Everyone in America will eventually depend on the healthcare system, yet getting the care a doctor recommends can mean fighting through weeks of insurance paperwork. We must fix prior authorization and demand healthcare AI that understands clinical evidence, insurance policies, and state regulations. Otherwise, we risk building a future where technology accelerates administrative mistakes while patients suffer preventable harm.
Prior authorization requires doctors to obtain an insurer’s approval before certain treatments or medications can be covered. The goal is safe and effective care while maintaining cost. In practice, the process too often turns physicians into clerical workers and patients into collateral damage. Modern medicine can perform extraordinary feats, but those capabilities mean little when administrative barriers prevent patients from accessing treatment.
The numbers are damning. According to the American Medical Association’s 2025 Prior Authorization Physician Survey, physicians and their staff spend an average of 13 hours each week completing authorizations, handling approximately 40 requests per physician. Ninety-five percent of surveyed physicians report treatment delays, 79 percent say authorization problems can lead patients to abandon treatment, and 94 percent say the process contributes to burnout. Most alarming, 26 percent report that prior authorization has caused a serious adverse event for a patient.
These are warning signs of a system that can turn a treatable condition into a medical crisis.
The underlying problem runs deeper than excessive paperwork. Healthcare decisions depend on context, and that context changes constantly. The appropriate treatment criteria may vary by insurer, insurance product, diagnosis, and state. A commercial health plan and a workers’ compensation policy from the same insurer can follow different rules.
Consider a physician treating a patient with a rotator cuff injury. Whether the tear is acute or chronic can affect the relevant treatment criteria. If the doctor consults the wrong guideline, misses a documentation requirement, or relies on an electronic health record checkbox that fails to capture the insurer’s exact criteria, a necessary treatment may be denied. The information exists. The challenge is finding the right information for the right patient at the right time.
Artificial Intelligence Could Help, Or Not
Healthcare leaders are dangerously tempted to treat the AI label as proof of competence. Generic AI tools can produce confident, convincing answers without understanding the specific rules governing a clinical decision. In healthcare, a plausible answer is not necessarily a correct one. When technology overlooks a critical distinction, it can reinforce denials, waste clinicians’ time, and delay treatment.
The promise of better technology is real, but so is the risk of deploying tools without the expertise needed to make them reliable. AI must connect clinical evidence to the precise policy, jurisdiction, and circumstances involved. It should show users the applicable criteria and supporting sources, flag uncertainty, and send exceptions to qualified human reviewers. Automation should remove routine friction, not replace clinical judgment.
Nor can we rely on insurers’ promises alone. The AMA survey reports that in June 2025, more than 60 insurers pledged voluntary reforms to prior authorization. Yet it found that only 33 percent of physicians expected the commitments to make a meaningful difference. Trust will not return through pledges. It requires measurable results.
Healthcare executives, insurers, and providers must act now. Rigorously test AI tools against real-world clinical and regulatory scenarios before deploying them. Demand transparent sources, accurate policy matching, clear escalation procedures, and independent measurement of denials, delays, and patient outcomes. Regulators must also require meaningful reporting and hold organizations accountable when administrative practices obstruct necessary care.
Get this right, and clinicians can reclaim time for patients, decisions can become more consistent and more transparent, and more people can receive appropriate care without navigating an administrative maze. Get it wrong, and we will automate confusion, deepen distrust, and leave vulnerable patients waiting while their conditions worsen.
Healthcare technology should make the path to treatment clearer, faster, and safer. The measure of success is whether a patient gets the right care before a preventable delay changes their life, not how impressive an AI demonstration looks.
Jamie LaPaglia, RN, CCM, is founder of Claim Clarity, an AI-powered workers’ compensation technology company. All opinions are the author’s own.