know, I think the way I look
• 2:05
  - at every problem
• 2:06
  - is always from a doctor's lens, which is
• 2:09
  - try to figure out what's going on as
• 2:11
  - in diagnose the problem, with the patient,
• 2:13
  - and then find out what to do and
• 2:15
  - how to treat them, therapeutically.
• 2:17
  - And that's kind of the lens that I
• 2:18
  - brought into revenue cycle, having sat in the
• 2:21
  - chair to kind of listen to all the
• 2:23
  - vendors and services and BPOs,
• 2:26
  - to pitch our hospital and health care system,
• 2:29
  - before in the past.
• 2:31
  - And, very challenging times during the pandemic,
• 2:34
  - while we were all being hailed as heroes,
• 2:36
  - we saw some of the highest, denial rates
• 2:39
  - and obviously hit the reimbursement side of physician
• 2:42
  - compensation. So,
• 2:44
  - having worked, with,
• 2:46
  - you know, our chief technology officer who's one
• 2:48
  - of the AI pioneers for foundational
• 2:50
  - architectures for,
• 2:52
  - AI and transformers,
• 2:54
  - he and I, as well as our COO,
• 2:56
  - decided to tackle this, for a true end
• 2:58
  - to end AI native RCM, agentic platform.
• 3:03
  - Thank you so much. I'll ask each of
• 3:05
  - you to introduce yourselves and give us just
• 3:06
  - a bit of the focus, then I'll come
• 3:08
  - back and start to ask you deeper questions
• 3:10
  - about what you're doing and where you're seeing
• 3:11
  - impact.
• 3:13
  - Nish Khanwala, can I ask you to take
• 3:14
  - a moment to introduce yourself and and tell
• 3:16
  - us about Bunker Hill Health?
• 3:18
  - Thanks, Scott. It's great to be here. My
• 3:20
  - name is Nish, CEO, cofounder of Bunker Hill.
• 3:23
  - I was previously
• 3:24
  - a researcher at Stanford on the API side,
• 3:27
  - building different algorithms
• 3:30
  - for use cases in health care, saw that
• 3:31
  - it was incredibly difficult to translate those ideas
• 3:33
  - into clinical practice.
• 3:35
  - So we spun out Bunker Hill, where our
• 3:37
  - goal is to work with operational and clinical
• 3:40
  - leaders to take their ideas
• 3:42
  - and build AI agents for them on our
• 3:43
  - platform.
• 3:45
  - And it all stems from the insight that
• 3:48
  - a lot of use cases are LLMs plus
• 3:50
  - x.
• 3:51
  - And as these LLMs get better,
• 3:53
  - x becomes smaller and smaller,
• 3:56
  - which really allows for there to be a
• 3:58
  - platform that can do a lot of the
• 3:59
  - clinical and operational
• 4:01
  - Excited to be here.
• 4:03
  - Thank you so much.
• 4:05
  - Steve Lew, let me ask you to take
• 4:07
  - a second to introduce yourself and tell us
• 4:09
  - just a little bit about Clariant.
• 4:11
  - Yeah. Sure thing. First of all, it's great
• 4:13
  - to be here, Scott. And,
• 4:14
  - Sam, great to share this podcast with you
• 4:16
  - too.
• 4:17
  - For those of you who don't know me,
• 4:18
  - I'm Steve Liu, founder and CEO of Clariium.
• 4:21
  - I spent the first sixteen years of my
• 4:23
  - career as a health care investor before founding
• 4:25
  - Clariium in 2020,
• 4:27
  - really in the early days of the pandemic,
• 4:29
  - which for better or worse, put a massive
• 4:31
  - spotlight on the inefficiencies,
• 4:34
  - disruptions,
• 4:35
  - and really lack of investment in health care
• 4:38
  - supply chain technology.
• 4:39
  - And really to summarize the problem that we're
• 4:42
  - solving,
• 4:43
  - as you know, health systems work tirelessly to
• 4:45
  - ensure a supplier drug shortage never impacts patient
• 4:49
  - care, but their technology and tools have always
• 4:52
  - been twenty years behind what their peers had
• 4:54
  - in, companies like Amazon and Walmart.
• 4:57
  - And Firem exists to give health care the
• 4:59
  - technology and AI advantages that other industries have
• 5:02
  - so that, really, every product SKU is visible,
• 5:05
  - every dollar is well spent, and patient care
• 5:06
  - is never put at risk.
• 5:08
  - Thank you so much. So so we've heard
• 5:10
  - so far from doctor Ku, revenue cycle end
• 5:13
  - to end,
• 5:14
  - Nish Kanwala,
• 5:16
  - in in in in working at with Bunker
• 5:18
  - Hill Health on
• 5:20
  - agentic AI and helping systems really accelerate their
• 5:23
  - use of agentic AI,
• 5:26
  - Steve Wu,
• 5:27
  - on supply chain and just doing brilliant work
• 5:30
  - there. Sam, let me tee it up for
• 5:32
  - you and and and tell us a little
• 5:33
  - bit about what Superdial is doing.
• 5:36
  - Certainly. Thank you, Scott. It's, great to join
• 5:39
  - all of you here today. Sam Schwager, cofounder
• 5:41
  - and CEO of Superdial.
• 5:42
  - We build voice AI agents for RCM teams.
• 5:46
  - And we actually got our start, as an
• 5:48
  - RCM company. We were called Superbill before Superdial.
• 5:51
  - And Superdial was actually born, you know, as
• 5:53
  - an internal automation at at Superbill. We set
• 5:56
  - up to solve our, biggest administrative problem, which
• 5:59
  - was,
• 6:00
  - the the phone calls that we had to
• 6:02
  - make to payers, to verify benefits, follow-up on
• 6:05
  - claims, and more. This this was, you know,
• 6:07
  - the biggest source of, you know, time and
• 6:09
  - and cost
• 6:10
  - expenditure in in our business. And, you know,
• 6:12
  - we we realized pretty quickly it was, you
• 6:15
  - know, generally applicable, you know, solution,
• 6:17
  - to other RCM companies and RCM teams.
• 6:20
  - So we ended up pivoting from SuperBuild to
• 6:22
  - Super Dial back in 2023,
• 6:24
  - and the rest is history. We are now
• 6:26
  - working with some of the top RCM teams
• 6:28
  - in the industry, but we've handled over 7,000,000,
• 6:31
  - phone calls.
• 6:33
  - And we've been even expanding the capabilities,
• 6:36
  - of our platform to, you know, solve other
• 6:38
  - problems, similar to the the pair phone call
• 6:41
  - space by our our Centimeters customers. So, it's
• 6:43
  - it's great to be here today and to,
• 6:45
  - you know, talk about, you know, what we're
• 6:46
  - seeing in the industry in general.
• 6:48
  - Thank you. When you talk about sort of
• 6:50
  - the the the phone calls, are the out
• 6:52
  - phone calls, inbound calls, how does that work
• 6:54
  - today, and and how does that interaction work?
• 6:57
  - I know and and is most of your
• 6:58
  - work
• 6:59
  - with other RCM companies or directly with providers
• 7:02
  - and payers? What does that look like, Sam?
• 7:05
  - Yeah. So the the the core of our
• 7:07
  - platform is,
• 7:08
  - the the outbound call capability to the payer.
• 7:11
  - So our our customers
• 7:13
  - our combination,
• 7:15
  - of the RCM teams within provider organizations and
• 7:18
  - then, yes, RCM companies,
• 7:20
  - themselves.
• 7:21
  - You know, companies like ThinkOmega,
• 7:23
  - which is one of the, you know, largest
• 7:25
  - RCM companies in the industry and, one of
• 7:28
  - our, one of our clients.
• 7:30
  - Yeah. The calls are, you know, to the
• 7:32
  - you can think of them as, you know,
• 7:33
  - calls to to gather information
• 7:35
  - from payers, that that could not be gathered,
• 7:38
  - via digital modalities.
• 7:41
  - So, you know, these are calls related to
• 7:42
  - benefits verification, prior authorization,
• 7:45
  - claim status, denial follow-up,
• 7:47
  - credentialing and enrollment.
• 7:49
  - There are calls where you you navigate a
• 7:51
  - phone tree, wait on hold, authenticate yourself to
• 7:53
  - a payer representative, potentially get transferred around to
• 7:56
  - a couple of different departments, and then, you
• 7:59
  - know, conduct a a very,
• 8:01
  - domain specific
• 8:03
  - conversation,
• 8:03
  - with the with the payer representative,
• 8:05
  - about one of the the use cases I
• 8:07
  - listed.
• 8:08
  - These calls can, you know, be up to
• 8:09
  - an hour long.
• 8:11
  - So they're they're very nuanced. And, you know,
• 8:13
  - frankly, before the advent of, you know, know,
• 8:15
  - LLMs and, you know, a lot of the
• 8:17
  - increases that we've seen in, you know, voice,
• 8:20
  - capabilities,
• 8:21
  - this just would not have been something that
• 8:22
  - was that was possible to handle. So, you
• 8:25
  - know, it's it's a very exciting time, to
• 8:27
  - see this automated working at such scale.
• 8:29
  - That's amazing. And in terms of, working with
• 8:32
  - the payers and so forth, this must take
• 8:34
  - a lot of the you know, typically, somebody
• 8:35
  - in doctor's office or revenue cycle office is
• 8:38
  - calling a payer.
• 8:39
  - In addition to needing information or trying to
• 8:41
  - get information through, they've got tremendous frustration. There's
• 8:44
  - emotion.
• 8:45
  - This must help cut to the chase some
• 8:47
  - too because you're taking the human element out
• 8:49
  - of that.
• 8:51
  - That's exactly right, Scott. When you think about
• 8:53
  - the core experience of the the conversation on
• 8:55
  - the, you know, payer representative side, Our voice
• 8:58
  - agents are, you know, trained in the revenue
• 8:59
  - cycle given we are a voice AI company
• 9:01
  - built for our Centimeters teams. Our our voice
• 9:03
  - agents are domain experts in our Centimeters.
• 9:06
  - They're very polite, friendly, and they don't spend
• 9:09
  - time, you know, looking up information. They have
• 9:11
  - it all in context so that the conversation
• 9:13
  - is faster, it's more pleasant,
• 9:15
  - and it's, you know, really expert to expert.
• 9:18
  - So the the payer rep has a, frankly,
• 9:20
  - good time or a better time talking to
• 9:22
  - our, voice agent, I think, than in a
• 9:24
  - lot of, you know, cases talking to a
• 9:25
  - person.
• 9:27
  - Simply remarkable. And and, Steve, will tell us
• 9:31
  - what you're seeing on the supply chain side?
• 9:33
  - Where is this working? I I know you're
• 9:35
  - both great
• 9:37
  - great,
• 9:39
  - investors,
• 9:40
  - I think general catalyst and some others,
• 9:43
  - great customers. And then then where is this
• 9:46
  - working so far? Where are people using your
• 9:48
  - solution, and what does that look like?
• 9:50
  - Yeah. Great question, Scott. Let me start by
• 9:53
  - giving a quick overview and recap Clarium.
• 9:55
  - So Clarium is an AI native supply chain
• 9:58
  - intelligence layer. We unify supply chain, finance, and
• 10:01
  - clinical data for health systems, allowing them to
• 10:04
  - automate and transform business processes, reduce unnecessary spend,
• 10:08
  - and prioritize patient care.
• 10:10
  - Supply chain, for those who don't know, is
• 10:12
  - the second largest expense for hospitals with over
• 10:15
  - 450,000,000,000
• 10:16
  - annually spent by US hospitals, and
• 10:19
  - we feel that approximately 60,000,000,000 of that is
• 10:22
  - what Clariem is looking to eliminate. And we
• 10:24
  - can categorize that as excess spend,
• 10:26
  - procedural waste, and unnecessary administrative
• 10:29
  - burden.
• 10:30
  - Now what I'm most excited about is Clariem's
• 10:33
  - rapid deployment of agentic AI, and we believe
• 10:36
  - that it's at its earliest stages in health
• 10:38
  - care, especially within supply chain.
• 10:40
  - Just last week, McKinsey published a report that
• 10:42
  - found
• 10:43
  - 85% of organizations are pursuing AI adoption in
• 10:46
  - health care supply chain today,
• 10:49
  - 79%
• 10:50
  - have adopted it, and 66%
• 10:52
  - are already seeing ROI from it.
• 10:54
  - And for us, the top use cases in
• 10:56
  - that survey,
• 10:57
  - for example, supplier performance, predictive demand planning, inventory
• 11:01
  - management are all areas that Clarion's agentic platform
• 11:04
  - is working across. And just to give a
• 11:07
  - few real world results,
• 11:09
  - of what we've seen at our health systems,
• 11:11
  - one at St. Luke's Health in Boise, Clarion's
• 11:14
  - AI agents have helped automate 79%
• 11:16
  - of tasks during critical supply disruptions
• 11:19
  - and have reduced their total disruption resolution time
• 11:22
  - by 71%.
• 11:23
  - At Yale New Haven Health, we identified more
• 11:26
  - than 3,000,000
• 11:27
  - in annual supply waste on preference cards
• 11:29
  - with, a little over 1,300,000
• 11:32
  - in savings already achieved from our platform. And
• 11:34
  - at Cleveland Clinic, Clarion's helped identify over 26,000,000
• 11:38
  - in net new annual savings, and we've achieved
• 11:41
  - 7,000,000 of that in less than twelve months.
• 11:43
  - And this really leverages our AI powered procedure
• 11:46
  - card optimization for system wide standardization
• 11:49
  - of supplies used in all surgeries.
• 11:52
  - Thank you. And and let me ask you
• 11:53
  - a a a follow-up question there. Now in
• 11:56
  - in
• 11:56
  - I don't wanna put you on the spot,
• 11:58
  - but Nish,
• 12:00
  - Sam, and doctor Ku, all are Stanford grads
• 12:04
  - in of one sort or another.
• 12:06
  - You're a Columbia grad.
• 12:08
  - And and and and how does Columbia compare
• 12:11
  - to Stanford today academically? And and any thoughts
• 12:14
  - there? Is there should be there should there
• 12:16
  - be an anti Stanford bias given how much
• 12:19
  - they're dominating the call?
• 12:21
  - Not at all. Stanford is one of our,
• 12:25
  - critical and and important clients and co development
• 12:28
  - partners, and so I have the deepest respect
• 12:30
  - for Stanford.
• 12:32
  - Clarion,
• 12:32
  - you know, I was a native New Yorker,
• 12:34
  - so born and raised in New York. It
• 12:36
  - was always my goal and dream to go
• 12:38
  - to Columbia, stay close to the family,
• 12:40
  - but also a great academic program. And the
• 12:43
  - other benefit it afforded me was,
• 12:45
  - it is the the finance hub of the
• 12:47
  - world. And,
• 12:48
  - as you know, Scott, I spent the first
• 12:50
  - sixteen years of my career as a health
• 12:52
  - care investor and worked at, large hedge funds
• 12:55
  - like Millennium, Citadel, point seventy two. And I
• 12:59
  - worked at JPMorgan in investment banking and health
• 13:01
  - care before that and also in health care
• 13:03
  - venture capital. And so having that home base
• 13:06
  - in New York and going to Columbia afforded
• 13:08
  - me a lot of opportunities,
• 13:10
  - from a job perspective that I may not
• 13:12
  - have gotten if I attended another school. Oh,
• 13:14
  - no. No. Simply remarkable. And I have a
• 13:16
  - daughter at Columbia grad school there now, so
• 13:19
  - I have great, great respect for Columbia.
• 13:22
  - Just, and and with the relationship to point
• 13:25
  - seventy two or back in the day, are
• 13:27
  - you able to get Mets tickets or not?
• 13:30
  - Unfortunately, I I don't have Steve Cohen on
• 13:33
  - speed dial anymore, so, probably not, but, I'll
• 13:36
  - try my best for you, Scott. No. No.
• 13:37
  - I I don't need them, but maybe we
• 13:39
  - could use super dial to get us tickets
• 13:42
  - through he could get, Steve on on on
• 13:45
  - auto dial. That's a good call.
• 13:47
  - Nish, let me ask you.
• 13:49
  - Bunker Real Health, talk a little bit about,
• 13:52
  - you know, sort of, you know, the thesis
• 13:54
  - and how it's going so far
• 13:56
  - and and and how it's rolling forward. And,
• 13:59
  - again, I mean, it's an amazingly bright group
• 14:01
  - of people we've got on the call today.
• 14:03
  - Nish, tell us a little bit about
• 14:05
  - where you're at with Bunker Hill Health and
• 14:07
  - how it's coming along.
• 14:09
  - Yeah.
• 14:10
  - So Bunker Hill is the enterprise AI agent
• 14:12
  - platform for hospitals.
• 14:15
  - We partner very closely
• 14:17
  - and become
• 14:18
  - their operating model and the governance around AI.
• 14:21
  - What I mean by this is when different
• 14:24
  - operating leaders and clinical leaders
• 14:26
  - at hospitals we partner with have ideas,
• 14:30
  - we work with them to implement those into
• 14:32
  - AI agents. And so we integrate a lot
• 14:35
  - with different EHRs, ERPs,
• 14:37
  - the Internet, and actually not only read but
• 14:39
  - also write back into them as well. And
• 14:42
  - so to to this date, we have partnered
• 14:44
  - with about two dozen health systems.
• 14:46
  - Some of the largest ones like the Cleveland
• 14:48
  - Clinic, Intermountain, Centara Health.
• 14:51
  - And at each of our health systems that
• 14:53
  - we partner with, we have more than 10
• 14:55
  - agents that are live across the enterprise. At
• 14:58
  - some health systems like UTMB, we're live with
• 15:00
  - over 20 different agents, some very clinical, some
• 15:03
  - very operational.
• 15:04
  - Like, a couple of examples to share,
• 15:07
  - we at one at Cleveland Clinic, we have
• 15:09
  - had agents identify patients who are at risk
• 15:12
  - of cardiovascular disease and get them in front
• 15:14
  - of the right care at the right time.
• 15:17
  - One of their other AI agents that are
• 15:18
  - live there,
• 15:20
  - is actually helping
• 15:21
  - the radiology department make sure that all patients
• 15:25
  - with actionable findings have have received appropriate follow-up.
• 15:28
  - We are processing over 3,000,000
• 15:30
  - radiology reports there annually
• 15:33
  - and helping over 60,000 patients get to the
• 15:35
  - right care.
• 15:37
  - At UTMB,
• 15:39
  - one of the agents that are live
• 15:41
  - is helping
• 15:42
  - different service lines ensure that patients who are
• 15:46
  - at high risk actually end up getting appointments
• 15:48
  - sooner. So we are looking at referral queues
• 15:51
  - and seeing which patients are high risk, which
• 15:53
  - patients are low risk, how do you match
• 15:55
  - the right patient to the right physician at
• 15:57
  - the right time.
• 15:58
  - And to give you a concrete example, in
• 16:01
  - the case of nephrology,
• 16:03
  - an average, sick patient would wait for about
• 16:06
  - seventy nine days for an appointment. After this
• 16:09
  - agent, that reduced down to,
• 16:11
  - less than less than thirty five days to
• 16:14
  - two x improvement.
• 16:15
  - And so we've just seen wonderful success
• 16:18
  - across clinical and operational use cases across different
• 16:20
  - health systems that we partner with.
• 16:23
  - No. Simple remark. I wanna come back to
• 16:25
  - you at you in a second and talk
• 16:26
  - about this from a health system perspective, a
• 16:28
  - CEO perspective.
• 16:29
  - In in how many different AI companies are
• 16:31
  - they gonna work with? Because I look at
• 16:34
  - over the years, a lot of systems had,
• 16:36
  - you know, five to 12 enterprise programs running,
• 16:40
  - and then, you know, it could be a
• 16:42
  - thousand to 3,000 apps or modules running of
• 16:44
  - different types, and then you see people trying
• 16:46
  - to sort of claw back on how many
• 16:48
  - different things are running. I'd love to, at
• 16:50
  - a moment, get with each of you on
• 16:52
  - some of your thoughts on a a CEO
• 16:54
  - perspective, a CIO perspective, an operating perspective,
• 16:57
  - and how many different AI companies is somebody
• 16:59
  - to work with? How much are you gonna
• 17:01
  - use them not just for a point, but,
• 17:02
  - like, for example,
• 17:04
  - Nisha's company works across a wide range of
• 17:07
  - areas with helping them develop agents.
• 17:09
  - And I wanna come back to that discussion
• 17:11
  - in a second. Before I do that,
• 17:13
  - Bo, tell us a little bit about, you
• 17:14
  - know, another Stanford Medical School graduate.
• 17:18
  - Tell us a little bit about, you know,
• 17:20
  - where you're working currently, how it's going, and
• 17:23
  - and and where things are going good. Yeah.
• 17:25
  - You know, I think every
• 17:27
  - co design question really starts with
• 17:30
  - a financial pain point and, only the outcome.
• 17:33
  - You know, I think
• 17:35
  - less and less are you really seeing point
• 17:38
  - solutions per se,
• 17:40
  - that's kind of heterogeneously,
• 17:42
  - distributed within the ecosystem,
• 17:44
  - whether it's revenue cycle or supply chain finance.
• 17:47
  - So I actually just had a whole long
• 17:49
  - two hour discussion with, one of the largest
• 17:51
  - health care system in the Northeast.
• 17:54
  - And
• 17:55
  - for them, what they really care about is
• 17:56
  - not just, you know, reducing your days in
• 17:59
  - AR, which we've done by fifty one days
• 18:01
  - and, you know, increase the appeal letter,
• 18:04
  - success,
• 18:05
  - by 90%.
• 18:07
  - But, really, if I spend a dollar on
• 18:09
  - Yubo,
• 18:10
  - would you be able to bring how many
• 18:12
  - dollars back? Right? So what is that true
• 18:14
  - ROI from,
• 18:16
  - overturn rate of denials,
• 18:18
  - whether it's,
• 18:19
  - really increasing that cash flow? And for us,
• 18:22
  - that number is five. And that's what we
• 18:24
  - have for the EagleFi yield coefficient,
• 18:26
  - for all our health care systems. And when
• 18:28
  - we do that cost analysis,
• 18:30
  - that's what we're looking at.
• 18:32
  - Secondly, from a technology
• 18:34
  - and from
• 18:35
  - a a broader vision,
• 18:36
  - they're looking for
• 18:39
  - how do you really anticipate things? How do
• 18:41
  - you predict things before they happen?
• 18:43
  - And, you know, one very, astute,
• 18:46
  - CRO was just saying that, hey. It can
• 18:48
  - be something like the minority report from that
• 18:50
  - movie where you anticipate a crime before it
• 18:52
  - happens.
• 18:53
  - So really building the predictive analytics, the revenue
• 18:56
  - integrity, and working more upstream,
• 18:59
  - where everything is interlinked within revenue cycle is
• 19:02
  - what we're building and have built, in terms
• 19:05
  - of metrics success.
• 19:07
  - Thank you. And and talk about it. I
• 19:09
  - know you got a huge vote of confidence,
• 19:12
  - from one of the most selective
• 19:14
  - early stage investors in the world.
• 19:17
  - Talk about how that impacts you, how that
• 19:20
  - impacts confidence in what you're doing,
• 19:22
  - and and and sort of what that means
• 19:24
  - for your growth,
• 19:25
  - as you continue to pursue growth and where
• 19:28
  - you're going. Yeah.
• 19:30
  - I think I'll be honest.
• 19:32
  - You know, when you have these conversations
• 19:34
  - with CEOs, CFOs, and, SVPs and VPs of
• 19:37
  - the health care system,
• 19:40
  - we we rarely talk about who our investors
• 19:42
  - are. I think there's a little bit of
• 19:43
  - personal validation. I think fundraising is a very
• 19:46
  - important piece of a founder success,
• 19:49
  - but
• 19:50
  - that doesn't really translate into why does it
• 19:53
  - matter for the CFO, why does it matter
• 19:55
  - for the CEO, and why does it matter
• 19:57
  - for the VP of RCM.
• 19:59
  - I think we're very
• 20:01
  - lean in terms of how we spend our
• 20:03
  - money.
• 20:04
  - We're very efficient.
• 20:05
  - So in terms of the capital, it's more
• 20:07
  - about scaling our growth, supporting,
• 20:10
  - our agentic,
• 20:11
  - deployments.
• 20:13
  - But I think from a personal standpoint,
• 20:15
  - you know, it's,
• 20:16
  - helping the confidence. But from
• 20:18
  - enterprise
• 20:19
  - deployment,
• 20:20
  - it still goes back to the pain point
• 20:22
  - and, financial outcome.
• 20:24
  - Thank you. And and who are you working
• 20:26
  - most closely with at systems? Is it the
• 20:29
  - CFO,
• 20:30
  - somebody in charge of revenue cycle? Who's the
• 20:33
  - core
• 20:34
  - contact point that you end up working with?
• 20:36
  - Yeah. And and and you know this better
• 20:37
  - than everyone else, Scott. You know, there's sometimes
• 20:39
  - always that segregation between finance,
• 20:42
  - as well as revenue cycle.
• 20:44
  - And in many ways, as we're kind
• 20:47
  - of relaying the success of revenue cycle, a
• 20:49
  - lot of those reports,
• 20:51
  - becomes more
• 20:52
  - understandable to the CFO and the finance team,
• 20:55
  - and they're not working against each other. But
• 20:57
  - the decision makers are the revenue cycle,
• 21:00
  - VPs, senior directors.
• 21:03
  - And a lot of times, we're working with
• 21:05
  - the billing,
• 21:06
  - the billers, the denial specialists on the ground,
• 21:09
  - from the forward deployment perspective.
• 21:11
  - But, ultimately, it's the CFO who sees that
• 21:13
  - in terms of,
• 21:15
  - budgetary,
• 21:16
  - benefits.
• 21:17
  - And and and let me ask each of
• 21:19
  - you this question, just thirty to sixty seconds
• 21:21
  - each. You had mentioned the word forward deployment.
• 21:24
  - I think that could be so critical to
• 21:26
  - make sure these things actually work as a
• 21:28
  - customer expects them to work. Give me thirty
• 21:31
  - to sixty seconds on the importance of forward
• 21:33
  - deployment and what that looks like, and then
• 21:35
  - I'm gonna ask,
• 21:35
  - Sam, Steve, and Nish the same question about
• 21:38
  - forward deployment and how that relationship looks like
• 21:40
  - with your customers, with your hospital systems.
• 21:43
  - I think the success
• 21:44
  - of forward deployment really comes down to,
• 21:47
  - are you able to speak the language?
• 21:50
  - Do you understand what it means in terms
• 21:52
  - of, like,
• 21:53
  - a denial subcategory of medical necessity? Do you
• 21:56
  - understand,
• 21:58
  - utilization management? Being able to speak the language,
• 22:01
  - being able to understand the clinical context,
• 22:05
  - being able to sit down next to them,
• 22:07
  - whether they're making a manual phone call to
• 22:10
  - the payers or whether they're trying to, come
• 22:13
  - up with an appeal letter
• 22:14
  - or trying to use a pivot table or
• 22:17
  - slice a dinosaur for Epic and understand,
• 22:20
  - literally with your eyes and ears,
• 22:22
  - of what that really means. And for deployment
• 22:24
  - is actually more listening and understanding rather than
• 22:27
  - doing.
• 22:28
  - So for me, I think that's our key
• 22:30
  - to success. And, I would say 90% of
• 22:33
  - our times, we're just listening.
• 22:35
  - And I think they understand that we have
• 22:37
  - deployed operators who have been in revenue cycle,
• 22:40
  - including myself, and that,
• 22:42
  - that conversation goes a lot smoother.
• 22:45
  - And and, Sam, take a moment on a
• 22:47
  - similar question of how important is that engagement
• 22:49
  - with the with the health system
• 22:51
  - so don't feel like you're throwing it a
• 22:52
  - solution at them, but you're working with them
• 22:54
  - to make sure it gets deployed correctly. Sam,
• 22:57
  - can you take a moment on that?
• 22:59
  - Certainly. Yeah. So this is something that we're
• 23:01
  - very, you know, proud of in our company
• 23:03
  - that's that's evolved a lot over the the
• 23:04
  - past year.
• 23:05
  - And I think Ford deployment is kind of
• 23:08
  - a term that gets used, you know, monolithically,
• 23:10
  - but, you know, means different things for different
• 23:12
  - settings. So I'd I'd say in the in
• 23:14
  - the SuperDuo context and the RCM context, what
• 23:17
  - we found works best is a a combination
• 23:18
  - of,
• 23:20
  - technical and RCM,
• 23:22
  - subject matter experts coming together,
• 23:24
  - to both identify,
• 23:26
  - you know, integration
• 23:28
  - needs and, you know, configuration requirements,
• 23:31
  - and then to, you know, build, you know,
• 23:32
  - project plans to, you know, execute,
• 23:35
  - that kind of last mile work that ends
• 23:37
  - up unlocking a a ton of value for
• 23:39
  - the customer. So, you know, we we built
• 23:41
  - a base. So back to the, you know,
• 23:43
  - core outbound,
• 23:45
  - payer calling solution. Right? We we have a
• 23:47
  - gen you know, a a voice AI agent
• 23:49
  - right now that, you know, has access to
• 23:52
  - hundreds of millions of, you know, data points,
• 23:54
  - from the, you know, millions of, previous calls
• 23:57
  - that we've that we've run. But, you know,
• 23:59
  - customers have, you know, specific preferences and requirements.
• 24:02
  - So, yes, we have a base system,
• 24:04
  - that is expert in, you know, making these
• 24:06
  - outbound calls to payers. But, also,
• 24:08
  - there there's a need for, you know, our
• 24:10
  - RCM experts to, you know, come meet virtually
• 24:12
  - or in person ideally with, you know, customers
• 24:14
  - to, you know, identify exactly the behavior, that
• 24:17
  - they're targeting. And then our, you know, technical
• 24:19
  - subject matter experts, so the kind of the
• 24:21
  - four deployed engineers, right, to to then execute,
• 24:24
  - on on those on those requirements. So I'd
• 24:26
  - I'd call that, you know, kind of config
• 24:28
  - rather than building something custom, right, you're configuring
• 24:30
  - what you already have.
• 24:32
  - But that configuration,
• 24:34
  - again, is absolutely critical,
• 24:35
  - because, you know, without it, you're trying, you
• 24:38
  - know, as as you mentioned to, you know,
• 24:39
  - kind of, you know, push a general purpose
• 24:41
  - product,
• 24:42
  - onto, you know, a customer that that has
• 24:44
  - very specific needs. And I I do think
• 24:45
  - that the overarching theme here, right, is that
• 24:48
  - we're we're doing work for these customers. Right?
• 24:50
  - We're not we're not just delivering, you know,
• 24:52
  - software. We're we're delivering, you know, outcomes and
• 24:54
  - and, you know, labor,
• 24:56
  - via these these AI agents. So in the
• 24:58
  - same way, you'd you'd want, you know, new,
• 25:00
  - employees, right, to to be trained and, you
• 25:02
  - know, ramp up on your systems and configurations.
• 25:05
  - You require the same thing from, you know,
• 25:07
  - an AI workforce.
• 25:08
  - So we've taken it very seriously, and it
• 25:10
  - it's definitely paid off,
• 25:12
  - across our across our deployments.
• 25:14
  - I mean, I mean, it's so so important
• 25:16
  - that you're actually delivering results, working so closely
• 25:18
  - with the customers. It's not just delivering a
• 25:21
  - product. It's really engagement.
• 25:23
  - Nish, I know you and I have talked
• 25:24
  - about this before, the four deployment model. Give
• 25:27
  - us a moment there, and, Nish, because you
• 25:29
  - work across a bunch of different agentic AI
• 25:32
  - situations.
• 25:33
  - Give us a moment also on how many
• 25:36
  - AI companies
• 25:37
  - when you guys talk to health systems,
• 25:39
  - how many different AI solutions are health systems
• 25:42
  - talking to as well?
• 25:44
  - Maybe not in your same space, but but
• 25:46
  - across their ecosystem.
• 25:48
  - I don't know if you have a sense
• 25:49
  - of that niche, Steve. You probably do as
• 25:51
  - well. But there's talk about forward deployment and
• 25:54
  - what that looks like engagement to customer. And
• 25:56
  - then, also, when you're dealing with customers, how
• 25:58
  - many different AI companies are they each dealing
• 25:59
  - with? Absolutely. So,
• 26:02
  - you know, we pride ourselves in the fact
• 26:04
  - that we offer the breadth of a platform
• 26:06
  - but the depth of a point solution,
• 26:08
  - and the forward deployed team
• 26:10
  - is the one responsible
• 26:12
  - and needs to be applauded for
• 26:15
  - definitely the latter half of that sentence, which
• 26:17
  - is the depth of a point solution offered
• 26:19
  - in a platform.
• 26:20
  - What I mean by this is when a
• 26:22
  - new use case, new problem is identified,
• 26:25
  - the forward deployed team builds the agents. That's
• 26:27
  - the most fun part and arguably probably the
• 26:29
  - easiest as well. Maintains the agent,
• 26:32
  - which is like you know, you're not building
• 26:34
  - a toy. You're building something that's mission critical
• 26:38
  - that is going to be deployed in a
• 26:40
  - production setting at an enterprise in an enterprise
• 26:42
  - manner.
• 26:43
  - So you really need to be responsible if
• 26:45
  - things go south at four in the morning,
• 26:47
  - if AWS goes south or something like that.
• 26:50
  - So they are responsible for building the agent,
• 26:52
  - maintaining the agent, but perhaps most importantly, the
• 26:56
  - the change management that comes with all of
• 26:58
  - these agents being deployed. And so we help
• 27:01
  - the health system with that as well. That's
• 27:02
  - what the forward deployed team,
• 27:04
  - is instrumental
• 27:06
  - and responsible for doing. And are those people
• 27:08
  - mostly
• 27:09
  - engineers?
• 27:10
  - Who who are other people in the forward
• 27:12
  - deployed team? Are they Customer people, engineering people,
• 27:15
  - a mix of people? It's gotta be really
• 27:17
  - bright people to be able to
• 27:19
  - take context, feedback, and and and make sure
• 27:21
  - you're delivering what's needed.
• 27:23
  - Yeah. This is where you have to tread
• 27:25
  - the line between being a platform play versus
• 27:28
  - a consultancy play.
• 27:30
  - Because we have architected our platform in a
• 27:32
  - way that is very composable, that is has
• 27:36
  - the mission critical components that you can repurpose
• 27:38
  - and reuse,
• 27:40
  - none of our forward deployed team
• 27:42
  - are engineers by training.
• 27:44
  - They are,
• 27:45
  - sort of x startup, x consultants, really
• 27:48
  - folks that are incredibly
• 27:50
  - nimble
• 27:51
  - that have the grid in them, and they
• 27:54
  - work on our platform and configure it for
• 27:57
  - different use cases that a health system is
• 27:59
  - interested in. And so that's how we manage
• 28:02
  - to bring scale to our work,
• 28:05
  - and and, that's why we offer the breadth
• 28:09
  - of a platform depth of a point solution.
• 28:12
  - And and that's so important because anybody's been
• 28:14
  - in the startup world knows that
• 28:16
  - most people
• 28:18
  - that do point solutions, niche solutions can do
• 28:21
  - them better than broader solutions.
• 28:23
  - But at the end of the day, systems
• 28:24
  - don't wanna work with a thousand different point
• 28:27
  - solutions, so everybody ends up morphing it to
• 28:29
  - be on the point solution at some point.
• 28:31
  - So if you could offer the breadth that
• 28:32
  - people need, but the strength and and the
• 28:35
  - agility of point solution, that's a fantastic combination.
• 28:39
  - Nish, one more question for you. When you
• 28:41
  - talk to health systems,
• 28:43
  - how many other AI companies are they working
• 28:45
  - with and dealing with
• 28:47
  - generally? And I know that's gotta be all
• 28:49
  - over the board, but what do you see
• 28:50
  - out there?
• 28:51
  - Yeah. We work with health systems that are,
• 28:53
  - you know, that are very forward with their
• 28:56
  - AI strategy. They have been early and, as
• 28:57
  - a result, have worked with a lot of
• 28:59
  - companies. We work with health systems that are
• 29:01
  - just trying to, get started on the AI
• 29:04
  - front. But, actually, we see that a lot
• 29:07
  - of use cases come from point solutions trying
• 29:10
  - to sell to health systems.
• 29:11
  - I'll give a very concrete example.
• 29:14
  - We had a point solution company approach
• 29:17
  - the chair of vascular surgery
• 29:19
  - saying, hey. We have a solution that can
• 29:21
  - help you track aneurysms over time.
• 29:24
  - It's good for the patient. It's good for
• 29:26
  - the top line for the health system.
• 29:28
  - You should really consider using it. The vascular
• 29:31
  - surgery group was very excited.
• 29:33
  - They brought it to the AI governance group,
• 29:36
  - which said, hey. We already have Carebricks as
• 29:39
  - our enterprise AI platform. Carebricks is Bunker Hill's
• 29:41
  - platform.
• 29:43
  - Why don't you, the chair of vascular surgery,
• 29:46
  - give Bunker Hill a shot? Why don't you
• 29:49
  - take the next six to eight weeks to
• 29:51
  - see if Careburex, the platform, can do what
• 29:54
  - this point solution can?
• 29:56
  - And if not, happy to entertain
• 29:58
  - onboarding that point solution.
• 30:00
  - Our forward deployed team worked with the vascular
• 30:02
  - surgery group,
• 30:04
  - implemented the aneurysm tracking solution.
• 30:07
  - And as a result, that health system was
• 30:09
  - able to,
• 30:11
  - solve that problem,
• 30:12
  - not need to onboard a new point solution,
• 30:15
  - and the vascular surgery group got what they
• 30:17
  - wanted, perhaps eight to twelve months ahead of
• 30:19
  - schedule.
• 30:21
  - One last question. You mentioned AWS.
• 30:25
  - Is everybody touching AWS,
• 30:27
  - and should we keep on investing in Amazon?
• 30:30
  - You you you don't have to answer the
• 30:31
  - second part of the question. But is everybody
• 30:33
  - touching AWS one way or another?
• 30:35
  - Perhaps. Yeah. What I meant was, I think,
• 30:38
  - last year, we had a couple of instances
• 30:40
  - where,
• 30:41
  - some of the major cloud providers went down,
• 30:43
  - for for some time.
• 30:46
  - And those are typically events where
• 30:49
  - it's not just
• 30:50
  - us as a company. It's many industries that
• 30:53
  - go down simultaneously.
• 30:55
  - And those are the types of instances where
• 30:58
  - the health system needs a partner
• 31:01
  - to know what's going on. It could and
• 31:03
  - that that's where when, you know, when you
• 31:04
  - think about build versus buy, this is the
• 31:07
  - downside of the build aspect, which is who
• 31:09
  - do you go to?
• 31:11
  - You can't go to a researcher. You can't
• 31:12
  - go to a physician builder and be like,
• 31:16
  - hey. It's gone down at four in the
• 31:17
  - morning. Can you go fix it? You really
• 31:19
  - need a partner that is responsible for the
• 31:22
  - day to day operations and maintenance of the
• 31:24
  - solution. No. That that needs a deeper team.
• 31:26
  - And my my question was not so much
• 31:27
  - about the the outage because you have that
• 31:29
  - everyone's want all these tech companies that just
• 31:31
  - happens. It was more about is everybody touching
• 31:34
  - AWS and should I invest, but we'll we'll
• 31:36
  - touch on that at a different time. The,
• 31:38
  - the Steve,
• 31:40
  - your thoughts
• 31:42
  - on CEO's biggest priorities now,
• 31:44
  - and how many different AI solutions are you
• 31:46
  - seeing when your team start to work with
• 31:48
  - systems?
• 31:49
  - You know, what does that look like?
• 31:52
  - And, also, you had mentioned that supply chain
• 31:54
  - is the second largest expense for health systems,
• 31:56
  - so, obviously, a priority or should be a
• 31:58
  - priority.
• 31:59
  - What are you seeing out there from a
• 32:00
  - CEO's perspective, and what are you also
• 32:03
  - thinking about when you think about systems and
• 32:05
  - how many different AI solutions to work?
• 32:08
  - Yeah. Great question.
• 32:09
  - When we talk to CEOs and importantly CIOs
• 32:13
  - at Leading Health Systems,
• 32:15
  - they
• 32:16
  - really stress that, they see a wave of
• 32:19
  - inevitable point solution consolidation happening over the next
• 32:22
  - three to five years. I just spoke to
• 32:25
  - the CIO of a leading system the other
• 32:27
  - day who said they have 1,800 different applications
• 32:30
  - and are looking to dramatically consolidate that. So
• 32:34
  - it's definitely,
• 32:35
  - something that's on everyone's minds.
• 32:37
  - I think that there will be the core
• 32:39
  - systems of record that everyone's invested in, that
• 32:41
  - they'll continue to use, and they will really
• 32:44
  - look to a small subset
• 32:46
  - of trusted vendor partners
• 32:49
  - that really can deploy
• 32:51
  - high value
• 32:52
  - AI solutions over the next twelve to eighteen
• 32:54
  - months that can deliver immediate ROI within the
• 32:57
  - first twelve to eighteen months of deployment.
• 33:01
  - In in in how important is it when
• 33:03
  - working with health systems
• 33:05
  - that they start to receive
• 33:07
  - ease of use
• 33:09
  - and results fairly quickly? How much patients do
• 33:12
  - they have? And when you mentioned 1,800, is
• 33:14
  - it 1,800 different AI solutions or 18 different
• 33:17
  - apps and modules of all sorts?
• 33:19
  - Applications overall, I
• 33:20
  - meant. Right. No. No. No. Absolutely. Absolutely. And
• 33:22
  - everybody's looking to consolidate some of that. But
• 33:24
  - but how important is it with your customers,
• 33:27
  - your health systems, that they start to see
• 33:30
  - results fairly quickly and ease of use? How
• 33:32
  - important is ease of use?
• 33:35
  - Both are incredibly important.
• 33:37
  - I would say that,
• 33:38
  - for us, we are dealing with our end
• 33:40
  - users being nurses, clinicians,
• 33:43
  - supply chain teams, inventory teams, incredibly busy with
• 33:46
  - their day to day operations and jobs. And
• 33:49
  - so you really need to show immediate,
• 33:52
  - improvement in terms of their workflows, in terms
• 33:54
  - of time savings, in terms of cost savings.
• 33:57
  - Otherwise, they will quickly move on from your
• 34:00
  - solution and look at the next best thing.
• 34:02
  - So that is really important.
• 34:04
  - And I would say, as Nish highlighted,
• 34:07
  - change management
• 34:08
  - is just as important as having a technology.
• 34:11
  - And for us, that's where we really focus
• 34:13
  - on two major things back to your initial
• 34:16
  - question, Scott, which is, you know, how important
• 34:18
  - is forward deployment for Clariem?
• 34:20
  - Forward deployment has been incredibly important for us
• 34:23
  - over the last five years. For every onboarding,
• 34:26
  - Clariem assigns a dedicated team of forward deployed
• 34:28
  - engineers and subject matter experts.
• 34:31
  - And as you know,
• 34:32
  - every health system IT team is incredibly bandwidth
• 34:35
  - constrained
• 34:36
  - with
• 34:37
  - a 100 different ongoing projects,
• 34:39
  - a massive EHR, ERP upgrade,
• 34:42
  - and they just don't have the time to
• 34:43
  - implement a new platform.
• 34:45
  - So we make that easy. We take on
• 34:46
  - all of that work for them in a
• 34:48
  - highly secure environment,
• 34:49
  - and that allows Clariant to deploy incredibly quickly
• 34:52
  - on average in eight to ten weeks, which
• 34:55
  - then allows us to deliver immediate hard dollar
• 34:57
  - savings within the first twelve months, which is
• 34:59
  - exactly what CIO is looking for these days.
• 35:03
  - Thank you. I'm gonna ask each of you,
• 35:05
  - one last question, and and and I'll focus
• 35:07
  - more on health systems
• 35:10
  - than on founders.
• 35:12
  - From
• 35:13
  - a health system perspective
• 35:15
  - and and, Paul, you're working with health systems
• 35:17
  - as well as practices. Correct? Not just health
• 35:19
  - systems, but practices too? Yeah. Correct. So p
• 35:22
  - platform, MSOs, large physician groups, and health care
• 35:25
  - systems.
• 35:26
  - Thank you very, very much. When you're working
• 35:29
  - with large practice groups, large PM,
• 35:32
  - PE funded practices, and and just and and
• 35:35
  - health systems,
• 35:37
  - what's the one thing that health systems should
• 35:38
  - be thinking about as they start to work
• 35:40
  - with an AI company, a a a vendor
• 35:42
  - in in the AI
• 35:45
  - world. Where should they what questions should they
• 35:46
  - ask? What's one or two things they should
• 35:48
  - be focused on and starting to work with
• 35:50
  - you? Then I'll ask each of you the
• 35:51
  - same question. You know,
• 35:53
  - Sam, Steve, Nish,
• 35:55
  - what should people thinking about as they start
• 35:58
  - to work with a vendor or particularly your
• 36:00
  - company?
• 36:01
  - Yeah.
• 36:02
  - I think the first thing that should pop
• 36:04
  - up in their mind, and, you know, same
• 36:06
  - thing in my mind, sitting at the opposite
• 36:08
  - side of the table and previously as a
• 36:10
  - health care executive,
• 36:11
  - is do I foresee a future, a long
• 36:14
  - term future,
• 36:15
  - with this vendor or,
• 36:18
  - a more, quote, unquote, partner?
• 36:21
  - Meaning that do they have
• 36:23
  - the ability to, quote, unquote, forward deploy?
• 36:26
  - Can they understand my pain points
• 36:28
  - and adapt,
• 36:31
  - to that pain point from financial to operational
• 36:33
  - or back to financial?
• 36:35
  - Do they have the technology
• 36:37
  - flexibility
• 36:38
  - to build more agents or to build a
• 36:41
  - more holistic
• 36:42
  - comprehensive platform?
• 36:44
  - And can they absorb,
• 36:46
  - honestly,
• 36:47
  - a lot of
• 36:48
  - what the previous,
• 36:50
  - BPOs were doing?
• 36:52
  - And can you identify that and show a
• 36:55
  - very measurable,
• 36:56
  - cost savings from FTEs being reduced or repurposed?
• 37:01
  - That's the first thing that's brought up to
• 37:02
  - us,
• 37:04
  - implicitly, and that's the first thing that I
• 37:05
  - was asked when I was, listening to the
• 37:07
  - vendors.
• 37:08
  - That's the first question that should pop up
• 37:10
  - to mind. And secondly, you know, can I
• 37:13
  - can I trust them? Right? Healthcare is always
• 37:15
  - built on trust, and trust is not just
• 37:17
  - a handshake. Trust is really, can you be
• 37:17
  - transparent when your metrics don't look look good?
• 37:25
  - Integration or h l seven data feeds, or,
• 37:28
  - you know, a cloud server being down? So
• 37:31
  - I think honesty is the basis of the
• 37:33
  - of the trust,
• 37:35
  - as well as adaptability and having a sustainable
• 37:37
  - future with, the CEO and health care system.
• 37:41
  - Well, no. I I absolutely love that because
• 37:43
  - the concept of anybody who's hiring somebody or
• 37:45
  - starting to contract somebody to business with them
• 37:48
  - doesn't wanna be going through 13 different vendors
• 37:50
  - over the next year. They want somebody who's
• 37:52
  - growing, who's thriving, who's who's continuing to improve
• 37:54
  - what they do so they could hopefully work
• 37:56
  - with them for a long time, not be
• 37:58
  - in and out. And and, obviously,
• 38:01
  - you want that
• 38:02
  - the the same as the company that's providing
• 38:03
  - services. You want that as well. So I
• 38:05
  - I I love that. Is this a long
• 38:07
  - term relationship? Can we sustain it? Can we
• 38:09
  - grow? Can we thrive together?
• 38:11
  - Sam, what what's the first thing a health
• 38:13
  - system should be thinking about as they start
• 38:15
  - to work with yourselves with Superdial or or
• 38:18
  - or or or another company
• 38:19
  - providing services?
• 38:21
  - Yeah. So, I mean, I I think especially
• 38:23
  - in the, you know, revenue cycle
• 38:25
  - context, RCM
• 38:27
  - is obviously a very,
• 38:28
  - you know, exciting area. There's a lot of,
• 38:31
  - you know, innovation and, you know, a lot
• 38:33
  - of a lot of promise
• 38:34
  - around, you know, the the impact that AI
• 38:36
  - agents can have in the revenue cycle. I
• 38:38
  - I do think when engaging with any, you
• 38:41
  - know, AI vendor,
• 38:42
  - a revenue cycle leader should say, are there,
• 38:45
  - you know, are there are there a handful
• 38:46
  - of other
• 38:47
  - revenue cycle leaders you have delivered for that
• 38:50
  - I could call that you've worked with for,
• 38:53
  - let's say, at least a year? Right? I
• 38:54
  - think that's a that's a good way to
• 38:56
  - weed out, you know, companies that that have
• 38:57
  - not really spent a lot of time, you
• 38:59
  - know, building the, you know, RCM expertise that
• 39:02
  - that that is required
• 39:04
  - to come into this this space. It it
• 39:06
  - it's extremely complex.
• 39:07
  - There are people with, you know, decades of
• 39:10
  - experience who are still, you know, learning a
• 39:11
  - ton,
• 39:12
  - in in the space. I I think the,
• 39:13
  - you know, the promise of agent
• 39:16
  - in RCM is is true.
• 39:18
  - Agent AI could have a massive impact, but
• 39:20
  - it's still really complicated and nuanced, and and
• 39:22
  - you need RCM expertise,
• 39:24
  - in in the company's DNA,
• 39:26
  - in order to be able to deliver. So
• 39:27
  - I'd say that's one thing. I'd say the
• 39:28
  - other thing, that that RCM leaders, you know,
• 39:30
  - should be asking is, you know, what what
• 39:32
  - is your company, you know, really the the
• 39:34
  - best in the world at? I I think
• 39:35
  - that, you know, that of course, you you
• 39:38
  - want to try to, you know, limit it,
• 39:39
  - you know, vendor bloat. It's it's good to
• 39:41
  - have, you know, vendors that can handle more
• 39:43
  - and more, but I I do see a
• 39:44
  - pattern of, you know, companies maybe that that
• 39:46
  - started out doing something, you know, very specific
• 39:48
  - and then have built around that. And I
• 39:50
  - think if a vendor can't really, you know,
• 39:51
  - answer very, very specifically,
• 39:53
  - you know, what what it is that, you
• 39:55
  - know, kind of,
• 39:56
  - is at the core of their of their
• 39:57
  - platform. May maybe it's a couple of things,
• 39:59
  - but, you know, something that that truly they
• 40:01
  - have, you know, more experience in other companies,
• 40:04
  - building more more data with which they can
• 40:06
  - train their agents.
• 40:08
  - You know, I I think if if you
• 40:09
  - don't have that, then it's it's hard to,
• 40:11
  - you know, believe that you're gonna, you know,
• 40:13
  - achieve that, you know, really fast time to
• 40:15
  - value, with the with the initial use case
• 40:17
  - that I think is is so critical because
• 40:19
  - I I do believe that, you know, starting
• 40:21
  - with, you know, a well defined, you know,
• 40:23
  - proof of concept or, you know, initial use
• 40:25
  - case or small set of use cases, really
• 40:27
  - delivering, you know, value quickly. So, like, within
• 40:29
  - the first, you know, few months, seeing the
• 40:32
  - true hard ROI and then expanding,
• 40:35
  - from there with a vendor is is better
• 40:36
  - than starting, you know, with 20 initiatives,
• 40:39
  - that, you know, can can lead to that,
• 40:40
  - you know, kind of failed deployment. We all
• 40:42
  - know that MIT, you know, article, 95% of
• 40:45
  - these, you know, enterprise AI deployments are are
• 40:48
  - failing. I think a lot of that comes
• 40:49
  - down to, you know, eyes bigger than stomach,
• 40:52
  - scenarios at the at the beginning where folks
• 40:54
  - wanna try everything all at once rather than
• 40:55
  - going deep on a couple of things and
• 40:57
  - then expanding once they're really seeing the results.
• 40:59
  - So, again, I'd I'd say two things, RCM
• 41:02
  - expertise, and then, you know, what is your
• 41:03
  - company, you know, really set itself apart, doing
• 41:06
  - at the core of the offering,
• 41:08
  - really important for RCM leaders to have in
• 41:10
  - mind as they evaluate,
• 41:12
  - vendors.
• 41:13
  - Thank you very, very much. And just a
• 41:15
  - couple quick
• 41:16
  - notes on the concept of
• 41:18
  - making sure that somebody's got references, somebody else
• 41:21
  - you could talk to that they've worked with
• 41:22
  - for a year or some period of time.
• 41:24
  - I I love that. Whenever I'm a buyer
• 41:26
  - technology, I never wanna be a first user.
• 41:29
  - I always wanted first question I always ask
• 41:31
  - is, who else is using it? How is
• 41:32
  - it going? Can I talk to them? The
• 41:34
  - other thing I'll note, I mean, both you
• 41:36
  - and Beau are in different parts of the
• 41:38
  - RCM business. We're people that I mean, people
• 41:40
  - are mostly aware of this, but health care
• 41:42
  - is about a 5,400,000,000,000
• 41:44
  - a year,
• 41:45
  - part of our economy.
• 41:47
  - And the latest that I saw was 200
• 41:48
  - to 220,000,000,000
• 41:50
  - spent somehow or another on revenue cycle billing
• 41:52
  - and collections with health care. So a huge,
• 41:55
  - huge,
• 41:56
  - you know, different areas within revenue cycle.
• 41:59
  - And and the wonder keeps on getting so
• 42:02
  - much attention because there's there's so much,
• 42:04
  - money to be saved, money to be moved,
• 42:06
  - and and so important making things work for
• 42:08
  - providers. So I I love both your inclusion
• 42:11
  - and comments. Thank you.
• 42:14
  - Nish, the first thing that a health system
• 42:15
  - should be talking to you about or or
• 42:17
  - anybody they're talking to as they start to
• 42:19
  - decide who to work with on these kinds
• 42:21
  - of solutions.
• 42:23
  - Yeah. I think, there needs to be a
• 42:25
  - sort of
• 42:26
  - we need to, as an industry, need to
• 42:27
  - be self aware that this is an incredibly
• 42:30
  - new field. And it's not just, oh, before
• 42:32
  - we didn't have LLMs, now we have LLMs.
• 42:34
  - Even as these LLMs
• 42:37
  - have progressed and gotten better, we just see
• 42:39
  - new capabilities
• 42:40
  - be unlocked. So this
• 42:42
  - the technology underlying all of this is moving
• 42:45
  - at a pace that is very fast. Now
• 42:47
  - what does that mean? It means that
• 42:49
  - you cannot be
• 42:51
  - of the of the opinion that, oh, you
• 42:53
  - can't afford to make mistakes. You can't afford
• 42:56
  - to take risks. You have to be how
• 42:58
  - do you set yourself up in a way
• 43:00
  - where the cost of iteration is very low,
• 43:03
  - where the speed of iteration is very fast,
• 43:06
  - where the cost of making mistakes is fairly
• 43:09
  - low?
• 43:10
  - And that's what I would encourage most health
• 43:12
  - system CEOs to think about in terms of
• 43:14
  - their AI strategy. Like, set it up in
• 43:17
  - a way
• 43:17
  - where you don't feel like you're getting into
• 43:21
  - a lock in period with a three to
• 43:22
  - five year contract,
• 43:24
  - you know, 6 to 7 figures annual basis
• 43:27
  - with no way to get out of that.
• 43:30
  - That makes it such that you can't afford
• 43:31
  - to be nimble. Every decision seems like a
• 43:34
  - one way door, not something that you can
• 43:36
  - never reverse.
• 43:37
  - And so the way we think about this
• 43:40
  - at Bunker Hill is we do not want
• 43:42
  - to lock in a hospital with something like
• 43:45
  - this. We offer health systems to buy AI
• 43:47
  - credits upfront
• 43:49
  - and decide
• 43:51
  - where do they want to spend those AI
• 43:52
  - credits on. So if you have an experimental
• 43:55
  - use case in mind,
• 43:56
  - you know, you can just try you doing
• 43:59
  - that use case. And if it doesn't work
• 44:00
  - out, you can stop spending AI credits on
• 44:02
  - it and just move on to a different
• 44:03
  - use case. And so that way,
• 44:06
  - rather than always feeling that it's a one
• 44:08
  - way door, you feel like the cost of
• 44:09
  - iteration is fairly low. And as a result,
• 44:11
  - you can move fast and the Right. At
• 44:14
  - which this the price of And I love
• 44:16
  - that because people don't have to think of
• 44:17
  - this. I was on the phone yesterday with
• 44:19
  - the CEO
• 44:20
  - who's doing another the latest epic transformation in
• 44:23
  - their system, and they've had great luck with
• 44:25
  - Epic. But every time they do it, it's
• 44:28
  - it's it it it's very gut wrenching, and
• 44:30
  - there's a lot that goes into it. Here,
• 44:33
  - you shouldn't think about it as though you're
• 44:35
  - you're, like, capturing a whale. You're starting with
• 44:38
  - pieces and moving from there. Nish, is that
• 44:40
  - a fair
• 44:41
  - so it's easy to get started. It doesn't
• 44:43
  - become you're not you're not planning for a
• 44:45
  - multibillion
• 44:46
  - dollar conversion. Yeah. It's easy to get started,
• 44:48
  - but it's also easy to expand and move
• 44:50
  - fast very quickly. Like, if you there are
• 44:53
  - many use cases that are just white spaces,
• 44:56
  - and you can easily see the application of
• 44:58
  - AI into those use cases and those problems.
• 45:01
  - And,
• 45:01
  - you know, to the earlier point, there are
• 45:04
  - there would be zero companies in that space
• 45:06
  - who can say, oh, we've been doing it
• 45:07
  - for three, four years. You can ring up
• 45:09
  - 30 customers, and they will tell you how
• 45:11
  - much they love you. It's just it's not
• 45:13
  - possible because it's so new. The capabilities are
• 45:16
  - so new that have been unlocked. And so
• 45:18
  - for those kinds of use cases or those
• 45:20
  - kinds of problems,
• 45:21
  - rather than say, oh, we're gonna wait three
• 45:23
  - to five years to solve that, you could
• 45:25
  - potentially get started now as long as the
• 45:27
  - cost of failing,
• 45:29
  - the cost of making mistakes is is low.
• 45:32
  - Thank you. No. I I I love that.
• 45:33
  - And and, Steve, let me wrap up with
• 45:35
  - you, Steve.
• 45:36
  - How should all systems be thinking about this?
• 45:38
  - I mean, again, as as you mentioned, just
• 45:40
  - like revenue cycle is a huge expense,
• 45:43
  - supply chain is the second largest expense in
• 45:45
  - health systems. How should all systems be thinking
• 45:47
  - about working with an AI supply chain
• 45:50
  - company like yours? What are the first things
• 45:52
  - they should be asking and thinking about?
• 45:55
  - Yeah. There are a few broader recommendations I'd
• 45:58
  - make and then a couple supply chain specific
• 46:00
  - ones. In terms of the broader recs, I
• 46:02
  - would say, one,
• 46:04
  - fix your data before you deploy AI.
• 46:06
  - Every health system is under massive pressure to
• 46:09
  - have an AI strategy,
• 46:10
  - as you know, but putting AI on top
• 46:12
  - of dirty fragmented data
• 46:14
  - only produces expensive noise faster. And the first
• 46:17
  - real move is cleansing, enriching, unifying all that
• 46:20
  - disparate data from all your systems of record
• 46:22
  - into one unified intelligence layer. And once you
• 46:25
  - have that strong foundation, then AI can be
• 46:27
  - incredibly impactful.
• 46:29
  - Two,
• 46:30
  - by outcomes, not software.
• 46:32
  - CEOs, CFOs, and CIOs
• 46:35
  - should demand that vendors show measurable ROI within
• 46:37
  - predefined window and walk away from anyone that
• 46:40
  - won't commit to those results in writing.
• 46:42
  - And three, separate the signal from the noise.
• 46:45
  - As Sam said, and you echoed Scott, CIOs
• 46:47
  - are getting pinged with thousands of AI vendors
• 46:49
  - claiming to do things they can't or simply
• 46:51
  - showing vaporware.
• 46:52
  - And, really, my advice would be to focus
• 46:54
  - on the startups that have built deep trusted
• 46:56
  - relationships with your peers and also have the
• 46:58
  - real world case studies to back up their
• 47:00
  - claims.
• 47:01
  - Now as it relates to supply chain specific
• 47:03
  - recommendations, I'd say,
• 47:05
  - one, elevate supply chain to a strategic function.
• 47:07
  - Amazon and Walmart treat their supply chain as
• 47:10
  - a competitive advantage, while many health systems still
• 47:13
  - treat it as a back office function, and
• 47:15
  - CEOs should give their supply chain leaders a
• 47:17
  - real seat at the executive table. And the
• 47:19
  - second supply chain specific recommendation would be
• 47:23
  - invest in technology now, not before the next
• 47:25
  - disruption.
• 47:26
  - Disruptions are not episodic anymore. They're the new
• 47:28
  - status quo. And if teams don't have the
• 47:30
  - right technology in place proactively,
• 47:33
  - this is gonna lead to significant financial strain
• 47:35
  - and inevitable workforce burnout.
• 47:38
  - Thank you very, very much. I I wanna
• 47:40
  - thank all of you. I know I took
• 47:41
  - more time than we'd anticipated, but just a
• 47:44
  - pleasure of visiting with the four of you.
• 47:46
  - For me, nothing better than getting to speak
• 47:48
  - with four brilliant people who are making a
• 47:50
  - huge difference in health care.
• 47:52
  - Doctor Gu, Steven Lu, Nish Kanwala, Sam Swagger,
• 47:57
  - Superdio,
• 47:59
  - Clariom,
• 48:00
  - Bunker Hill Health, UFI.
• 48:01
  - I I wanna thank all four of you
• 48:03
  - for joining us on the special episode of
• 48:05
  - the Becker's Healthcare,
• 48:06
  - and Becker Business podcast. Thank you so much
• 48:09
  - for taking the time with us today.
• 48:11
  - Thanks, Scott. Pleasure to be here. Thank you,
• 48:13
  - Scott. Thanks, Scott. Great to be here.