Written by Chuck Feerick, CEO of Latitude Health
Why AI in Prior Authorization Must Focus on the Manual Review Process
The conversation around automation in prior authorization (PA) has advanced considerably in the last five years. Yet much of the attention continues to center on improving the submission experience — auto-approvals, standardized transactions, and digital handoffs between providers and payers. While these are necessary areas of focus, they are not sufficient for solving the operational bottlenecks that affect clinical quality, administrative cost, or regulatory compliance.
Current data from the 2022 CAQH Index shows that only 25 to 30 percent of prior authorization requests are submitted electronically using the standard transaction. That means 70 to 75 percent of requests still arrive through non-integrated channels such as fax, phone, or health plan portals. These submissions, by their nature, require downstream manual processing, and they represent the majority of prior auth volume in the U.S. healthcare system.
Even when electronic submissions are used, the majority of utilization management (UM) tasks remain manual. Preparing a case for review typically involves navigating multiple systems, locating relevant clinical documentation, comparing requests to plan-specific medical policies, and generating decision rationales that meet internal and regulatory standards. Across all lines of business, this work is most often performed by nurses and medical directors — clinical resources whose time is expensive and increasingly constrained.
Estimates from McKinsey suggest that roughly 50 percent of these manual UM tasks could be automated using modern AI tools, particularly if those tools are designed to support the review stage, rather than just triage or submission. Automating intake or status tracking is useful, but the highest leverage lies in augmenting the decision-making process itself: transforming unstructured records into structured data, guiding reviewers toward relevant criteria, and accelerating the documentation of determinations.
New regulations from CMS, taking effect in 2026, will further increase the urgency of addressing these bottlenecks. The proposed rules mandate the use of FHIR-based APIs for electronic prior authorization across Medicare Advantage, Medicaid, and Exchange plans. They also introduce tighter turnaround requirements — seven calendar days for standard requests, and 72 hours for expedited ones — alongside new transparency obligations. Plans will be required to publish prior auth metrics, including approval rates and average decision times. These changes are designed to reduce delays in care and improve accountability, but they also compress the operational timeline for plans that are already facing high volumes and limited staffing capacity.
Most AI solutions on the market today are not built to meet this reality. They are either submission-focused tools that require provider behavior change, or general-purpose language models that lack alignment with plan-specific policies and data environments. Neither approach is well-suited to the task of reviewing complex medical documentation and rendering a clinically appropriate determination in a compliant format.
At Latitude, we’ve taken a different approach. Our platform is designed to operate at the center of the UM workflow, supporting reviewers directly by automating the most time-consuming and repetitive parts of their process. This includes structuring inbound documentation, surfacing the appropriate policy criteria, highlighting missing information, and generating evidence-backed rationale. The intent is not to replace clinicians, but to allow them to work more efficiently and consistently — especially in high-volume specialties or during periods of constrained capacity.
AI in UM must evolve beyond static models and pre-scripted rules. It must incorporate context — including prior decisions, eligibility constraints, plan design differences, and emerging clinical guidance — and apply that context dynamically at the point of review. This is the direction the field is moving, and it’s where meaningful gains in turnaround time, quality, and consistency will come from.
In short, accelerating prior authorization requires more than digital submissions. It requires intelligent systems that improve how decisions are made. As volume increases and regulatory scrutiny intensifies, the ability to scale the clinical review process — without compromising standards — will be the key to success.
About Latitude Health
Latitude Health is transforming Utilization Management (UM) and Prior Authorization with AI-native technology that drives better outcomes for organizations. Our platform empowers UM teams to work smarter and faster - reducing costs, increasing review volume, and improving decision quality. By equipping teams with tools and insights, Latitude supercharges your greatest asset—your people—while delivering impact across all UM operations.
Want to see Latitude Health in action? Reach out to our team here for a Demo!