Artificial intelligence is rapidly changing how health insurers review claims, authorize services, and determine coverage. While payers argue that automation speeds up and standardizes work, many physician practices are experiencing a different reality: rising denial rates, growing administrative workloads, and longer reimbursement cycles.

The truth is, claim denials are increasing across provider organizations, particularly among practices with significant Medicare Advantage (MA) patient populations. At the same time, physicians and revenue cycle teams are spending more time managing prior authorizations, appeals, and documentation requirements.

The issue is not simply that practices are making more billing mistakes. Instead, many providers are exploring a new claims environment where automated review systems evaluate claims against increasingly complex documentation, coding, and authorization requirements.

Understanding how these systems work and how practices can adapt has become essential for protecting revenue and reducing avoidable denials.

The Numbers Say Denials Are Getting Worse

The trend data is consistent enough that it is hard to dismiss. Experian Health’s State of Claims 2025 survey, a quantitative poll of 250 healthcare professionals, 41% of providers report denial rates exceeding 10% of submitted claims, up from 38% in 2024 and 30% in 2022.

MGMA’s March 2024 Stat poll asked 235 medical group leaders whether denial rates had changed over the prior year. Sixty percent said denials had increased. Only 11% said they had gone down.

MDaudit’s 2025 Benchmark Report, drawn from real-world data across more than 1.2 million providers and 4,500 facilities, found that average denied inpatient claim amounts rose 12% year over year, while denied outpatient amounts rose 14%. The average dollar amount per denial tied to requests for information and medical necessity rose by 70%, reaching $450 per denial.

Practices experiencing this trend are not making more billing errors. Payer behavior has changed. Understanding why matters for how you respond to it.

What Payer AI Actually Does

The most important point is not that insurers use software. It is how they use it. A 2024 Senate Permanent Subcommittee on Investigations report said UnitedHealthcare’s prior authorization denial rate for post-acute care rose from 10.9% in 2020 to 22.7% in 2022, and that multiple automation initiatives were implemented during that period. The same report says UnitedHealthcare also explored AI and machine learning to predict which denials were likely to be appealed.

That does not prove every denial is caused by AI, and it would be inaccurate to say that it does. But it does support a more cautious inference: as payer review becomes more automated, the burden shifts to providers to anticipate exactly which data elements, codes, and documentation patterns the payer system expects to see. In practice, that makes denials harder to prevent and appeals harder to win quickly.

Why Medicare Advantage Is the Pressure Point for Physician Practices

If there is one place where the denial problem is most visible, it is Medicare Advantage. KFF’s 2024 analysis found that Medicare Advantage insurers processed nearly 53 million prior authorization requests in 2024, up from 49.8 million in 2023. They denied 4.1 million requests, or 7.7% of the total, and more than 80% of appealed denials were overturned.

That overturn rate is the most revealing number in the set. It suggests that a meaningful share of initial denials is not the end of the clinical discussion. They are the beginning of an administrative one. For physician practices, that means staff are often doing a second round of work just to get a result that should arguably have been reached earlier.

The burden is especially important for practices with large Medicare Advantage panels, because the issue now sits at the point of routine care: referrals, imaging, outpatient procedures, post-acute care, and medication approvals. The financial impact is not just delayed reimbursement. It is also time lost to follow-up, resubmission, peer-to-peer review, and appeals.

The Administrative Load Is Quantifiable and Growing

The AMA’s 2024 Prior Authorization Physician Survey of 1,000 practicing physicians found that the average practice completes 39 prior authorization requests per physician per week, consuming approximately 13 hours of physician and staff time. Forty percent of practices now employ staff dedicated exclusively to prior authorization management.

The same AMA report found that only one in three physicians believes the latest insurer promises will lead to meaningful change. That skepticism is not abstract. It reflects years of promises to simplify prior authorization without a corresponding drop in administrative friction.

The burden is still measurable, still expensive, and still shifting onto provider organizations rather than disappearing.

Why Traditional Appeal Workflows Are Becoming Less Effective

Traditional appeal processes were designed for an environment where human reviewers evaluated claims against written medical policies. Today’s automated review systems introduce a different challenge.

When a denial is triggered by missing documentation elements, coding combinations, authorization requirements, or system-defined criteria, providers often must address both procedural issues and clinical justification simultaneously.

The challenge is compounded by limited visibility into how proprietary review systems evaluate claims. Because providers generally do not have access to the exact criteria used by these algorithms, identifying the root cause of a denial can become more difficult.

As a result, appeal processes may become longer, more resource-intensive, and less predictable.

What Practices Can Do Now

Because regulatory solutions remain uncertain, practices should focus on operational strategies within their control.

• Know Your Payer Performance

Track denial rates, appeal overturn rates, authorization requirements, and reimbursement timelines by payer and plan. Medicare Advantage plans can vary significantly in their administrative burden.

• Move Authorization Upstream

Treat prior authorization as a front-end workflow responsibility. Verification at scheduling and pre-visit stages can prevent avoidable denials later in the revenue cycle.

• Document for Clinical and Administrative Review

Clinical documentation must satisfy both medical necessity requirements and increasingly structured payer review processes. Standardized templates and payer-specific documentation guidance can help reduce avoidable denials.

• Analyze Denial Trends

Monitor denial reason codes by payer, specialty, service line, and provider. Granular reporting helps identify patterns that broad denial metrics often miss.

• Monitor Policy Changes

Authorization requirements, payer policies, and documentation standards change frequently. Establishing a formal process for monitoring and communicating changes can help reduce compliance gaps.

Will Regulatory Pressure Change Things?

In June 2025, sixty health insurers signed a voluntary prior authorization reform pledge that included commitments to streamline authorization processes and require clinician involvement in medical necessity denials. However, physician organizations remain cautious about the potential impact.

CMS transparency requirements are providing greater visibility into insurer denial and appeal patterns, offering practices more data to evaluate payer performance and operational risk.

While regulatory oversight may continue to evolve, most physician practices will likely need to rely on internal process improvements to address denial-related challenges in the near term.

Conclusion

AI-driven payer review systems are reshaping the landscape of claims and authorizations for physician practices. Whether these technologies ultimately improve healthcare administration remains a matter of ongoing debate. What is clear is that denial management can no longer be treated solely as a back-office billing function.

Practices that combine strong documentation standards, proactive authorization workflows, payer-specific performance tracking, and disciplined revenue cycle management processes will be better positioned to protect reimbursement and reduce administrative burden.

As payer AI becomes more sophisticated, physician practices need to respond with equally sophisticated revenue cycle processes. Strong documentation, payer-specific workflows, proactive denial tracking, and experienced revenue cycle management partners can help protect reimbursement in an increasingly automated claims environment.