AI vs. AI: The New Revenue Cycle Arms Race
The next fight over hospital revenue will not happen across a negotiating table.
It will happen between algorithms.
Health plans are increasingly using artificial intelligence to review claims, challenge coding, identify medical-necessity issues and accelerate denials. Hospitals are responding with their own AI tools for clinical documentation, coding, claim preparation and appeal development.
Welcome to the new revenue cycle arms race.
The stakes are getting higher
Initial claim-denial rates have climbed to nearly 12%, according to data cited by HFMA. Every denial delays cash, adds administrative expense and forces hospitals to prove—often repeatedly—that medically necessary care was appropriately delivered and billed.
This is no longer just a back-office efficiency problem. It is a direct threat to liquidity, operating margin and the organization’s ability to invest in patient care.
AI is accelerating the conflict.
Healthcare organizations spent approximately $1.4 billion on AI in 2025, with health systems accounting for roughly $1 billion of that investment. HCA Healthcare expects its AI initiatives—including revenue-cycle automation and clinical documentation—to produce approximately $400 million in 2026 savings.
Payers are investing heavily, too.
The result is an uncomfortable reality: One algorithm helps construct the claim while another looks for a reason not to pay it.
A paid claim is not necessarily a correctly paid claim
Denials receive attention because they are visible. They land in work queues, receive reason codes and generate appeal activity.
Underpayments are far quieter.
A payer can process a claim, issue payment and still reimburse the hospital incorrectly. Unless the payment is tested against the contract, the variance may be posted as a contractual adjustment and disappear.
That makes underpayments one of the most overlooked forms of revenue leakage in healthcare.
Hospitals must validate payments against:
- Contracted reimbursement terms
- DRG and APC calculations
- Stop-loss and outlier provisions
- Carve-outs
- Bundling and multiple-procedure logic
- Transfer and readmission policies
- Payer-specific reimbursement rules
The question is no longer simply, “Did the payer pay us?”
The right question is, “Did the payer pay us everything it contractually owed?”
More technology will not automatically solve the problem
Hospitals are being offered AI for nearly every part of the revenue cycle: eligibility, authorization, CDI, coding, claim edits, status follow-up, denial prediction, appeal writing and payment variance.
Some of these tools can produce tremendous value.
Others simply create another platform, another work queue and another vendor invoice.
Before approving an AI investment, CFOs should demand answers to five questions:
- What specific problem are we solving?
- What is the current financial baseline?
- How much cash or expense improvement should the tool produce?
- How will results be independently validated?
- Can the output survive payer, compliance and regulatory scrutiny?
AI-generated revenue that cannot withstand an audit is not sustainable revenue.
The real advantage is intelligence, not automation
The strongest revenue-cycle organizations will not be those with the most bots.
They will be the ones that connect information across patient access, utilization management, clinical documentation, coding, billing, contracting, payment posting and denials.
That intelligence should reveal:
- Which payers are creating the most preventable friction
- Which services generate the greatest authorization risk
- Where DRG downgrades and medical-necessity denials are concentrated
- Which underpayments are systemic rather than isolated
- Which appeals produce recoverable cash
- Which workflows consume resources without generating a return
Automation handles transactions.
Intelligence changes decisions.
What acute-care CFOs should do now
The most important step is to establish an enterprise payer-performance command center.
It should combine denials, underpayments, authorization failures, payment variances, appeal outcomes and payer escalations into one financial view.
At a minimum, CFOs should monitor:
- Net collection rate by payer and service line
- Initial and final denial rates
- Denied dollars ultimately lost
- Expected versus actual reimbursement
- Authorization-related write-offs
- Appeal success and cash yield
- Cost to collect
- Days from discharge to final payment
The goal is not to produce another dashboard. It is to identify where payer behavior is preventing the organization from converting earned revenue into cash.
The bottom line
AI will not eliminate the conflict between hospitals and health plans. It will make that conflict faster, more complex and more expensive.
Hospitals cannot win by adding bots to broken processes. They need stronger revenue integrity, better contract intelligence, defensible documentation and clear financial accountability.
The organizations that succeed will use AI to prevent avoidable leakage, identify incorrect reimbursement and focus human expertise on the accounts that matter most.
In the AI era, the winner will not be the organization that submits the most claims or generates the most appeals.
It will be the organization that can prove—quickly, accurately and at scale—exactly what it is owed.
CycleIQ Health — Turning revenue-cycle complexity into actionable intelligence.