I've spent enough time in healthcare revenue cycle management to know what a billing department under pressure looks like. The team is smaller than it was last year. The claim queue is longer. And somewhere in that queue, there are accounts that will never get worked, not because anyone dropped the ball, but because there simply aren't enough hours in the day. That's not a failure of effort. That's a failure of infrastructure.
The numbers confirm what revenue cycle leaders are already living.
According to PMMC's 2025 analysis, 77% of hospital accounts receivable now extends beyond 90 days, a 21% increase from 2024.
The Healthcare Financial Management Association (HFMA) benchmarks healthy AR days at 35 to 45 for commercial payers. Many organizations are sitting well above 55 days, and at that point, it's not a warning sign anymore. That's a systemic breakdown.
And it's getting harder to solve the way we've always tried to solve it, by adding people.
The Workforce Math Is Broken
I understand the instinct to hire. When accounts are aging and cash flow is slipping, the immediate answer feels obvious: get more staff on the phones. But the math doesn't hold up anymore.
A 2025 benchmark survey conducted with Becker's Healthcare found that 100% of the hospital and health system organizations surveyed reported workforce shortages had negatively impacted at least one area of their RCM operations. Every single one. Half reported delayed reimbursements. Nearly half, 47%, saw claim denials rise directly because of inexperienced staff.
Replacing a billing team member costs between $9,000 and $12,000, per MGMA DataDive. Turnover in revenue cycle roles runs close to 20% annually. That means organizations are spending significant money training people who leave before they reach full productivity, while unpursued claims continue to age in the background.
Even a fully staffed team hits a ceiling. A standard AR follow-up queue can hold hundreds of accounts. With two or three people working it, every claim is underprioritized by default. Decisions about which accounts get attention are often made informally, by whoever has bandwidth, based on whatever criteria feels urgent that day. Some claims quietly expire past their filing deadline, and no one notices until it's too late. That's not a personnel problem. That's a prioritization problem.
Denials Are Making It Worse
The AR backlog doesn't just sit there. It actively grows, because denials keep feeding it.
Initial claim denial rates hit 11.81% in 2024, up from 10.2% in prior years, according to Kodiak Solutions' analysis of provider data. Experian Health's 2025 State of Claims survey found that 41% of providers now face denial rates above 10%. Medicare Advantage plans alone saw denial-related issues spike 59% in 2024.
Every denied claim is a clock. Payers set appeal windows. Filing deadlines approach. The longer a claim sits unworked, the lower the probability of collecting on it. An AHA and Syntellis report found that half of the 1,300 hospitals and health systems surveyed were carrying $100 million or more in AR for claims older than six months. That's earned revenue, for services already delivered, being written off not because the documentation was wrong, but because no one got to it in time.
Here's what strikes me about the top denial causes: missing or inaccurate data accounts for 50%, authorization issues for 35%, and incomplete patient registration for 32%, per Experian Health. These aren't clinical judgment calls. They're structured, rule-based errors, exactly the kind that should never reach the denial stage if the right checks are in place upstream.
What AI Is Actually Built For Here
I want to be careful about how I frame AI's role in this conversation, because the hype often overshoots the reality. AI isn't going to replace skilled billing professionals. What it can do is absorb the structural, rule-based workload that currently consumes most of a billing specialist's day, and do it at a scale no human team can match.
That means triaging hundreds of accounts by recovery probability, claim age, payer behavior, and appeal deadline. It means automating payer status inquiries. It means generating appeal letters drawn from historical insurer data and contracted terms. And it means surfacing the right accounts to the right people before the filing window closes.
The American Hospital Association has noted that generative AI can assist AR follow-up with automated outreach and fact-based appeals built from payer policy manuals and historical performance data. Organizations using AI-powered claims review have reported meaningful results: a 22% decrease in prior authorization denials in one California network, an 18% reduction in non-covered service denials, and a 40% productivity increase for coding teams at a New York hospital system, according to HFMA-cited reporting.
HFMA data also shows that practices maintaining AR days below 25 outperform peers in overall cash flow by 22%. That gap doesn't close through effort alone. It closes through better prioritization, faster action, and consistent follow-through on every account in the queue, and that's exactly what well-implemented AI delivers.
The Shift That Has to Happen
I think the deeper problem is how the industry has historically categorized AR management, as a back-office function, important but not strategic. That framing has to change.
Aging receivables are a direct measure of how effectively an organization captures the revenue it has already earned. When 77% of AR sits beyond 90 days, that's not a billing team failing to keep up. That's an organization that hasn't built the infrastructure to close the loop between care delivery and reimbursement.
The question I'd put to every revenue cycle leader reading this isn't "can we afford to invest in intelligent automation?" It's "how much longer can we afford not to?" Net revenue leakage increased 25% across more than 2,300 hospitals analyzed by Kodiak Solutions in 2026. Denial rates are still climbing. AR days are still widening. The cost of inaction now exceeds the cost of transformation.
AI isn't a silver bullet. Getting it right requires clean data, thoughtful workflow design, and a clear understanding of where automation ends and human expertise begins. Complex appeals, payer negotiations, and edge cases that fall outside standard rules still require experienced people making experienced decisions. But the volume problem, the velocity problem, the prioritization problem, those are solvable with the technology that exists right now.
The organizations that close the AR gap over the next few years won't do it by hiring faster than they can retain. They'll do it by building systems that follow up at scale, prioritize intelligently, and reach every account before the clock runs out. The claims are there. The revenue is there. The only question is whether the infrastructure exists to collect it.
Sign in to leave a comment.