When AI Becomes a Payer Advantage, Hospitals Cannot Afford to Stay Reactive

For decades, hospital revenue integrity teams have fought payer denials as if they were isolated transactions. A claim gets denied; we appeal, we resubmit, we track outcomes, rinse, wash and repeat. That approach is no longer tenable. 

According to the American Medical Association, in 2023, 11% of all claims were denied by payers, up from 8% in 2021. This increase translates to approximately 110,000 unpaid claims for an average health system, underscoring the financial burden of denials.

Today’s payers are not just automating them. They’re strategically pinpointing which claims to deny based on highly sophisticated analytics that predict provider response or perhaps, ideally, lack thereof. Increasingly, insurers are using AI and advanced analytics not just to automate claims decisions, but to predict provider behavior. The question isn’t always “Is this claim payable?” It’s often, “Will this be worth the provider’s effort to appeal this claim?”

The goal isn’t merely to enforce rules about medical necessity or coding clarity. It is to predict the claims that won’t be appealed or, similarly, where underpayments will go unnoticed. That’s a subtle but profound shift in the reimbursement landscape that many hospitals aren’t prepared to address.

From rules-based denials to behavior-based strategy

Payers have always had access to large volumes of claims data. What’s different now are the tools at their disposal to analyze it. Modern, AI-enabled analytics allow payers to look back across years of claims and see patterns in how hospitals respond.

They know, for example:

• Which denial codes hospitals appeal most (and least).
• Which service lines specifically see low appeals volume due to resource constraints.
• Where underpayments fall below internal review thresholds. 

They can see appeal timelines, overturn rates, and even how often providers do not pursue appeals because the reimbursement dollars don’t justify the work.

Now, when that data is fed into AI-driven decision engines, insurers are able to assign “denial risk scores” that factor in the likelihood of appeal and historical appeal outcomes, not just clinical or other administrative criteria. In effect, some denials are designed to succeed not because they are accurate, but because they are unlikely to be challenged.

A denial that has a low chance of being appealed becomes a low-risk decision, even if it’s questionable. A small underpayment that consistently goes unnoticed becomes an easy win. Multiply those decisions across thousands or millions of claims, and the impact adds up quickly. What makes this especially challenging for hospitals is that it doesn’t always show up as a spike in denials. In many cases, denial rates stay flat. Revenue erosion happens incrementally, spread across claims in ways that are hard to spot without deeper analysis.

Why many denials “work” even when they’re wrong

Most revenue cycle teams triage denial workloads based on timely filing deadlines and expected yield. If the cost in staff time and operating budget to appeal one denial is greater than the likely recovery, it makes sound business sense to let it go. According to Becker’s Hospital Review, the cost per denial worked jumped from roughly $44 in 2022 to more than $57 in 2023 (30%), collectively costing the industry tens of billions.

Payers understand this calculus just as well as providers do. When AI models start to factor in provider capacity and historical appeal patterns, payers can tilt decisions toward cases where the appeal rate is low. It’s not that those denials are always correct. It’s that they are profitable when left unchallenged. In those cases, denials succeed not because they’re valid, but because they’re inconvenient and costly to address. 

From a payer standpoint, it’s operational optimization. From a provider standpoint, it’s revenue loss that often goes unexplained because no single claim looks significant on its own.

Over time, those “not worth it” decisions quietly become baked into expected reimbursement, establishing a new status quo.

What hospitals should be paying attention to

As payers increase their use of AI and automation to issue denials faster, at greater scale, and with increasing complexity, hospitals are falling behind. Overwhelmed by volume, constrained by budgets, and lacking the specialized expertise and technology required to consistently overturn denials, traditional approaches to denials management and tracking performance are no longer enough. 

While automation can improve standardization and speed, studies show many hospitals still struggle with rising denial volumes and the burdens they place on staff and cash flow. According to a recent poll conducted by MGMA, nearly 60% of providers reported year-over-year increases in claim denials, with providers collectively spending upwards of $20 billion annually trying to overturn them.

If hospitals are going to stay ahead, revenue cycle leaders must shift from reaction to strategy. Here are key areas of focus:

• Which denial categories have the lowest appeal rates, and are they actually valid?
• Where do underpayments consistently fall just below review thresholds?
• Which payers show patterns of partial payment or “soft denials” that don’t trigger workflows?
• How often do hospitals choose not to appeal due to effort, not merit?

These questions aren’t about blaming teams. They’re about understanding where payer behavior and provider capacity intersect and where that intersection creates risk.

Using data to level the playing field

The good news is that providers can use analytics to counter these dynamics, too. Leading organizations are starting to analyze denials and underpayments through different lenses. Not just by reason code, but by cost to appeal, probability of recovery, and true financial impact.

Instead of asking, “Why was this denied?” they’re asking: 

• Why does this denial persist?
• What claims are most likely to be underpaid systematically?
• Which denials have the highest ratio of overturn to effort?
• What payer policies correlate with high rates of low-yield denials?

That shift allows hospitals to identify denial patterns that are strategically designed to go unchallenged and bring those issues forward, whether through targeted appeals, contract discussions, or payer negotiations.

When providers can show data-backed patterns rather than isolated examples, conversations with payers change. What once felt like anecdotal frustration becomes measurable, defensible insight.

The revenue you don’t see is the revenue you lose

Hospitals today operate under unprecedented margin pressure. Even as median health system operating margins reached 1.2% in the last quarter of 2025, marking the strongest performance of the year, there remains little room for error. In this environment, the biggest risk isn’t the denial that lands squarely in a work queue, but the one that quietly blends into routine operations and never gets questioned.  

As AI becomes more embedded in payer workflows, revenue loss will increasingly be driven by what goes unnoticed rather than what gets rejected outright. With hospitals investing heavily in their own analytics and automation, now is the right time to broaden the conversation. Not just about speeding up appeals but about understanding payer strategy. Not just about working denials but about seeing the patterns behind them.

Because when decisions are designed to stay below the radar, visibility becomes one of the most valuable tools a provider can have.

Photo: digicomphoto, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.