The CFO’s Role in Healthcare AI–Cost, Value and Governance [PODCAST]
The CFO’s Role in Healthcare AI–Cost, Value and Governance
In this episode, Alina Henderson, Vice President of Healthcare Solutions at Strata Decision Technology, discusses the CFO’s role in healthcare AI: cost, value, and governance.
Highlights of this episode include:
- What healthcare finance leaders should be thinking about when it comes to the true cost of implementing and maintaining these AI tools
- How CFOs can build a more complete business case for AI investments
- What questions finance and operational leaders should be asking before making an AI investment
- Intervention and evidence approach to evaluating AI
- How organizations should think about ongoing governance and monitoring
- How finance leaders can determine where to prioritize investment and where a particular technology can have the greatest impact
- What some of the biggest mistakes or challenges are that health systems make when evaluating/investing in AI
- How health systems should think about evolving people roles alongside AI
Subscribe Today!
Kelly Wisness: Hi, this is Kelly Wisness. Welcome back to the award-winning Hospital Finance Podcast, we’re pleased to welcome Alina Henderson. Alina is the Vice President of Healthcare Solutions at Strata Decision Technology, where she leads strategic engagement with healthcare finance executives across the country. She connects health system leaders facing common challenges and translates insights from these conversations into Strata’s product and commercial strategy. Alina brings more than 15 years of experience across healthcare finance and technology. Prior to Strata, she worked at R1 RCM in revenue cycle leadership roles spanning patient accounting operations, product management, and go-to-market strategy. Prior to R1 RCM, she led JPMorgan Chase’s healthcare receivables product, HealthcareLink, where she owned the product road map, end-user engagement, and pipeline growth. Alina holds a Bachelor of Science from the University of Illinois at Urbana-Champaign and an MBA from the Kellogg School of Management at Northwestern University.
In this episode, we’re discussing the CFO’s role in healthcare AI: cost, value, and governance. Welcome, and thank you for joining us, Alina.
Alina Henderson: Kelly, it’s so great to be here. Thank you for having me.
Kelly: Yeah, we’re happy to have you. Well, let’s go ahead and jump in. So as hospitals and health systems continue investing in AI, what should healthcare finance leaders be thinking about when it comes to the true cost of implementing and maintaining these tools?
Alina: Yeah, absolutely. I mean, obviously they’re looking at the actual cost of the tools, right? Licensing, subscriptions. And a lot of conversation right now happening around kind of this idea of tokenomics: how are those tools actually going to be tracked around utilization and resources? But beyond that, there’s really kind of four key areas that I think are important for CFOs to consider. They’re thinking about the true cost of implementing, and then, really importantly, maintaining these tools because it’s not just about getting it set up. First of all, it’s training and change management. This is not a one-time thing. We had an organization that really talked a lot about it at a recent event, investing in cross-functional teams and tactics that really drive stickiness across those teams. That means training folks across clinical, financial, operational areas to make sure they understand how those tools can be used in today’s environment, and then how they may change the workflow that they’re leveraged in across kind of other uses and other use cases. Training and change management is really critical because AI does not simply accelerate things we do today; it really changes a lot of the workflow.
Second is the context and semantic layer. And that’s really an example of kind of a core focus on maintenance. Think of it as an agent assistant that you’re trying to stand up for yourself through ChatGPT or Claude. You have to invest a lot of time in educating it so that it understands the context of what you do, how you talk, how you interact, the knowledge that you have. And all of that maintenance is really important. As CFOs, think about that collective picture and the true cost of implementing and maintaining, managing that semantic layer. And we’re seeing an increasingly bigger cost than most CFOs had expected, especially when it comes to understanding that healthcare data. And then the last piece is really kind of thinking about the bottleneck and preparing for the next bottleneck. When you think about AI, it accelerates a lot of things, right? But most of the things that it accelerates are not a singular step in one process. They are going to lead to other things down the line. So, if you’re starting to develop reports a lot faster, you’re starting to develop insights a lot faster, or getting to information, what is the downstream impact of that, and then where could there be potentially another bottleneck that could need additional investment, additional support?
Between that and then finally kind of the core architecture and IT, those are really the areas outside of the core licensing that I think are important for them to consider: training, context and semantic layer, understanding and preparing for the bottleneck, and then the core architecture of their system.
Kelly: Yeah. It makes a lot of sense when it comes to the true cost of implementing and maintaining those tools. Those are great explanations. And I loved what you said about the stickiness factor because that truly made a lot of sense to me. So how can CFOs build a more complete business case for AI investments, particularly when some of the value may not immediately show up as direct cost savings or revenue?
Alina: Yeah, and it’s been interesting to watch this evolution over the last few years, both as a society and then within our market. I think a couple of years ago, there was this sense that AI was just kind of the thing that was going to save all of us. ROI, value, dollars were almost going to fall out of the sky once you put AI in and accelerate things. And I think we’ve seen kind of some of that newly wet glow start to wear off. When we really think about cost savings and margin, we’ve got to think about that it can be translated through more than just hard dollars.
And I think that’s where CFOs are starting to see the business case as more than just dollars in and dollars out. In fact, I recently heard the leader of a large enterprise revenue cycle team that said, “We brought in AI tools, and we did. We were able to replace 10 people with 4 people. But guess what? The labor expense was the same, if not higher, because we’ve upskilled those folks. We now need different resources to work with those AI tools.” So, when CFOs are thinking about kind of that business case, it’s not that there’s just going to be a basic ROI calculation.
In addition to the cost we just talked about, when we think about kind of translating that value, a couple of different lenses to consider. First of all, improved team morale and job satisfaction, where, again, earlier we saw, “Oh, well, AI can maybe replace people quickly, and we can see some value there in savings.” It’s really pivoted from that to how can AI support and, with the human in the loop, be making ways for people to do their jobs easier, driving lower attrition, lower recruitment costs. That’s where you’re going to see that improvement. It’s not going to be immediate. But by way of not having to rehire and retrain folks, you’re going to see some savings.
And then accelerated financial visibility. Bringing AI into the fold, there’s a cost to the tool. But the fact that you’re able to get to data earlier, you’re able to see your prediction in the month earlier, you’re able to understand how your margin picture is going to come together over a quarter earlier in that timeframe, allows you to make moves and pull levers that you can’t do in arrears. You can’t change how something is happening after the month is closed, where volume is moving, what spend is occurring, how you’re supporting the demand that’s coming into the facility.
But doing that mid-month and understanding those options you have, can accelerate some of the margin management picture as well. I think it’s really important, though, just to define what success looks like as they think about those business cases. Because it’s not always going to be just pure ROI and dollars.
Kelly: Right. I mean, what does that success look like? That is a great question. And I really loved what you said about the AI newlywed glow. I think it is definitely gone, and I just loved that. I wrote it down because I just thought that was so clever. So, yeah, Alina, when a health system is considering a new AI use case, what questions should finance and operational leaders be asking before making the investment?
Alina: Yeah, and this is where I think I’ve seen that glow exactly wear off. There are institutions, not only CFOs, healthcare leaders across finance, who are being inundated with opportunities, with point solutions, and ways that AI can make everything better. What I’m hearing from folks that have been really ahead of this, both in areas like ambient listening, in radiology, in revenue cycle, is that there really needs to be time spent up front in getting alignment on the problem and the solution. First of all, do we agree that this is a problem that needs to be solved?
The thing that we’re going to buy this AI tool for, is this a problem that we agree needs to be solved? And do we agree that this tool is going to solve it? And then it’s really thinking about what’s the timeframe and what’s our process to check if we achieved what we planned. One of the things that healthcare is incredibly good at is accommodating and adjusting and workarounds, right? I often say we are the industry that have plugged and unplugged and replugged back in a fax machine. And so, we will continue doing– and our industry will continue doing whatever it needs to do because, at the end of the day, the people that work in hospitals and health systems are focused on their mission. They’re focusing on whatever needs to be done to deliver care. Because of that context, that mindset, there’s not always capacity, and there’s not always accountability to do lookbacks. Hey, let’s look back and decide. Did we actually do what we– did this do what we said it was going to do? Do we need to cut bait here, or do we need to actually move forward? So, making sure that as they’re considering those new AI cases, they’ve got folks that are dedicated to documenting and defining the problem, really defining what does it look like when that problem has been fixed, did we see something earlier, did we get to something faster, and then being able to measure what that means and whether you’ve achieved it or not. Really being clear about kind of who’s doing that work and the roles and responsibilities around it. It does come back a little bit to change management, but it starts at the very beginning. When you’re looking at that AI use case, let’s be very transparent and open about what we expect it to do.
Kelly: Yeah, no, those are fantastic questions to ask, Alina. And you’ve talked about an intervention and evidence approach to evaluating AI. Can you explain what that looks like in practice?
Alina: Yeah. Yeah, absolutely. It was interesting. I was at this revenue cycle and AI cycle conference last month. And there was a lot of conversation around pilots and kind of the lookback on pilots and understanding that there needs to be an operating rhythm that makes pilots successful, this process that I described of identify the problem, validate what you think is going to solve the problem, or actually document what you think is going to solve the problem, validate whether it happened, rinse and repeat. When we think about that process and when we look at where pilots have gone in the last couple of years, they’ve largely started in places where we’ve seen clinical resources at the table, whether they’re direct end users in things like ambient and radiology or whether they’re human in the loop in things like charge capture and other revenue cycle work. And when we’ve seen those clinical stakeholders at the table, they take a very empirical approach to that work, right? So, think about what I described: identify the problem, define what you think is going to solve it, and then test it. That is identify a problem to solve, define a hypothesis, and then test that hypothesis. So, I think what we’re starting to see is this kind of merging of clinical approaches of intervention and evidence-based testing with AI listening, right? They’re being asked to evaluate something. They’re thinking about it through the lens of select, implement, and evaluate. And that’s what we’re seeing work really well. So, I think it’s really exciting that we’re leveraging the proven evidence-based care that makes our clinical industry so special and applying that to kind of the business practice of these AI tools.
Kelly: Yeah, it’s an interesting way to approach it, for sure. And once an AI tool is deployed, how should organizations think about ongoing governance and monitoring? Where does the finance team fit into that process? And are there any particular indicators they should be tracking along the way?
Alina: Yeah. Yeah, absolutely. I think at its core, it’s really important to understand the contracts and understand the licensing behind the tools that are being deployed. Are they based on tokens? And so, then you really need a governance model that kind of aligns with both the use cases and the business case expectations. How are you going to be tracking user utilization? It’s not dissimilar as we think about kind of broader analytics strategies, right? Who is going to get access to the data? How is that going to be governed? And how do we make sure that both the data itself is secure, but also, really importantly, the insights and the decisions coming out of that data are secure? AI has democratized and made kind of the democratization of data a lot easier. You can get it in the hands of anybody. And that’s a great thing. But especially as we think about healthcare data – healthcare finance, administrative, operational data – there’s a lot of ways to slice it. And in the hands of folks who may not be familiar with what it’s telling them, you could get to a lot of different conclusions that might not be fully accurate. So, I think, in addition to understanding kind of the utilization and who’s getting it, it’s both from a security standpoint and a cost standpoint. And then really thinking about kind of how are your enterprise vendors bringing it in? Are there ways to bring AI in without the tokenomics piece? That’s really been our approach. We’ve seen AI become really much more integrated within daily, weekly, monthly workflows. So, we don’t see it as being something of an ad hoc tool that you kind of keep on your shelf and pull down when you need for reference.
Our approach at Strata has been to really look at non-token-based utilization and the use of our intelligence tools to be fully embedded within the way that organizations work. I heard a story from an organization in the Southeast who had a physician with access to a clinical AI tool and went in and posted some pretty esoteric question around things that he needed to understand around a pretty isolated clinical case. And it spent a long time cycling and ended up being about a $20,000 query that came back with an inconclusive answer. That’s not what anybody wants, and that’s not even what the clinician wanted. So, I think it’s putting in the boundaries to avoid things like that.
Kelly: Yeah, no, that makes a lot of sense. And I kind of really took in what you said about managing the use case and the business case expectations. So that seems to make a lot of sense here too. With so many potential AI use cases across a health system, how can finance leaders determine where to prioritize investment and where a particular technology can have the greatest impact?
Alina: Yeah, I really think that it’s about looking at– right now, at least, the most that I hear is really looking at where do your teams see the most opportunity. So, I think about organizations like Mayo Clinic, who have rolled out their Mayo Clinic platform. And they’ve been arguably, probably, leading the way in terms of understanding the value of bringing their data together and then putting AI capabilities on top of it. One of the ways that they decided where to prioritize their really significant investment in this space was by socializing it within their teams. They started internally within Mayo Clinic and said, “Where do we need– where can we get help? Where can we provide accelerated insights? Where can we make things faster? Where can we remove barriers?”
I think that healthcare often has looked at the world that way. And healthcare administration and finance certainly has, but we simply haven’t had some of the tools that are available now to triangulate data across different data sets to get deeper into insights. So, I think historically, where there’s been conversations around process improvement and optimization, the options have been kind of all vetted and attempted. And now, there’s really a new set of tools that are available. Understanding from those that use them and that will be required to adopt them is probably the fastest way to determine where to prioritize.
Sit with your clinicians and understand what’s needed. Sit with your finance, with your administrators, your operating and nursing leadership over facilities. Sit with the provider, the ambulatory space, your non-acute leaders, and then finance and operations, and you will hear, very clearly, feedback from those groups. I think the other piece is that will be a way to tell you where your organization is in the change management maturity cycle. If you’re hearing a lot of pushback from people saying, “I’m not ready. I’m not ready. Here are the challenges,” that gives you a sense of what you need to do to remove some of the blockers around AI adoption.But we have to remember that AI is not a tool like another SaaS enterprise solution that replaces an Excel spreadsheet. It’s really going to change the way that we work. And that’s really how we’re looking at it as well. It’s accelerating the activities so that some things don’t have to happen after the month ends, and it’s accelerating orchestration of some of these decisions so you don’t have to start from scratch when you’re doing things like planning, when you’re doing things like strategic kind of scenario modeling.
Kelly: Yeah, I mean it definitely is accelerating so many things within our workspace. And I really loved what you said. Where do the teams see the most opportunity? That really resonated with me because, as someone who has kind of brought AI into the marketing space, it just resonated. So, what are some of the biggest mistakes or challenges health systems may make when evaluating or investing in AI, and how can finance leaders help avoid them?
Alina: Yeah. And I think it comes back to what we were just saying. I think the risk is really when you start with the technology rather than the problem, adopting AI because it sounds awesome, and the founders are really great, and they really seem to understand you and your organization. I think healthcare leaders and administrators have always been overwhelmed with the amount of adjacent support that exists within the healthcare industry. I think that has only grown exponentially with AI solutions. And so, strengthening that filter of, is this a really cool, neat feature that we think is useful, or is this something that will actually move the needle for my organization? And those things are not always the same. They’re not always mutually exclusive either, but just continuing to look at it through the lens of team and people and understanding that. I think the other part is also just treating the implementation as the finish line. Like I said, we really look at AI adoption not only in the how can we bring it and adjust our workflow, how can we actually accelerate the things that we’re doing, but also in incremental pieces throughout the tools, ways that you can make small improvements. If I’m doing all this analysis on my budget variants, how can I quickly get to a narrative summary rather than having to dig through all the data? How can I use natural language processing to build this table versus selecting all of the fields manually? When you implement a tool, that’s really the starting point of the tool’s ability to add value. So, I think making sure that there is a cadence, that there are resources, that there’s investment in thinking about how that tool is used, refining, revisiting its use case, revisiting the processes that leverage it, that’s where you’re really going to get ROI. ROI is not going to come when you replace a step in a process. It’s when you wholesale shift the way you do something from A to Z.
Kelly: Yeah. No, I love that perspective. And thanks for sharing those challenges with us. How should health systems think about evolving people roles alongside AI? And from your perspective working with healthcare leaders, what are you hearing from teams about how AI is changing financial planning and decision-making? And what do you anticipate this evolution to look like over the next few years? I know that’s a loaded question.
Alina: Yeah. Well, and I think it’s a loaded question because this is such a fascinating time because we’re all living through it as well, right? Like you said, you’re bringing in our marketing tools. I’m building out kind of agents that can help with almost a kind of personal assistant type chief of staff work. So, in addition to trying to think about how we’re innovating the industry and support, we’re also living through this kind of in a personal time. And I think that’s really fascinating. I think we are going to certainly see the evolution. Listen, the most, I would say, kind of obvious one is the upskilling conversation that I hear CFOs having on a regular basis, which is kind of twofold. A, how do we train and make it both an opportunity, but almost a bit of a requirement to move forward in the AI adoption within our enterprise? It might seem heavy-handed, but the organizations that have seen kind of the best stickiness in adoption and momentum there have really required the use of AI tools. So, there’s an organization in the Northeast that has a really fantastic governance and training process, and they have isolated folks within that kind of group that are part of the committee that are using the tools, piloting the tools, and they are quite literally required to use them for a certain amount of time every day. That’s the way you’re going to be really thinking about driving change, right? These are different shifts, different movements in our day, different clicks, different interactions with the data, and that doesn’t come organically, especially in a place where we’re always moving so fast and trying to get to kind of deadlines. So, I think making sure that we’re thinking about giving finance teams kind of that training and support, but also letting them define what that looks like, “Hey, right now, what’s data that you get too late in your month and in your quarter? If I can get that to you earlier, what’s the next thing that you need after that? Do you have the next step or person to hand that off to who can analyze it and take action?” So, thinking about what the entire process looks like, because upskilling may not simply say, “Hey, let’s do your job differently.” It might say, “Hey, we’ve got a need for a different job down here now.” Now I’ve got a really different focus on making sure my upstream data cleanliness is there. And where I used to look at that as an ad hoc process that was largely exception-based, now I really want to focus on that data cleanliness, and that’s where I want to dedicate more resources. Or maybe it’s, I need now more higher-level business analyst thinking. And so, we’re going to start to really train people up on kind of the strategic initiatives of our organization. That piece is really critical. And I think the core of it is really helping people kind of make better decisions. So, as you’re evolving around AI, what AI is going to commoditize and has already commoditized is access to data. And the question is then, what data do you put in front of people? And what do you make it easier to do with that data? The way that we look at it as, how do we get that data into the hands of people who can make changes, and who can influence ways to affect the margin? And how can we bring that data earlier? So, it’s not just about software. It’s about understanding where people can ask questions and have data to support decision-making and where we can add intelligence to that decision-making. The crux of what we’re seeing in our industry is that the decisions at a health system level from executive down to a department manager almost always have a finance lens to it. And so, thinking about how we can accelerate that, how we can help people understand what’s to do with that data is where our focus is. And it’s not just in software, it’s in the people use of it as well.
Kelly: Well, thank you so much for those perspectives and for that awesome advice. And thank you so much, Alina, for sharing your insights with us on the CFO’s role in healthcare AI, cost, value, and governance. And if a listener wants to learn more or contact you to discuss this topic further, how best can they do that?
Alina: Yeah, absolutely. Please feel free to look me up on LinkedIn, Alina Henderson. And if you’d like to email me directly, you can email me at alina@stratadecision.com. Thank you so much, Kelly.
Kelly: Thank you, and thank you for providing that. And thank you all for joining us for this episode of The Hospital Finance Podcast. Until next time…
[music] This concludes today’s episode of The Hospital Finance Podcast. For show notes and additional resources to help you protect and enhance revenue at your hospital, visit besler.holdings/podcasts. The Hospital Finance Podcast is a production of Besler Holdings.
If you have a topic that you’d like us to discuss on The Hospital Finance Podcast or if you’d like to be a guest, drop us a line at contact@besler.holdings.






