Blog
Article
With one-quarter of health care spending classified as waste, AI systems built to reason—not just automate—may be key to restoring clinical capacity and operational trust.
Ganesh Padmanabhan: ©Autonomize AI
America’s health care system is suffering from a chronic condition that no prescription can cure: administrative waste.
Despite spending over $5.1 trillion annually—nearly 19% of GDP—the U.S. continues to experience systemic inefficiencies that add no clinical value. According to a 2019 JAMA study, between $760 billion and $935 billion—roughly 25% of total health care spending—is considered waste, with administrative complexity named the single largest contributor.
The cost shows up not only in budgets but also in burnout, workflow breakdowns, and care delays. Health plans and providers alike are impacted—forced to devote vast resources to documentation, compliance, and review tasks that sap time from higher-impact activities.
While technology has long promised relief, the real question is whether health care is investing in solutions that actually understand the complexity they’re meant to solve.
Because this isn’t a software problem. It’s a systems problem. And the stakes go far beyond dollars—they affect access, quality, and trust.
No one enters medicine to spend hours on paperwork. But for care managers, nurses, and physicians across the system, the administrative burden has become inescapable.
Take prior authorization—a vital process for ensuring appropriate care and managing utilization. Yet even when done well, it often takes too long. Each request can take 35 minutes or more to process, requiring the assembly of data from EHRs, scanned PDFs, faxes, and handwritten notes. Multiply that across tens of thousands of requests per day, and organizations face thousands of hours lost to manual review.
This isn’t about blame—it’s about bandwidth. Plans are trying to ensure adherence to guidelines. Providers are trying to move fast for patients. And both sides are working within legacy systems that simply can’t keep up.
The result is friction for everyone. And that’s where policy and technology are beginning to converge.
This issue has sparked a wave of recent legislation. Most notably, Congress introduced the Reducing Medically Unnecessary Delays in Care Act of 2025, which aims to reduce administrative bottlenecks in clinical decision-making.
The Centers for Medicare and Medicaid Services has also signaled action, exploring proposals that would reduce the number of medical services subject to prior authorization and streamline the process itself. The objective isn’t to eliminate review—it’s to make it faster, more consistent, and less burdensome for providers and patients.
Health plans are already aligning around this goal. But even the best policy changes can’t solve for fragmented systems, siloed data, and overloaded reviewers.
What’s needed is intelligent infrastructure—technology that doesn’t just digitize tasks, but understands and augments them.
Over the last decade, health plans and providers have tried to address administrative challenges with optical character recognition tools, robotic process automation, and traditional large language models. And these solutions have brought incremental improvements.
But they often fail in real-world conditions.
Health care generates more unstructured data than any other industry. In fact, up to 80% of health data is unstructured, according to the JAMA Health Forum. Faxes. Handwritten notes. Disparate EHR systems. General-purpose AI simply isn’t equipped to ingest this kind of chaos—much less reason through it.
That’s where Agentic AI enters the conversation.
Agentic AI refers to a new class of intelligent software agents that can interpret context, reason across criteria, and collaborate with humans in real time. These are not just automation scripts. They’re digital co-workers—task-specific agents that embed directly into workflows and get smarter as they go.
In the case of prior authorization, an AI Copilot built on agentic principles can transform a traditionally slow and manual process into an intelligent, collaborative workflow. It pulls relevant data from charts, scanned PDFs, EMRs, and even handwritten forms; summarizes the patient’s medical history to accelerate clinical review; and adjudicates requests against guidelines such as InterQual, Carelon, or internal plan criteria. Most importantly, it generates clear, explainable recommendations that link directly back to the source data—ensuring transparency and trust. Clinicians remain firmly in control, making the final decisions, while the copilot lifts the administrative burden that once consumed hours of valuable time.
A Fortune 100 health plan recently implemented an Agentic AI copilot across its utilization management workflow. The outcome?
55% faster processing times
Up to 36,000+ clinical hours saved monthly
Significant FTE time savings, allowing for greater focus on patient care activities
16% improvement in Medicare turnaround compliance
These aren’t theoretical gains. They’re measurable benefits—helping plans operate more efficiently, while supporting providers in making timely care decisions. And the ultimate winner? The patient.
By accelerating review, aligning documentation, and reducing clerical workload, Agentic AI helps unlock faster access to necessary treatments—without compromising clinical rigor or compliance.
Behind the metrics are real people. Nurses toggle through 10 systems to complete a single review. Patients caught between office staff and insurers. Care teams trying to stay compliant while moving fast.
Agentic AI doesn’t remove the human—it removes the noise.
It returns time to providers. It reduces stress for plan reviewers. It allows both sides to focus on what matters: making high-quality care decisions, supported by data, aligned with policy, and made with confidence.
As Dr. Atul Gawande famously put it:
"The complexity of medicine is only increasing. The answer isn’t more checklists. It’s systems that can absorb complexity and return clarity.”
The administrative reckoning has arrived. Legislative, economic, and workforce pressures are all converging on one conclusion: the way health care does work must change.
But not through more portals. Not through more paperwork. Through smarter systems built for health care, not retrofitted to it.
Agentic AI is not a buzzword. It’s a shift in how clinical and operational teams partner with technology. It’s a way to scale capability, not just capacity. Plans and providers who embrace this shift won’t just see cost savings or compliance boosts. They’ll create a better experience for patients—one where care moves faster, decisions are more informed, and burnout gives way to bandwidth.
Ganesh Padmanabhan is a technology entrepreneur and AI strategist focused on healthcare transformation. He is a leading voice on the role of agentic AI in optimizing clinical and administrative workflows and is the CEO and founder of Autonomize AI.