The Hardest Part of Healthcare Might Be Getting Paid

Transcript

Imagine you run a hospital. Every day, physicians see patients. Nurses provide care. Procedures are performed. Tests are ordered. Medications are administered. Patients go home. The healthcare happened. Now you just have to get paid for it. Simple, right? Except first you have to document exactly what happened. Translate that care into the right codes. Determine what the patient’s insurance covers. Make sure any required authorization was obtained. Create the claim. Submit it. Get it through the payer’s edits. And then wait. Maybe the claim gets paid. Maybe part of it gets paid. Maybe it gets denied. Maybe someone has to figure out why.

Maybe the payer needs additional documentation. Maybe the denial requires a clinical review. Maybe somebody has to write an appeal. Maybe the patient owes the remaining balance. And then somebody has to collect that too. Suddenly, providing the care was only half the story. Because in healthcare, delivering a service and getting paid for that service can be two very different things. This is Healthcare Power Ups, a podcast about how healthcare actually works. In each episode, we’re going to take one part of the healthcare system, pull it apart, and figure out what really happens underneath. Not the PowerPoint version. The real version. The people. The processes. The technology. The economics. And most importantly, why it matters.

Today we’re talking about the provider side of healthcare. And one deceptively simple question: Why is it so hard for healthcare providers to get paid? We recently sat down with Titus Leo, Sagility’s provider leader. I started by asking him what pressures are forcing healthcare providers to rethink their administrative and revenue cycle operations. His answer wasn’t AI. It wasn’t technology. It wasn’t even staffing. It was something much more fundamental. Margin. And Titus described the problem as being caught between two hammers. On one side, the amount of money coming in is under pressure. On the other, the cost of operating the organization keeps rising. And the provider is stuck in the middle.

That’s important because revenue cycle management, or RCM, as you’ll hear it called, isn’t simply the department that sends bills. It’s the financial engine that converts healthcare delivery into cash. And that cash has to fund everything else. Salaries. Facilities. Equipment. Technology. Innovation. The operation itself. You can provide exceptional care, but if the financial machinery behind that care doesn’t work, the organization eventually has a very different kind of healthcare problem. So let’s pull apart that machinery. The revenue cycle begins earlier than most people realize. It can start before the patient ever walks through the door. Is the patient eligible? What does their insurance cover? Does the procedure require prior authorization? Has that authorization been obtained?

Then care happens. Now it has to be documented accurately. The services have to be coded. Charges have to be captured. A claim has to be generated. The claim needs to be accurate enough to get through the payer’s systems. Then it gets submitted. And that’s when things get interesting. Because the provider and the payer enter the same transaction from two very different directions. The provider is saying: We delivered this care. Pay us. The payer is asking: Was this covered? Was it authorized? Was it medically necessary? Was it coded correctly? Is this the amount we actually owe? That tension creates enormous friction.

And Titus points to two places where you can see it particularly clearly. At the beginning: prior authorization. And at the other end: denials. Prior authorization essentially asks whether a service meets the requirements for coverage before it happens. A denial happens when the payer doesn’t pay all or part of the claim after it has been submitted. Between those two points sits an enormous amount of administrative work. And here’s the interesting part: The administrative work on one side often creates administrative work on the other. A payer asks for additional information. A provider retrieves it. A claim is denied. A provider investigates it. A provider calls. The payer answers. The provider appeals. The payer reviews the appeal.

Every additional interaction consumes resources on both sides. Which is why payer provider friction isn’t just frustrating. It’s expensive. Now add another party to the equation. The patient. Over time, patients have taken on greater responsibility for portions of their healthcare costs through deductibles, copays and other cost sharing arrangements. That means providers increasingly have to collect not only from insurance companies, but from the people receiving the care. And patient collections create an entirely different challenge. Imagine receiving a medical bill. It is competing with your mortgage or rent. Groceries. Utilities. Childcare. Car payments. Everything else a household has to pay.

At the same time, a hospital can’t necessarily approach a patient relationship like an ordinary debt collection exercise. That patient may return next month. Their children may receive care there. Their parents may receive care there. The provider has to collect what is owed while preserving a relationship built around healthcare. That’s why technology has become so important on the patient side of the revenue cycle. Digital portals. Text notifications. Electronic statements. Simple payment links. Mobile payment options. The objective isn’t merely to make collections cheaper. It’s to remove friction. Make the bill easier to understand. Make the payment easier to complete. Make the entire interaction feel less like a collections process and more like a normal digital experience.

And that leads to a bigger question. If the revenue cycle contains this many steps, shouldn’t technology be able to fix it? Yes. But maybe not in the way you think. For years, healthcare organizations have continuously improved pieces of the revenue cycle. First through better systems. Then through workflow tools. Then robotic process automation. Then machine learning. Then generative AI. And now agentic AI. Each wave can make individual tasks faster or more efficient. And Titus makes an important point here: Incremental improvement still matters. If a process changes, improve it. If a repetitive task can be automated, automate it. If analytics can help someone make a better decision, use them. Healthcare operations don’t have to wait for some grand transformation before getting better.

But there’s another possibility emerging. Instead of improving each step independently, what if we started connecting them? Think about what happens today when a claim isn’t paid. A person receives the account. They investigate. They look at information from different sources. They determine why the claim wasn’t paid. They decide what action to take. They may adjust something. They may submit additional documentation. They may follow up. They may escalate. Eventually, hopefully, the account gets resolved. Now imagine an operating model where technology can perform more of that sequence. Not simply suggesting the next step. Actually moving the work forward. That’s where agentic AI becomes interesting.

An AI enabled workflow might analyze an account. Understand its status. Retrieve the appropriate information. Determine the next permitted action. Execute defined steps. Monitor what happens. And continue moving the account toward resolution. For some accounts, a person might never need to touch it. For others, technology might do the research and preparation before handing the case to an experienced associate. And for the most complex situations, particularly those involving clinical judgment, the person remains central. That creates a very different model. Instead of giving every account to a person and using technology to make that person slightly faster, you begin asking: Which work actually requires a person? That is a much bigger transformation.

And interestingly, Sagility has been moving in this direction for years. Long before today’s generative AI boom, analytics were already being used to help determine which accounts deserved attention. Think about a hospital with thousands of outstanding accounts. They’re not all equally likely to be collected. They’re not equally complex. And they’re not equally valuable. So why treat them all the same? Predictive and propensity models can help score those accounts. Which ones appear highly collectible? Which require more expertise? Which should be prioritized? Which may ultimately become write offs? Now you can match the work to the resource.

Experienced people focus on complicated, higher value problems. Less experienced people can concentrate on more defined tasks. Technology helps orchestrate who, or what, should handle the work. That’s one form of intelligence. Clinical denials take the idea even further. Suppose a claim is denied for a reason involving medical necessity. Now we’re not simply talking about an administrative discrepancy. Someone may need to understand the clinical record. Historically, a nurse might manually review the documentation, determine whether an appeal is appropriate, write the appeal and then begin the follow up process. That’s expensive expertise. And more importantly, it’s expertise you don’t want spending unnecessary time on administrative work.

Now AI can help prepare that review. It can analyze information. Surface relevant details. Assist with drafting correspondence. Organize the evidence. The nurse still provides the clinical judgment. But the nurse doesn’t necessarily have to perform every step required to get to that judgment. That’s a critical distinction. AI isn’t valuable because it replaces expertise. It can be valuable because it allows expertise to concentrate on the work where expertise actually matters. And that same principle can extend across the revenue cycle. An associate’s worklist can be prioritized by AI. Knowledge tools can help answer questions in real time. Generative AI can assist with documentation. Predictive models can identify patterns. Agentic workflows can increasingly complete defined activities.

Some people working inside an AI enabled operation may not even think of themselves as “using AI.” It’s simply embedded in how the work gets done. That’s probably an important sign of maturity. Because eventually, we may stop talking about AI as a separate tool altogether. Nobody walks into an office and announces: I’m using the internet today. It’s simply part of the environment. AI may increasingly work the same way. And for a hospital CEO or CFO, that matters because they don’t necessarily need another AI product. They need better financial performance. They want more of the money they’re legitimately owed. They want it faster. They want fewer denials. They want less rework. They want a lower cost to collect. And they want the patient experience protected along the way.

So the technology conversation has to connect back to those outcomes. That can also mean working with the technology the provider already has. Many healthcare organizations have invested heavily in modern clinical and patient accounting platforms. Those platforms continue getting smarter. They’re adding automation. They’re adding AI. They’re improving documentation and billing workflows. That’s good. A transformation partner shouldn’t need the client’s technology to be bad in order to create value. In fact, the more interesting opportunity may be connecting what happens between systems and across workflows. Look at denial trends. Why are they happening? Are some of those problems actually originating upstream? Could something change in documentation? Could a rule be improved? Could a workflow be redesigned? Could intelligence from the back end prevent a problem at the front end?

That’s where RCM stops being simply a collection operation. It becomes a learning system. Every denial tells you something. Every underpayment tells you something. Every repeated follow up tells you something. Every patient call tells you something. The question is whether the organization captures those signals and uses them to improve what happens next. And that’s where provider transformation starts connecting to a much bigger healthcare story. Because many of the problems we’re describing aren’t really “provider problems.” And they aren’t really “payer problems.” They’re problems created between providers and payers. Remember that denied claim? The provider spends money investigating it. The payer spends money responding. The provider sends documentation. The payer reviews documentation. The provider calls. The payer staffs a contact center to answer.

What if some of those interactions simply didn’t require people anymore? Imagine a provider’s technology communicating directly with a payer’s technology. A claim needs information. The systems exchange it. A status needs to be checked. The systems resolve it. A defined discrepancy needs action. Agents communicate behind the scenes. No hold music. No manual follow up. No person on the provider side calling another person on the payer side just so two systems can eventually arrive at the same answer. Not every interaction can work this way. Some issues will remain complicated. Some will require judgment. Some will require human conversation. But what if a meaningful portion didn’t? Now something interesting happens. The payer saves money. The provider saves money. The claim gets resolved faster. And potentially, the patient never experiences the administrative friction at all.

That’s why Titus believes the future of provider operations isn’t simply about automating today’s work. It’s about reducing the amount of unnecessary work that exists between organizations. And there is another force pushing healthcare in that direction. Workforce shortages. Healthcare organizations cannot assume that every administrative or clinical challenge can be solved simply by adding more people. In some markets and specialties, the people aren’t available. In others, they’re simply becoming too expensive. That means technology enabled operating models will increasingly be necessary. But again, the goal shouldn’t be replacing people for the sake of replacing people. The better question is: Where does a person create the most value? If technology can handle a routine administrative interaction, let it. If AI can prepare information for review, let it. If an agent can move a straightforward account toward resolution, let it.

Then put people where judgment, empathy, expertise and accountability matter. That may become one of the defining ideas of the next generation of healthcare operations. Because Titus’s view of the future isn’t a healthcare system without people. It’s a healthcare system where people spend less time doing work technology can handle. And where technology increasingly handles the invisible administrative transactions connecting the pieces of healthcare. The interesting part is what happens to the economics when that begins at scale. Today, healthcare organizations spend enormous effort managing friction. Following up. Correcting. Calling. Reviewing. Resubmitting. Escalating. Collecting. Imagine turning even a portion of those activities into intelligent, connected workflows. Suddenly, improving revenue cycle performance isn’t only about collecting harder. It’s about designing a system that creates fewer reasons to chase the money in the first place.

And that may be the most important idea in this entire conversation. Because the traditional question in RCM is: How do we collect more? The better question may eventually become: Why was this so difficult to collect in the first place? Fix that problem. Then fix the next one. Use analytics to understand where the friction originates. Use technology to remove unnecessary work. Use AI where it improves the economics. Use people where judgment matters. And connect the pieces so that what happens at the back of the revenue cycle makes the front of it smarter. That’s how you start moving from revenue cycle management toward revenue cycle intelligence. And perhaps, eventually, toward something even bigger: A healthcare ecosystem where providers and payers don’t have to spend quite so much money arguing about money.

Because here’s the irony. Every dollar spent managing unnecessary administrative friction is a dollar somebody in healthcare had to spend. The payer. The provider. Sometimes the patient. Making the system more efficient isn’t simply about one side winning. Sometimes the biggest opportunity is eliminating work neither side particularly wanted to do in the first place. And if we can do that? Maybe the hardest part of delivering healthcare won’t always be getting paid for it. And that is today’s Healthcare Power Up. One piece of healthcare operations. Pulled apart. Put back together. And hopefully, a little easier to understand than it was twenty minutes ago. Until next time, keep asking the question that tends to make healthcare a lot more interesting: What actually happens next?