The Case for Accountability in Prior Authorization Automation
In our recent post, , we explored how legacy technology is quietly draining healthcare budgets and stalling the modernization that CMS-0057 demands. But there’s a second, equally important thread running through that conversation: artificial intelligence.
AI is everywhere in healthcare conversations right now. When it comes to prior authorization specifically, the discussion has taken on a sharper edge. Congressional hearings. State legislation. Physician groups raising alarms. The promise of AI-driven efficiency is real, and so is the concern about what happens when automation gets it wrong.
The challenge for the industry is understanding where AI genuinely helps, where it introduces risk, and what responsible adoption looks like in practice.
Adoption Is Growing, But Scaling Remains Elusive
West Monroe’s recent insurance modernization research offers useful context. Nearly 60% of insurers surveyed had moved past the pilot stage for generative AI, but most deployments remained small-scale and fragmented. In claims, one of the most automation-ready functions in the industry, only 29% had reached production at scale.
The Barriers Are Human, Not Technical
What’s holding organizations back? The top barriers are not technical. Twenty-four percent cited resistance to change or cultural inertia. Twenty-three percent struggled with limited perceived value or unclear benefits. Only 13% pointed to actual technology performance issues. The tools are increasingly capable, while the organizational conditions for adoption are still catching up.
In prior authorization specifically, deploying AI responsibly requires more than a capable algorithm. It requires clean, structured medical policy data, deep integration with existing utilization management and claims platforms, and clinical governance processes that ensure human judgment is applied where it matters most.
The Regulatory Environment Is Tightening
The regulatory environment around AI in prior authorization has intensified considerably. In 2025, more than 250 AI-related healthcare bills were introduced across 47 states, with 33 signed into law. The consistent thread across those measures: prohibiting the use of AI alone to deny care, requiring human review of algorithm-driven decisions, and mandating transparency when automated systems are used in coverage determinations.
WISeR: A Federal Test Case
At the federal level, CMS launched the Wasteful and Inappropriate Service Reduction (WISeR) model in January 2026, deploying AI-powered prior authorization for select Medicare services across six states. While the model aims to reduce waste and improve efficiency, it has drawn concern from provider groups and lawmakers who worry about erroneous denials and inadequate safeguards for patients. Major physician groups, including the AMA, have called for greater transparency and stronger accountability mechanisms as the program gets underway.
The regulatory direction is clear: automation is welcome, but it has to be accountable.
Automate Approvals. Protect Denials.
The practical implication is a straightforward principle: automate approvals and preserve human review for denials.
Approvals that meet clear clinical criteria can move faster with automation, reducing turnaround times, cutting administrative burden for providers, and freeing up clinical staff for cases that genuinely require their expertise. That’s where AI delivers undeniable value.
Denials carry clinical, legal, and ethical weight. Routing every denial through human clinical review reflects the right governance model. It protects patients, protects the organization, and aligns with the accountability standards that state and federal regulators are actively enforcing. Organizations that have already built human review into their denial workflows are better positioned as oversight requirements continue to expand.
Building the Right Foundation
When asked to identify their top technology priorities, only 29% of survey respondents selected AI, automation, or analytics, and just 10% chose enabling the workforce with modern tools. Core application modernization and cloud infrastructure dominated the list. The risk is that organizations focused narrowly on foundational upgrades may find themselves underprepared when it comes time to deploy AI at scale.
Accountability Pays Off
Organizations that modernize their prior authorization infrastructure now, digitizing policies, implementing compliant FHIR-based workflows, and deploying AI with appropriate human oversight, are not just checking a compliance box. They are building operational capabilities that compound over time. Faster turnaround times improve provider relationships. Touchless auto-approvals reduce administrative costs. Transparent, auditable processes reduce regulatory exposure.
AI belongs in prior authorization. The priority for organizations today is building the right foundation to deploy it well, and doing so before the January 2027 deadline demands it.
Want to see how responsible prior authorization automation works in practice? Schedule a demo with the Itiliti Health team.