Artificial intelligence has become healthcare’s favorite scapegoat, especially in insurance. Every denied claim, every delayed approval, every inexplicable “not medically necessary” response seems to get blamed on an algorithm in the background.
Much of the frustration lands on prior authorization, the process that insurers use to approve medical services before they’re delivered. When prior authorization works, it confirms coverage and prevents surprise bills. When it fails, it delays or denies care that clinicians have already determined is necessary.
That failure is felt most acutely by physicians and patients. According to a 2024 American Medical Association survey, nearly six in ten physicians believe automation has made prior auth worse, not better. At the same time, prior authorization already delays care for 94% of patients and forces more than 80% of physicians to watch patients abandon treatment altogether because the approval process is too slow or too confusing.
But blaming AI misses what’s actually broken. The problem isn’t that technology is involved in prior authorization, but how only one side is really using it.
Insurers process millions of claims annually. According to industry surveys, 84% report using artificial intelligence and machine learning to help flag, route, and deny requests at scale. Providers, meanwhile, are still manually assembling documentation, navigating dozens of payer portals, and responding to denials that offer little explanation. Automation exists, but the imbalance has created a system optimized for payors and punitive for everyone else.
Why the imbalance matters
Prior authorization was never meant to be a punishment. It’s supposed to be a checkpoint: confirm coverage, confirm medical necessity, protect patients from surprise bills. Instead, it has evolved into a highly optimized process for insurers and a painfully manual one for providers. When only one side of a transaction operates at machine speed, friction becomes inevitable.
Insurers have invested heavily in systems that accelerate review and rejection. Providers are stuck with fax machines, PDFs, phone calls, and inconsistent requirements that vary not only by insurer but by plan, state, and procedure.
What’s the result? Predictable requests are denied not because care is inappropriate, but because something small is missing: a mismatched code, a missing signature, outdated eligibility information, the wrong form version, or documentation submitted out of sequence. Industry analyses show that administrative and technical errors account for a majority of claim denials with estimates ranging from 40% to nearly 60% depending on specialty and payer mix.
These errors clearly reflect a process failure. It’s a dynamic that has been played over and over again. Clinician burnout is a similar story. Physicians aren’t leaving healthcare because they disagree with utilization review. They’re leaving because they’re drowning in administrative tasks that basic automation could prevent.
Federal regulators are forcing change
In 2024, the Centers for Medicare & Medicaid Services finalized the Interoperability and Prior Authorization Rule (CMS-0057-F), a sweeping effort to modernize how payers handle approvals.
Since January 1st, Medicare Advantage, Medicaid managed care plans, CHIP plans, and ACA marketplace insurers have been required to respond to standard prior authorization requests within seven days, and urgent requests within 72 hours. By 2027, they must also support standardized electronic APIs that allow providers and patients to track authorization status and access decision rationales digitally.
The regulation matters for two reasons.
First, it creates enforceable expectations. Delays can no longer be shrugged off as operational complexity. Second, it forces transparency. Insurers will have to show how decisions are made, how long they take, and how often requests are denied. This data will become measurable and visible, which is vital for understanding claim denials.
While employer-sponsored commercial plans aren’t directly covered by these requirements, market pressure will likely push the entire industry toward similar standards over time. Organizations prepared to operate in a more interoperable, data-driven environment will be rewarded. Those still relying on manual workarounds will be exposed.
Where automation actually helps providers
AI is poorly suited to making medical necessity decisions in isolation. That concern is legitimate, and regulators should continue to draw clear boundaries around clinical judgment.
Before a request ever reaches a payer, technology can identify whether an authorization is required, verify eligibility in real time, confirm benefit limitations, and ensure the correct documentation is attached in the correct format. It can flag inconsistencies before submission rather than after denial.
This is about removing administrative friction so physicians can focus on patient care. Getting documentation right at intake now matters more than ever, because faster payer decisions mean less margin for error on the provider side. Inaccurate submissions will be denied more quickly.
It also addresses one of the most under-discussed sources of patient harm: administrative delay. The AMA reports that 18% of physicians say prior authorization has led to a serious adverse event for a patient.
Those harms are rarely caused by a disagreement over clinical evidence. They are caused by time.
What this means for healthcare organizations
Some form of utilization review will always exist in a system balancing cost, access, and outcomes. Can prior authorization function without causing preventable delays and administrative exhaustion? At this current time, there isn’t a simple solution.
Regulated automation can level the playing field. By mandating standardized data exchange, decision timelines, and transparent explanations, CMS is building infrastructure where technology can be audited, compared, and improved. Tools that reduce errors and speed approvals will stand out. Tools that simply accelerate denials will become harder to justify.
Other industries have modernized complex approval systems without sacrificing accountability. Banking standardized underwriting. Aviation embedded safeguards into every workflow. Healthcare can do the same.
Technology alone won’t solve prior authorization. But refusing to adopt it guarantees the status quo: a system where insurers operate at machine speed, providers operate at human speed, and patients pay the price in delayed care.
Healthcare organizations should investigate systems that reduce friction between discharge and follow-up, that verify documentation completeness before submission rather than after denial, and that give clinical teams visibility into the status of requests.
We need to move past “it’s always been done this way.” Prior authorization is broken. Regulation alone won’t fix everything overnight, but it creates the conditions for technology to be used responsibly: catching errors early, reducing friction, and getting patients into care faster.
Photo: nevarpp, Getty Images
Jeffrey Morelli is the Co-Founder and CEO of Silna, the industry’s first Care Readiness Platform built to eliminate the administrative barriers that delay care. Silna automates benefit checks, eligibility, and prior authorizations across 1,000+ payors, delivering 99%+ accuracy and cutting authorization turnaround times from weeks to hours.
Backed by $27M from Accel and Bain Capital Ventures, Silna now supports 100,000+ patients nationwide and is trusted by leading healthcare organizations for its ability to reduce denials, accelerate treatment, and remove front-end revenue cycle bottlenecks. Before Silna, Jeff spent a decade in San Francisco’s tech sector building and scaling products for highly regulated industries, including leading go-to-market at Truework (Series C; acquired by Checkr).
This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.
