William Blair’s group head of healthcare technology and services analyst, Ryan Daniels, explores how artificial intelligence is revolutionizing healthcare revenue cycle management by automating financial workflows, reducing costs, and reshaping the industry’s future for providers and vendors.
Subscribe here: Spotify | Apple Podcasts | Amazon Audible
Podcast Transcript
00:21
Chris
Welcome back to William Blair Thinking Presents. Today is Tuesday, November 25th, 2025. We’re gearing up for Thanksgiving. I'm joined by Ryan Daniels, a William Blair equity research analyst specializing in healthcare technology and services. You all know him very well. He's been with us no less than, I think, three or four times. So, thank you for joining.
00:40
Ryan
Thank you, Chris. I always look forward to our quarterly chats, so thanks again for having me on the podcast today.
00:45
Chris
So, we're diving into your latest industry report. It's called “The Growing Importance of AI in the Revenue Cycle Management Marketplace.” So, this one explores how artificial intelligence is reshaping the financial backbone of health care providers. I figured you could start by giving us a high-level overview of what revenue cycle management is, you know, it's called RCM, so I'll probably be referring to it as RCM moving forward.
But, also wanted to know why AI is becoming so important in this space right now. You could argue it’s becoming important in every space right now but would love to know more about that. So, Ryan, go ahead and kick it off.
01:22
Ryan
Sure. I'd be happy to do so. First, maybe a quick programing note for the listeners. The report you mentioned is part of our ongoing health care mosaic series, where we take a deep dive into an emerging topic we think is, really, going to have broad implications across the entire health care space. This is actually now our 39th consecutive quarterly mosaic. This topic on RCM and AI is really one of the most exciting and impactful pieces we've written, in my view. So, I definitely encourage everyone to reach out for a copy of that if they don't have it already. Now, onto your question on what revenue cycle management or, as you said, RCM is, and why AI is becoming so critical for the space.
First off, RCM is really the lifeblood that is behind the entire financial payment infrastructure between patients, providers and health plans or payers. And, despite its importance in the space, it's not very well understood, given the huge amount of complexity in the system. So, I think for our listeners today, I would really break it down into three main phases across the care delivery life cycle, with those three phases being: one - pre procedure, number two- mid-cycle RCM and then three - post encounter or RCM.
So, in the first stage, or pre procedure, this is what occurs before a patient obtains care, meaning putting their insurance information into the system, attempting to identify their coverage limits. And items like copays or deductibles, maybe even collecting that financial responsibility from the patient at the point of care, and then working with payers, if necessary, to get pre-approval for services or certain procedures, which in health care is referred to as a prior authorization. And those, frankly, are just a few examples of some key RCM steps in this first stage.
Looking at stage two, that is after the care occurs. So, this could be as simple as an office visit all the way to a complex surgical procedure and subsequent hospital stay. And this is where the provider needs to document and validate all the care that was delivered, you know, including supplies, time spent, procedures done, days in the facility, etc, and then bill a payer for the charges. The charge related documents then flow over a clearinghouse to the health plan, where the health plans check the submitted claim against their payment rules and data requirements, and then they determine the reimbursement amount.
And then this leads to the third point in the RCM life cycle, or post procedure. And that's where payers either pay or deny the claim and send information and funds back to the provider. And the provider then must take this remittance and match it against the claim they submitted to see if they were fully reimbursed, what the patient balance may be, and then the patient is billed for the remainder. Also, if a claim is denied at this stage, there's entirely novel set of processes that begin where the deny claims appeal, which requires, again, a whole new set of data collection and submission back to the payer.
And, again, I've gone through a lot here, but I'm actually really oversimplifying the process. It's actually markedly more complicated and time-consuming than these examples, but hopefully that provides our listeners with a good overview of, kind of, the basic RCM life cycle. And then, to the second part of your question, this is also why AI can be so impactful here.
Just to give you one very distinct example, think about what's needed to document a certain level of charge or to get a prior authorization, to get a procedure approved from a payer. A provider needs to analyze and adhere to all these payer rules. It could be dozens of different insurance companies, dozens of different lines of business like Medicare, Medicaid, commercial managed Medicaid.
And for each of those, they need to package data and codes for that specific payer. And the provider may need to review current and past medical records, submit lab results, images, and medical justification letters simply to get a prior authorization to do a procedure. So, hospitals historically have had literally RCM teams in the hundreds of people to pull charts and do this.
But with AI, we're seeing the ability for agentic agents to do all this in minutes for them in a nearly fully automated manner, which is a huge workflow burden reduction for providers. And, again, this is one example only, but hopefully shows how impactful RCM can be. And, you know, in all, we think it could save nearly $20 billion a year in the RCM space for providers. So, clearly a huge opportunity here.
05:54
Chris
Yeah, and the report argues that AI could fundamentally change the balance between outsourced RCM vendors and internal software solutions. What are the main advantages of AI driven RCM over traditional approaches?
06:09
Ryan
Yeah, it's a great question. Today, if you look at the RCM space, it's about $100 billion in total spend. And, remarkably, 80% of that or $80 billion is to large BPO’s or business processing organizations. And, these are entities that have historically employed hundreds of workers outside of the U.S., mainly in India, to help providers with RCM activities.
So, those are the activities I talked about before, pulling charts, getting medical records combined, writing letters, etc. Now, if you think of agentic AI, the key here is actually creating software solutions that can emulate those human activities more efficiently, like the chart pulls, the report submissions, etc. with more limited human oversight. So, while the BPO’s traditionally could leverage a labor arbitrage opportunity to drive savings for RCM operations, we now believe that will be leapfrogged by AI enabled RCM solutions.
So, the next level of arbitrage will be taking human labor in general and moving it to agentic AI. And, in turn, that could cause more dollars to shift from these BPO’s to the software companies, you know, over time, in our view. And, as I mentioned a minute ago, this is a huge part of the market. so we could see dollars flow from outsourced RCM to software over the next three to five years, in our view. So, it's definitely creating a ton of interest in the space. And it could create an acceleration growth for those software vendors.
07:33
Chris
All right. Regulation and compliance are always big topics in health care. What are the main regulatory or compliance considerations for AI adoption in RCM?
07:46
Ryan
Yeah, this is always a huge topic in health care and HCIT. But, you know, maybe a little bit less so in revenue cycle management, to be honest, as you're not actually dealing with care delivery, but it's rather a financial transaction. So, we see this as an area where AI adoption can likely progress at a much more rapid pace than, say, you know, something like clinical decision support that's actually directing patient care.
Still, that said, there are needs for things like data privacy, the protection of patient information. So, in health care, there's a specific rule called HIPAA, which stands for the Health Insurance Portability and Accountability Act, and that governs patient privacy. So, anything with AI is going to have to follow those security and patient privacy standards. There also has to be a high level of transparency for providers, for patients, for payers in the AI RCM process.
And, you know, as in all use cases, there has to be checks and balances here. I would say providers can't solely rely on AI for 100% of the process. So, there does need to be a very strong talent pool to help implement and oversee the use of solutions. And, there has to be, you know, very strong data upon which to train these AI models.
And then lastly, I think a big challenge here is just a lack of AI talent. And, this is, again, every market is seeing this given how fast and explosive AI is growing. For example, I think we reference it in the report, but about 80% of respondents to a recent survey regarding the use of AI in health care said a lack of talent is the single largest barrier for implementation. So, I think those are some of the key issues there, Chris.
09:21
Chris
Okay, got it. And let's talk about use cases. The report suggests that denial management, prior authorization, and also patient engagement are the most compelling applications for AI and RCM. Why is that, and what's the size of the opportunity?
09:38
Ryan
Yeah, I think Willie Sutton said it best when he was asked why he robbed banks… because that's where the money is. And that's the same thing here. These are, you know, these are the areas where providers risk leaving a lot of money on the table if the RCM process breaks down or isn't fully functional.
So, let's take claims denial, you mentioned. It's crazy to think about this, but despite everything providers do before submitting a claim for payment, in getting prior authorization, more than 15% of every or of all claims, I should say, are denied initially by payers. So, if you think about $5 trillion in health care spending, that's about $750 billion in claims that get denied upfront.
But, with AI, you can actually more accurately predict why a claim might be at risk of denial. For example, a system can analyze one denial from a payer and then adjust all future claims submissions in order to avoid making the same mistake or submitting something wrong for the same reason. So, it can solve a huge problem.
Prior authorization is another great example you brought up. If any of the listeners on the podcast today are in the health care field, they all have to deal with this. It's one of the biggest pain points in health care today and gets a lot of press in the media. Here, a lot of times, providers see a patient and they just want to do a surgical procedure.
They want to run a diagnostic test. But, before doing so, they have to effectively submit data to the payer to prove it's medically necessary and then get the okay. So, in effect, they're asking for the okay to run a test to do a procedure. And, if they don't do that in advance, the pair will deny their claims, at the critical step.
But it does require doctors to perform a bunch of administrative tasks. They have to work through all these provider loopholes to get the okay, which isn't just a burden for them, it can really delay necessary care. So, this is a huge pain point and, as I said, AI can help do all this for the provider with very little oversight.
And then the last thing I think you mentioned was patient engagement. And for RCM, that's really, you know, taking each patient's individual insurance information, and then looking at the provider rates or pricing, whether they're in or out of network, has the patient met their deductible this year, what their co-pay might be for the service, and then providing an upfront estimate of the cost of a procedure or visit.
So, it adds a new element of transparency with patients really wanting like. And it's also key for RCM, as data shows time and time again, providers can collect almost all of the patient financial responsibility at or before the point of care. But, collections dropped precipitously if you bill that same patient in the rear. So, its really an area that's a win for patients, and it probably won't see any friction as it rolls out more broadly.
And here, you know, another interesting use case. You can even use artificial intelligence to idea if a patient might have trouble paying and then automatically offer a payment plan to help, again, improve the experience, to increase collections, reduce friction. So, a few big opportunities here.
And then, final comment, I think you asked about the size of the opportunity. So, today and we have this in our report, we peg the software sales for provider revenue cycle management at about $25 billion. But, we do think it's going to continue to grow at a pretty rapid rate, probably, high single, low double digit.
And that's just in of itself. And then if you take that BPO market, which is $80 billion, and assume some of that flows to software as these agentic agents roll out in AI, we could see it grow to around $40 billion by 2023. So, again, a $25 billion market today, we think it's $40 billion by 2030. So, we clearly think it’s a very big opportunity here, over the next five years or so.
13:26
Chris
And, what are the biggest risks or uncertainties that could slow down AI adoption and RCM?
13:30
Ryan
Great question. It's probably not dissimilar than a lot of the other industries. And, I definitely hit on some of that earlier. But, you know, I think the need to train the AI models on very good data to get accuracy and then, in turn, drive confidence and providers to use it more broadly is probably key.
And then the AI talent issue, as well. There's going to be a need for more and more talent to be hired to implement this in the RCM market, the health care market more broadly. We also get questions, and this might be a second order derivative, in the provider space. Yeah, there's a lot going on in health care today that could pressure these providers. You've got Medicaid work requirements coming up in 2027, 2028. You've got a lot of the payers pushing back on providers, because they've seen such cost pressure and higher utilization. You've got government funding cuts. And, I think the question there is, will hospitals, under these pressures, therefore pull back on any capital expenditures? And we don't think that's a risk. Again, RCM is the lifeblood of collecting cash for providers. So, we think that there's m-market pressures, it could actually enhance the demand for RCM solutions. And, again, if AI can really drive even more value, we think it could enhance that. So, we don't see that as a big risk, although I do get asked that a lot.
14:49
Chris
What should listeners watch for as indicators of real-world AI adoption RCM over the coming months, would you say?
14:56
Ryan
Well, you know, near-term we're focused on a few things. One is just product announcements. We're seeing almost daily launches of AI based revenue cycle management solutions, both from the large incumbents, from larger tech providers, and even some of the industry startups that offer point solutions. So, we're tracking that very closely.
And, then, subsequently what we want to see is use cases. So, how are these solutions being used in practice, and what are the case studies that show how the products are creating value and generate ROI for providers? Because once that's the case, you have providers using it, they're talking about the ROI, the time to value the improved cash flows. The more these surface, the more it creates, kind of, a flywheel effect, if you will, where other providers will seek out that solution and really drive a virtuous growth cycle. We outlined a lot of this in our report and, kind of, go through the product launches.
Second, I would say there's a lot of money flowing into the space, a lot of M&A activity, like private equity dollars, and that's actually helping fund these product launches and sales cycles.
And again, the funding here is creating the opportunity for bigger vendors to buy some of these point solutions. So, we follow that closely. And, you know, how these assets are coming together to provide a more integrated RCM solution. And then, lastly, and this is probably relevant to AI in all industries, you know, obviously very closely monitoring the regulatory front. And we watch the regulatory front because developments there could either stimulate or stymie growth in the sector. So, I think that's kind of a final thing that that everyone needs to keep a close eye on.
16:29
Chris
Okay, got it. All right, well that's all the time we have for today. But, Ryan, is there anything else you'd like to highlight from the report that we didn’t cover?
16:35
Ryan
No, I would say, as usual, you clearly reviewed the report in detail. I think you hit on all the key topics. I would just remind everyone we do have a much larger piece out there. If anyone who wants that piece, they can call or email us. We're happy to send it. It does list a lot of the public and private entities in the sector, so that may be of interest. And, again, we're happy to provide a copy to anyone. Other than that, I think we covered the key points. I really appreciate it.
16:59
Chris
Great. So, for those interested in reading the full mosaic report, it's called “The Growing Importance of AI and the Revenue Cycle Management Marketplace.” You can request copy by reaching out to us at WilliamBlair.com/contact-us or to Ryan directly. All right, Ryan, thanks again for joining us.
17:18
Ryan
Thanks for having me. Happy holidays to everyone. You too, Chris. Thanks.