Health systems cannot cut their way to revenue cycle sustainability. That is not an opinion — it is the consensus conclusion from the HFMA 2026 Annual Conference executive roundtable on revenue cycle performance. CFOs from Essentia Health, LifeBridge Health, critical access hospitals, and multi-facility health systems arrived at the same conclusion independently: the cost-cutting playbook that defined the last decade of RCM strategy has hit a structural wall. Denial rates are at 9% and climbing. AR days have stretched to 42. Combined denial losses and uncompensated care hit $48 billion across 2,300+ hospitals in 2025 — a 25% increase year-over-year. You cannot cut your way out of a revenue problem this large. You have to build your way out.
AI is not the cost-cutting tool the industry has been marketing. It is the sustainability infrastructure that replaces misaligned vendor partnerships with accountable, transparent, results-driven automation — at a fraction of the cost and without the structural conflicts that make traditional billing companies a shrinking-returns proposition.
The HFMA Consensus: Cost-to-Collect Alone Suboptimizes Revenue
Kevin Boren, East Region CFO at Essentia Health Duluth, framed the problem precisely at the HFMA 2026 roundtable: "You can't manage solely on cost to collect without suboptimizing revenue." The margin compression is real — every health system in the room acknowledged it. But the CFOs who have been managing exclusively on cost-to-collect have discovered that cutting billing staff, reducing follow-up intensity, and shifting to the cheapest vendor available produces a predictable result: revenue leaks faster than costs shrink.
"Health systems cannot cut their way to sustainability. They must protect and grow revenue while using technology responsibly, addressing workforce constraints and creating more accountable partnerships." — Matt Leshy, Signature Performance
Adam Conley from a critical access system put it even more directly: "It's impossible for us to cut our way to sustainability. There aren't enough people and contracts to cut. We have to grow our way to sustainability."
This is the strategic pivot that the healthcare RCM industry is making in 2026. The question is no longer "how do we reduce costs?" It is "how do we protect and grow revenue at a sustainable cost?" Those are fundamentally different problems — and they demand fundamentally different tools.
The Partnership Problem: Your Billing Company's Incentives Are Misaligned
Traditional billing company partnerships operate on a model designed for the vendor's sustainability, not the practice's. The typical arrangement: a billing company charges 5–9% of collected revenue, takes over some or all of the revenue cycle, and reports monthly on performance metrics they define and measure themselves.
Three structural misalignments make this model unsustainable:
1. Volume Over Quality
Percentage-of-collections models reward claim volume regardless of claim quality. Submitting more claims — even dirty claims that will be denied — generates activity. The vendor bills on collections, not on clean claim rate. The practice absorbs the denial rework, the delayed cash flow, and the staff time spent reconciling discrepancies the vendor created.
2. The 20% Denial Tax
When practices outsource denial management, vendor partners capture up to 20% of claim reimbursement on recovered denials, according to HFMA roundtable participant Dennis Jones. That is a structural tax on the practice's revenue — paid on claims that should never have been denied in the first place. AI that makes claims clean on the first submission eliminates the denial pipeline entirely, turning that 20% recovery fee into retained revenue.
3. Short-Term Profitability Over Long-Term Outcomes
Matt Leshy identified the core failure mode: "The partner's operating model rewards short-term profitability over long-term outcomes." The hospital absorbs the downstream cost — inconsistent service, governance friction, diminished institutional knowledge. Every time the vendor turns over a team or reassigns accounts, the practice loses accumulated payer-specific intelligence that took months to build. The vendor's P&L improves; the practice's revenue cycle degrades.
The Labor Crisis: You Can't Hire Your Way Out Either
If the billing company model is broken, the obvious alternative is to build an in-house team. Except you can't — not at any cost that makes mathematical sense.
Desmond Jackson, serving a rural health system, described serving a "labor desert" with no revenue cycle talent pipeline. Jeff Costello confirmed the same dynamic in a different market: "In our market, we can't hire problems away. There's just too much need and not enough revenue cycle talent available."
The numbers validate the anecdotes. HFMA's 2026 staffing benchmarks document 30–40% annual turnover in billing positions — meaning a 10-person billing department replaces 3–4 employees every year. Each replacement cycle costs 50–90% of the role's salary in fully loaded recruiting, onboarding, and productivity ramp-up costs. And the replacement often arrives with no institutional knowledge of the payer-specific rules, quirks, and workarounds that the departing employee carried in their head.
Adam Conley crystallized the AI implications: "AI is going to lead to the replacement of people who don't know how to use it." The workforce is not disappearing — it is bifurcating. Staff who can work alongside AI to handle complex exceptions and relationship-dependent tasks become more valuable. Roles that consist of repetitive data entry, status checking, and form submission are already being displaced.
The Offshore Displacement: AI's Real Target
Dennis Jones offered the most direct assessment of where AI displacement actually lands: "AI-driven layoffs are going to affect offshoring companies more. I've been to Mumbai. At 7 p.m., you see thousands of people pouring into buildings to do repetitive, high-volume, sometimes low-balance work. That's the AI target."
This is not speculative. The offshore RCM model is built on labor arbitrage — paying workers in lower-cost markets to perform tasks that American workers price too high. AI eliminates the arbitrage entirely. An AI agent processes claims, verifies eligibility, submits prior authorizations, and manages denials at a cost structure that undercuts offshore labor by an order of magnitude — without time zone lag, training ramp-up, or communication overhead.
The displacement is already underway. Practices that previously outsourced to offshore partners are discovering that AI handles the same work with higher accuracy, faster turnaround, and zero coordination cost. The billing company's competitive advantage — cheap labor — is no longer cheap enough to compete with AI agents that operate 24/7 at fixed pricing.
The Takeback Trap: Fighting to Capture AND Keep Revenue
Even when claims are paid, the fight is not over. Kenneth Hogue from the HFMA roundtable flagged a pattern that every billing manager recognizes but few vendors address: takebacks are rising even as initial denial rates drop.
Payers are paying claims faster — then auditing them retrospectively and demanding repayment 30, 60, or 90 days after the check clears. The practice records revenue, allocates it, and then receives a recoupment notice that claws it back. The cash flow impact is devastating. The administrative burden of fighting takebacks is often higher than fighting the original denial would have been.
Hogue's principle: "Trust but verify." Audit the services your vendor partners are delivering. Verify that denial follow-up is actually happening. Confirm that paid claims are documented thoroughly enough to survive post-payment audit. Continuity is critical — and it is exactly what high-turnover billing companies and offshore teams cannot provide.
AI creates an auditable record from the moment of claim creation. Every modifier choice, every medical necessity linkage, every payer-specific rule application is documented and defensible. When the takeback demand arrives, the defense is pre-built — not scrambled together from whatever notes the departed billing specialist may have left behind.
The AI Accountability Model: What Replaces the Billing Company
The replacement for the broken billing company model is not a better billing company. It is a structurally different approach — one where the incentives, the economics, and the operational model align with the practice's actual goal: maximum clean revenue at minimum sustainable cost.
Here is what that looks like:
| Dimension | Billing Company Model | AI Accountability Model |
|---|---|---|
| Pricing | 5–9% of collections (rewards volume) | Fixed pricing (rewards accuracy) |
| Denial approach | Reactive recovery at 20% of reimbursement | Prevention — claims clean first time |
| Staffing | 30–40% annual turnover | Zero turnover, 24/7 operations |
| Transparency | Vendor-defined, self-reported metrics | Real-time dashboards, auditable decisions |
| Knowledge retention | Walks out the door with each departure | Accumulates permanently across every claim |
| Takeback defense | Reactive, often undocumented | Pre-built, auditable from submission |
| Labor market dependency | Competing for shrinking talent pool | None — operates in labor deserts |
The transition is not theoretical. Dennis Jones framed the build-vs-buy decision practically: Epic-integrated AI solutions may take 18 months to deploy; smaller, purpose-built AI companies can deliver working solutions now. The key is choosing AI that addresses the full revenue cycle — not another point solution that adds a layer of complexity without removing the underlying dependency on human-intensive processes.
From Reactive Denial Management to Upstream Prevention
The HFMA September 2026 analysis — "From Reactive Denial Management to Upstream Prevention" — quantifies why the cost-cutting model fails structurally:
- 15% of claims are initially denied across the industry
- Hospitals spend $19.7 billion per year on the appeals process — and only half of appeals succeed
- 68% of RCM leaders identify inaccurate or incomplete patient data at intake as the primary denial driver (Experian 2025)
- OIG's 2026 federal review found 97% of appealed Medicare Advantage SNF denials from one major authorization contractor were overturned — meaning the denials should never have been issued
- Combined final denials and uncompensated care reached $48 billion in net revenue loss across 2,300+ hospitals — up 25% year-over-year
The critical concept: move from retrospective visibility to in-flight intervention. Billing companies operate in the retrospective model — they see the denial after it happens and try to recover. AI operates upstream, catching the data error at intake, the authorization gap before the procedure, the coding mismatch before submission. The claim never enters the denial pipeline because the denial trigger was eliminated before the claim was built.
That is not a cost reduction. That is a structural change in how revenue is captured — the kind of change that makes sustainability possible without endless cutting.
What This Means for Your Practice in Q4 2026
Nearly half of healthcare executives identify revenue cycle as their top IT investment area for the coming year. The HFMA roundtable consensus is clear: the practices and health systems that survive the current margin compression will be those that replace volume-dependent, turnover-prone, misaligned vendor relationships with AI infrastructure that makes claims clean the first time, operates 24/7 without labor market dependency, and retains every piece of payer intelligence permanently.
The question is not whether AI replaces the billing company model. It is whether your practice makes the transition before Q4 denial rates, open enrollment coverage changes, and takeback campaigns compound into a revenue gap that no amount of cost-cutting can close.