Prior Authorization Built for Control, Not Speed
Prior authorization was created to manage cost and utilization, but in modern care delivery, especially same-day and on-demand, it often feels misaligned with how healthcare actually works. Care moves in real time, while authorizations crawl through portals, faxes, and phone calls.
In a recent Office Hours conversation, Evan Martin, VP of Revenue Cycle Management at ZoomCare and co-host of the Wilshire IT RevCast Podcast, shared what this disconnect looks like in real-world operations. He also discussed how intelligent, AI-powered prior authorization and agentic workflows can begin to bring structure and speed to the chaos.
On-Demand Care Meets Outdated Authorization Systems
ZoomCare operates around 50 primary care clinics offering urgent and specialty services throughout the Pacific Northwest. The model is patient-centric by design. Roughly 80 percent of patients self-schedule, often for same-day appointments that may include advanced imaging.
From an access perspective, this is ideal. From a prior authorization standpoint, it creates a daily clash between clinical urgency and administrative bottlenecks. The team often faces a difficult decision: prioritize timely, appropriate care and pursue retro authorization support afterward, or delay care while chasing an urgent auth first.
ZoomCare chooses the patient. They deliver medically necessary services and then race to secure retro authorizations within strict payer timelines. With clinics open from 7 a.m. to 10 p.m. and prior auth staff working standard business hours, authorizations for early morning or late-night visits often start at a disadvantage. One authorization specialist supports all 50 clinics, with backup staff pulled in only when volume spikes. The real bottleneck is not submitting the request. It is locating, assembling, and presenting the right documentation fast enough.
How Agentic AI Reshapes the Workflow
Traditional automation, such as screen-scraping bots, can move data from point A to point B. But Evan Martin believes that prior authorization now requires more. He sees a need for agentic AI in healthcare workflows that behaves like a capable teammate, not just a tool.
In practice, this means AI agents that understand payer rules, organizational logic, and task sequencing. Rather than a human pulling data manually, an agent could pre-gather documentation for a retro auth, apply payer-specific rules, flag missing elements, and assemble a complete package for a specialist to review and submit. This is the next step in prior authorization automation.
The greatest value comes from timing. For pre-scheduled authorizations, AI agents can often manage the process end-to-end. For urgent imaging that needs to happen within 20 minutes, agents can instantly retrieve relevant histories, clinical indications, prior imaging, and decision-making context, while a human team member remains on the phone with the payer. Evan envisions a future where prior auth specialists are “co-piloted” by multiple AI agents, each managing a specific part of the process, with ambient AI listening and launching background tasks in real time. Humans remain in control. AI eliminates the scavenger hunt.
Injectables, Coverage Gaps, and Real-Time Decision Making
Injectables and in-room procedures introduce another layer of complexity due to the split between medical and pharmacy benefits. ZoomCare bills most services under place of service 11 to keep system costs lower. While financially efficient, this increases prior auth requirements for injectables and raises the chance of white bagging or brown bagging scenarios.
Providers are trying to deliver care during the same visit. Patients, however, face financial trade-offs. Some are willing to pay out of pocket for immediate treatment, while others prefer to wait for a covered alternative. Evan sees a clear opportunity for AI to assist at this moment of decision. A smart benefits agent could evaluate the patient’s coverage in real time, determine if prior auth is required, flag specialty pharmacy rules, and present clinical and financial options to both the provider and the patient. That is true patient access optimization.
The goal is to give providers the information they need to lead with clinical judgment, and to give patients full transparency about costs and timing—similar to the philosophy behind ABNs in Medicare, but modernized with real-time intelligence.
Bridging the Gap Between Clinical Judgment and Policy
The most persistent challenge remains the gap between what clinicians consider medically necessary and what payers are willing to cover. Providers base decisions on evidence and training. Payers use policy to manage cost, often applying a different definition of necessity.
Evan does not expect that tension to vanish, but he believes agentic AI can minimize its impact on care teams. Rather than relying on staff to memorize every payer rule, integrated AI agents could continuously update policy logic, cross-reference current orders, and surface guidance at the moment of care. Inside the EHR, providers would see a plain-language prompt showing whether a service is likely to be approved and what alternatives may be required.
If Evan were advising CMS or a major payer, his wish list would start with shifting prior authorization from a pre-service barrier to a post-service review model anchored in national guidelines and published evidence. Let clinical judgment set the course, and let utilization review and refine the care plan, not dictate it.
While policy may take time to catch up, agentic AI can begin reshaping this environment today. By organizing data, surfacing payer rules, and coordinating tasks, AI gives healthcare teams the time and clarity they need to refocus on the patient, not the paperwork.
To see how intelligent AI orchestration can reduce prior authorization chaos, improve patient access, and support your clinical teams, request a demo if you’re interested.