AI isn’t coming to healthcare finance. It’s already here. And while providers dream of AI unlocking efficiency and better patient care, payers are deploying AI agents with a very different mission: deny more claims, faster.
They call it “payment integrity.” In reality, it’s denial automation at scale. And the numbers show the stakes couldn’t be higher.
AI Denials, the UnitedHealthcare Case Study
According to a U.S. Senate report, UnitedHealthcare’s denial rate for post-acute care doubled between 2020 and 2022 as it rolled out automated tools like the nH Predict model from its subsidiary naviHealth.
- Denials for skilled nursing facility claims increased nine-fold.
- Post-acute denial rates jumped from 10.9% to 22.7%.
- A CBS News lawsuit report alleges United knowingly used a flawed AI model because it knew only 0.2% of policyholders ever appeal, despite 80–90% of denials being overturned when challenged.
As Fox Business reported, UnitedHealthcare is now accused of systemically denying claims with AI automation.
Not Just United, an Industry Trend
United isn’t alone. A Senate subcommittee investigation revealed that CVS Health and Humana have followed similar playbooks:
- CVS’s “Post-Acute Analytics” report projected $77.3M in savings by increasing denials (Healthcare Dive).
- Humana’s denial rate for long-term acute care rose 54% in two years (Healthcare Dive).
As Fierce Healthcare summarized:
“UnitedHealth, CVS and Humana used technology to increase Medicare Advantage prior authorization denials for post-acute services, boosting profits.”
The Human Cost of AI Denials
While payers profit, patients and providers pay the price.
- The Guardian reported claims being denied “every 1.2 seconds,” often reversed months later — but not before patients went bankrupt or went without care.
- Financial Times detailed how United’s NaviHealth tool reduced human oversight while producing “unfair denials.”
- Investopedia noted that 61% of physicians are worried about AI’s role in denying coverage.
- Even Wikipedia’s entry on nH Predict now documents how AI-driven denial models have been scrutinized for accuracy and fairness.
As Senator Richard Blumenthal stated in the Senate investigation:
“Medicare Advantage insurers are intentionally using prior authorization to boost profits by targeting costly yet critical stays in post-acute care facilities.”
“This Doesn’t Affect Me — I’m Not an Acute Care Provider” Rebuttal
At first glance, it may seem like AI-driven denials are only a concern for post-acute care providers, like skilled nursing facilities or long-term acute care hospitals. But that assumption is dangerously short-sighted.
Here’s why this impacts every provider organization:
- AI denial engines don’t stop at one specialty. Once trained and validated on post-acute claims, payers will inevitably expand them across other service lines including office visits, outpatient procedures, diagnostics, and more.
- Patterns are transferable. If AI models learn that denials yield profit (because appeals are rare), the same logic will be applied to ambulatory care and specialty practices.
- Denial strategies evolve faster than regulation. As NORC’s research found, regulators can’t keep up with payer AI adoption. Today it’s post-acute; tomorrow it could be the procedures your practice depends on.
Bottom line is that even if your practice hasn’t felt the full brunt of payer AI yet, it’s coming. The best defense is to get ahead now, before these denial strategies target your claims.
Providers Can Fight Back by Leveraging the White Plume Advantage
If payers are using AI to deny claims, providers must use Analytics + AI to fight back. At White Plume, our STAR² Ai platform is built for exactly this battle.
Here’s how we level the field:
- Data First = Advantage
By analyzing encounter data in real time, STAR² Ai identifies revenue at risk before the payer’s AI ever sees the claim. - Coders as Editors, Not Creators
Our AI does the heavy lifting, empowering coders to review, edit, and validate, thereby ensuring claims go out accurate, defensible, and fully optimized. - Visibility into Denials
With analytics tracking overturn rates and payer trends, providers finally have visibility into how payer AI behaves and the evidence to push back. - Continuous Improvement
Just as payers update their denial engines, STAR² Ai continuously learns from encounter data to sharpen its revenue protection capabilities.
The Takeaway: Don’t Let Payer AI Win by Default
The evidence is clear: payers are already deploying AI to deny claims. They’re counting on providers to stay reactive: to fight denials manually, appeal slowly, or accept revenue leakage as the cost of doing business.
But providers don’t have to play defense. By leveraging White Plume’s data, analytics, and AI, organizations can get ahead of payer AI, not just react to it.
The future of RCM isn’t about asking whether AI will be used. It’s about who controls it, and whose side it’s on.
