The $70 Complaint

Transcript

Healthcare is complicated. Not because the industry does not know what it wants to accomplish. Better outcomes. Better experiences. Lower costs. Smarter operations. The complication is everything that has to happen behind the scenes to make those things possible. This is Healthcare Power Ups, a podcast that looks underneath healthcare operations to understand the people, processes, technology, and economics that make the system work. Not the PowerPoint version. The real version. Because sometimes the smallest part of healthcare operations can reveal something much bigger about the entire system. And this story begins when someone tells a health plan, I think you got it wrong.

Imagine a patient goes to the doctor. The doctor orders an MRI. The claim goes to the insurance company. And somewhere inside a massive healthcare operation, a decision is made. Denied. Maybe the service was not considered medically necessary. Maybe an authorization was missing. Maybe the documentation did not support the procedure. Maybe there was a coding issue. Or maybe the provider disagreed with how the contract was applied. Whatever the reason, someone looks at the decision and says, No, take another look. That is where an appeal begins.

It is tempting to think of an appeal as simply another transaction. A claim comes in. A decision is made. Someone disagrees. The health plan reviews it again. Case closed. Except that is not really what happens. Because to understand an appeal, the health plan often has to reconstruct everything that happened before it. And that turns an appeal into something surprisingly complicated. Someone may need the original claim, the member history, provider documentation, the authorization, health plan policies, clinical criteria, contract terms, previous correspondence, standard operating procedures, and potentially several other pieces of information sitting in completely different systems.

An appeal is not really one transaction. It is an investigation. And investigations are expensive. Consider the economics. A claim might cost somewhere around six or seven dollars to process. An appeal can cost closer to seventy or eighty dollars. A disagreement about a transaction can cost roughly ten times as much as processing the original transaction itself. Why? Because claims are designed for scale. Appeals are designed for exceptions. And exceptions require judgment.

An appeal arrives. First, someone has to determine what it is. Who submitted it? What claim does it belong to? Is it a member appeal or a provider appeal? Is it clinical or administrative? Pre service or post service? Then the organization has to acknowledge that it received the appeal. Classify it. Route it. Research it. Determine which rules apply. Review the documentation. Make a determination. Generate the appropriate communication. Send it. And document everything that happened.

Throughout the entire process, the clock is ticking. Because appeals are not simply an operational process, they are also a compliance process. There are deadlines, reporting requirements, audit requirements, and rules governing what happens when certain timelines are not met. Now imagine managing all of that manually. In some organizations, people still do. Sagility appeals experts have seen large health plans manage portions of their appeals operations with something remarkably familiar, an Excel spreadsheet.

The claim system was not designed to manage the appeal. So someone tracks the appeal in a spreadsheet, then goes into the claim system, then perhaps another system for documentation, another for clinical policies, another for correspondence, and then back to the spreadsheet. Healthcare has some of the most sophisticated technology in the world, and sometimes the workflow connecting it all together is still a person clicking between windows.

This is where artificial intelligence becomes interesting. For years, healthcare organizations have used automation. If this happens, do that. Move this information here. Open this screen. Populate this field. Send this document. That is incredibly useful. But an appeal presents a different challenge. Because the hardest part is not necessarily moving the information. It is understanding the information.

Imagine an appeal arrives with several attachments. Someone or something has to determine: What is the provider actually questioning? What evidence did they submit? What information is missing? What does the member claim history reveal? Which medical policy applies? Which clinical criteria matter? What does the health plan own operating procedure say should happen? And ultimately, what should happen next? That is not simply a robotic automation problem. It is a reasoning problem. And that is what makes appeals such an interesting use case for AI.

Imagine several intelligent agents working together. One reads the incoming documents. Another reconstructs the claim history. Another searches clinical and administrative policies. Another searches the organization procedures and knowledge base. Then those pieces come together to give the person reviewing the appeal something they have not historically had: A recommended next best action. Not simply, here is the document, but, here is what happened, here is the relevant policy, here is the evidence, here is what is missing, and here is what should be considered next.

That distinction matters. Because healthcare AI is not necessarily about removing the human from difficult decisions. In many cases, it is about changing what the human spends time doing. Instead of searching, they review. Instead of assembling, they evaluate. Instead of navigating five systems, they make a decision. And there is already an interesting glimpse of what that can mean. In one appeals operation discussed by Sagility team, improving guidance retrieval helped increase productivity from roughly four cases per hour to approximately seven and a half.

Not because people suddenly worked twice as hard, but because they spent less time searching for the information they needed to do their jobs. That is an important distinction. Technology did not make the person faster, it removed some of the things making the person slow. But this is where the story gets more interesting. Because if AI is only used to process appeals faster, it may be solving the wrong problem.

Remember what an appeal represents. Something happened upstream. A claim was submitted. A decision was made. And someone disagreed with that decision. Maybe the decision was correct. Maybe it was not. Either way, the appeal is evidence of friction somewhere in the system. And when thousands or millions of those signals are viewed together, very different questions become possible. Why are providers appealing this particular procedure? Why does this type of claim repeatedly generate disputes? Are providers misunderstanding a documentation requirement? Is a policy confusing? Is something happening during authorization? Is claims configuration producing avoidable denials?

Are the same members calling the contact center afterward? Are certain appeals eventually becoming complaints? Suddenly, the goal is no longer simply process the appeal faster. The goal becomes understand why the appeal is happening at all. Healthcare operations teams increasingly describe this idea as shifting left. Move intelligence earlier in the process. If appeals repeatedly reveal a problem with claim adjudication, improve the adjudication. If providers repeatedly submit the wrong documentation, improve provider education. If an administrative policy is generating unnecessary disputes, examine the policy. If something happening in utilization management eventually becomes an appeal, connect those dots.

Because every appeal that can be appropriately prevented does not just eliminate an appeal. It can potentially eliminate everything surrounding it. The intake, the research, the clinical review, the correspondence, the provider call, the member call, the rework, the compliance burden, and the seventy or eighty dollar investigation. That changes the economics considerably.

And those economics are becoming even more interesting because providers are getting smarter too. Healthcare providers increasingly have access to analytics capable of identifying claims that may be worth appealing. Imagine a system looking across thousands of denied claims and identifying one with a meaningful probability of being overturned. From the provider perspective, submitting that appeal may be relatively inexpensive. If enough appeals succeed, the economics make sense. But the equation looks very different on the payer side. The provider incurs the cost of submitting the appeal, while the health plan incurs the cost of investigating it.

That asymmetry matters. And it helps explain why health plans are increasingly focused on appeals automation and intelligence. As the tools for identifying appeal opportunities become more sophisticated, the operational machinery for resolving them has to become more sophisticated too. Healthcare may be entering a strange kind of technological arms race. AI identifies something worth appealing. AI helps submit it. AI helps the payer understand it. AI retrieves the relevant evidence. AI supports the determination. And eventually, intelligence generated from those appeals flows upstream so fewer of them need to exist in the first place.

That last step may ultimately be the most important. Because the measure of a great appeals operation might not always be how efficiently did the organization process the appeal. It might eventually become how many avoidable appeals did the organization prevent. That is a very different way of thinking about healthcare operations. For decades, healthcare has organized work around transactions: claims, calls, appeals, authorizations, clinical reviews, provider inquiries. Each has its own team, its own technology, its own metrics, its own workflow.

But a member does not experience six workflows. Neither does a provider. They experience one healthcare system. An incorrect claim can become a provider call. The provider call can become an appeal. The unresolved appeal can become a complaint. Suddenly, what appears to be four different operational processes is actually one story. The more interesting question is whether the health plan can see it that way. That is where intelligence capabilities such as CoreIQ begin to matter differently. Not simply because they provide another dashboard, but because they create the possibility of connecting what happened in one part of the operation with what happens next.

An appeal does not have to remain an isolated transaction. It can become a signal connected to claims, provider behavior, member interactions, clinical processes, and downstream outcomes. Sagility broader Synchrony approach takes the idea another step by bringing together workflow, AI enabled intelligence, healthcare domain expertise, and operations. Rather than optimizing isolated steps, the organization can increasingly manage toward the outcome. The underlying model brings together platform capabilities, agentic technology, and people rather than treating them as disconnected components.

Because perhaps the most important question is not how quickly did the appeal close. It is why did the appeal exist. Think back to the MRI. The claim was denied. The provider disagreed. An appeal was submitted. Someone gathered the documentation. Someone searched the medical policy. Someone reviewed the claim history. Someone made a determination. Someone sent a letter. The case closed. Traditionally, that would be considered success.

But now imagine the system learns something from it. Maybe it discovers that hundreds of providers are making the same mistake. Maybe the same documentation is repeatedly missing. Maybe the same claims configuration is producing the same denial. Maybe the same policy is generating unnecessary confusion. Now that single appeal has done something much more valuable than simply getting resolved. It has taught the healthcare system something. And if the system is intelligent enough to listen, the next appeal might never happen.

That is this Healthcare Power Up. One piece of healthcare operations pulled apart, put back together, and made a little easier to understand. Because in healthcare, the most important question may not always be how efficiently a problem was resolved. Sometimes the better question is why did the problem exist in the first place. And perhaps the most powerful question of all: What actually happens next?