From Denials to Approvals: Why the Time to Act on AI in Prior Authorization Is Now

Last week, I had the honor of speaking at the American Rhinologic Society Summer Sinus Symposium in San Diego. What began as a conversation about AI’s promise quickly turned into a candid reckoning with its misuse—specifically, how insurance companies are now deploying generative AI to block access to medically necessary care.

The conversations I had—with surgeons, medtech executives, and key opinion leaders—kept circling back to one pivotal question: How deeply is this hurting our patients, and how broadly is it damaging our industry?

The answers were unequivocal.

The Silent Crisis: AI Denials Are Slashing Access and Slowing Innovation

Across surgical practices, we’re seeing 20–40% of cases denied—not by human reviewers, but by AI algorithms operating behind opaque black boxes. These rates are highest in practices that lack mitigation strategies and in regions where insurers have deployed more sophisticated denial technologies.

Let’s be blunt: This is not theory. It’s already here.

• UnitedHealth reportedly has over 1,000 active AI use cases across its claims infrastructure. (Becker’s Payer Issues)
• A 2023 OIG report found 13% of Medicare Advantage denials violated CMS coverage rules—a number likely amplified by algorithmic triage. (OIG Report)

For innovative medical device companies, the stakes are just as high. Internal modeling shows prior authorization denials—especially those accelerated by AI—pose a material and rising threat to U.S. medtech innovation. In real terms, they eliminate $3–11 billion in sales annually, hitting the high-tech segment hardest (Nemedic White Paper). When access is blocked at scale, pipelines slow, and innovation stalls.

The Playbook Insurers Are Using—and the Legal Reckoning Ahead

Here are just a few examples now playing out in courts and Congress:

• Cigna’s “PxDx” Engine: Allegedly denied claims in batches without physician review. Class action approved in California. (PPI Benefits)
• UnitedHealth’s naviHealth Model: Flagged Medicare Advantage patients for early discharge—regardless of clinical readiness. Certified class action, Feb 2025. Healthcare Finance News
• Multi-Payer Legal Scrutiny: Bloomberg Law is tracking a surge of lawsuits challenging the legality of AI-driven “black box” denial tools. Bloomberg Law

And according to the AMA, 93% of physicians report care delays due to prior authorization—a number now supercharged by AI operating at machine speed. AMA Survey

Countermeasures: What Surgeons and Device Teams Can Do Now

While regulatory and legal remedies are essential, they take time. Here are AI-enabled strategies being deployed today to stem denial rates and accelerate care:

Source: AAO-HNS Bulletin – A Novel AI-Based Method of Mitigating Prior Authorization Denial

Emerging tools like Gemini 1.5 Pro and Grok are redefining what’s possible—giving early adopters real-time access to payer policy shifts and longer-context medical reasoning.

From Defensive to Offensive: Building AI Muscle Responsibly

Whether you’re a surgeon or a device executive, now is the time to shift from defense to offense.

Insurers are not just experimenting—they’re industrializing. To respond, we need a wartime mindset that embraces intelligent automation while safeguarding patient trust.

Compliance guardrails are essential:

• Public LLMs like ChatGPT and Gemini don’t sign BAAs—keep PHI out of prompts.
• The OCR (June 2024) guidance warns IP+deidentified text can still be PHI.
• Even when training is toggled off, your prompts may train the model. (OpenAI, Google)
• PCI/PHI/PII data must never enter LLMs—use redaction or secure enterprise deployments.
• Default to de-identification, enterprise tools, audit logs, and prompt training.

Checklist: De-identify | Use enterprise-grade tools | Enable audit logging | Train staff on AI-safe workflows

Final Thought: Don’t Let AI Deny Your Future

At last week’s symposium, one truth emerged above all others: AI is now the new table stakes in prior authorization. On both sides of the chessboard.

The clinicians and device leaders who embrace AI responsibly will streamline approvals, reduce administrative friction, and protect the IP fueling our next-generation breakthroughs.

Those who don’t? Risk having their patients, products, and hard-earned innovations denied by someone else’s algorithm.

To learn more about how AI denials are impacting your business, please access our white paper AI, Prior Authorization, and the Hidden Drag on U.S. Device Innovation.

Let’s move—together—from denial to approval.