MGMA Takeaways: Where AI Actually Fits in Denial Management
At this year’s MGMA conference, revenue cycle leaders gathered to discuss a question many healthcare organizations are asking right now: What role will AI really play in denial management?
The conversation surfaced a mix of excitement, skepticism, and realism. While AI is appearing across the healthcare technology landscape, panelists were clear about one thing: denial management is not a problem that automation alone will solve.
The real opportunity lies in understanding where AI can strengthen revenue cycle operations - and where human expertise still matters most.
Denials Are Both Preventable and Inevitable
The panel opened with a discussion of denial drivers across healthcare organizations.
Many denials remain highly preventable. Front-end errors such as: eligibility verification, documentation gaps, and coding issues - continue to account for a large share of lost revenue. Technology and analytics can help surface payer rules and policy requirements earlier in the process so that claims are submitted correctly the first time.
But not every denial can be prevented. Even well-run organizations will still face a portion of denials driven by payer policy changes, adjudication quirks, and the sheer fragmentation of the healthcare payment ecosystem.
In other words, prevention matters, but investigation and recovery remain essential parts of revenue cycle operations.
As Joyful Health co-founder, Eliana Berger, noted during the discussion:
“There will always be denials that pop up no matter what, and they’re going to take time, energy, and attention to resolve.”
The Science and the Art of Revenue Cycle
One framework that resonated during the discussion was the idea that revenue cycle work contains both science and art.
Some parts of revenue cycle operations are highly structured and repeatable:
These tasks follow defined processes and can be supported by automation.
- eligibility checks
- rules-based coding validation
- policy lookups
- claim submission requirements
But other parts of revenue cycle require judgment and expertise:
- interpreting payer responses
- investigating claim history
- constructing appeals
- resolving ambiguous denials
Understanding that distinction is critical when evaluating where AI fits into denial management.
As our co-founder, Eliana Berger, explained during the panel:
“There are pieces of revenue cycle that are science, and pieces that are art. The art requires expertise. The science is where technology can really help.”
A New Framework for AI in Revenue Cycle
Rather than thinking about AI as a single solution, the panel explored a more nuanced model emerging in healthcare technology.
Today, organizations are beginning to deploy AI in three different ways:
- AI Agents: These systems execute well-defined tasks that follow clear rules. They are most effective in highly structured workflows, such as eligibility checks or data validation.
- AI Copilots: These tools assist human operators inside their workflow. They surface insights or recommendations while still relying on human judgment for final decisions.
- AI-Supported Services: Some areas of revenue cycle remain too complex to fully automate. In these cases, AI supports specialized teams who investigate and resolve claims.
Denial management and aged accounts receivable often fall into this third category.
Because payer rules change constantly—and because claims involve multiple data sources and contextual interpretation—these workflows still require human expertise working alongside AI-driven investigation tools.
Why AI Is Becoming Practical Now
If denial management has been complex for decades, why is AI gaining traction today?
One answer surfaced repeatedly during the discussion: interoperability.
Revenue cycle data is scattered across numerous systems:
- EHR platforms
- billing systems
- clearinghouses
- payer portals
- remittance data
- bank records
- payer contracts
When information lives in separate silos, it becomes difficult to reconstruct what actually happened to a claim. AI is becoming valuable because it can help connect these fragmented data sources and reason across them.
Instead of relying on manual investigation across multiple systems, AI can help reconstruct how revenue should have moved through the payment lifecycle—and where the breakdown occurred.
Joyful Health co-founder, Eliana Berger, highlighted this challenge during the panel:
“Our financial data is scattered across seven or more different places… when you can’t track a visit end-to-end, it becomes very difficult to understand what’s happening in revenue cycle.”
The Technology Arms Race in Healthcare
Another theme that surfaced during the discussion is the growing asymmetry between providers and payers.
Payers have invested heavily in automation and AI-driven adjudication engines. In some cases, claims are being processed or denied by automated systems at scale.
For providers, this means that revenue cycle operations must evolve as well. Technology alone won’t eliminate denials, but it can provide the investigative intelligence needed to identify patterns, respond faster, and recover revenue that might otherwise go uncollected.
As one panelist noted:
“Stop using people to fight robots.”
The Reality of Technology Adoption
Despite the promise of AI, the panel also acknowledged a practical challenge facing many organizations: technology fatigue.
Healthcare providers often operate with complex technology stacks and fragmented systems. Adding new tools can create additional operational burden if they require major workflow changes.
As a result, the most effective technologies today are those that integrate directly into existing workflows rather than forcing organizations to rebuild their operations around new systems.
Increasingly, healthcare leaders are looking for solutions that can operate within their existing infrastructure—augmenting teams rather than replacing them.
Denial Management Is Still One of the Hardest Jobs in Healthcare
Perhaps the most important takeaway from the discussion is that denial management remains one of the most difficult areas of healthcare operations.
Teams are asked to navigate:
- changing payer policies
- fragmented data environments
- complex adjudication systems
- increasing denial rates
All while maintaining financial stability for their organizations.
AI can help. But the goal is not to eliminate the people doing this work—it is to equip them with better tools to investigate, understand, and resolve revenue challenges.
As our co-founder, Eliana Berger, noted in closing:
“This is a really hard job… and unfortunately it’s getting harder day by day.”
The Path Forward
Conversations like the one at MGMA reflect a broader shift underway in healthcare revenue cycle.
The question is no longer whether AI will appear in denial management, it's how organizations will combine technology and expertise to manage the growing complexity of healthcare payment systems.
The most successful models will likely be those that treat AI not as a replacement for revenue cycle teams—but as the investigative infrastructure that helps those teams recover the revenue they’ve already earned.
For organizations looking to better understand where revenue is breaking down—and how new approaches to denial investigation and recovery might fit within their existing operations, Joyful Health is always open to continuing the conversation.
If you’d like to learn more about how Joyful works with healthcare organizations to investigate denials, recover unpaid claims, and improve financial visibility, you can reach out to our team here.
