Two trends crossed paths this winter, and billing teams are caught in between. The first: payers are adjudicating claims with AI, and they are getting fast at it. The second: initial denial rates keep climbing anyway. An HFMA analysis published January 30 puts the 2025 initial denial rate at 11.65 percent through November, up from 11.41 percent in 2024, per Kodiak Solutions data. The same piece quotes a Stanford Health Care revenue cycle leader saying payers are becoming more sophisticated with AI and that denials are coming faster than providers can keep up - sometimes within seconds of submission.
For most of the last decade, a denial arrived days or weeks after the claim. There was time to batch the work, gather charts, and respond in weekly cycles. When denials arrive in seconds, weekly cycles lose by default. By the time a human reads the rejection, the claim has already aged, the timely filing clock keeps running, and the payer's own documentation requirements have moved on.
The speed gap is the problem, not the AI
It is tempting to frame this as an AI story, but the mechanism is simpler. Payers automated adjudication; most provider organizations still staff follow-up manually. A denial that lands in seconds gets read in days, worked in weeks. Every hour of that gap is a claim sitting in AR and a rework cost accruing at $25 or more per claim.
The asymmetry also changes what "keeping up" means. If the payer's system reads the denial, checks the policy, and generates a documentation request instantly, then a provider response assembled from memory and faxed records is competing against a system optimized to reject incomplete first responses. Speed is not just convenience. It determines whether the first response wins or the appeal cycle begins.
The asymmetry shows up concretely in aging buckets. A denial caught on day zero is a correction; the same denial caught on day 20 is a project - records have been archived, staff have rotated off the account, and the payer's appeal window has started closing. Organizations that measure time-to-first-touch on denials routinely find it is measured in days, while the payer's decision was made in milliseconds. Every day between those two numbers is margin given away.
What keeping up actually requires
Matching machine-speed adjudication does not mean automating everything at once. The practical sequence looks like this:
- See the denial the moment it posts. Portal checks should run daily, not when someone gets to it. Most billing teams already know which payers deny fastest; those portals deserve first priority.
- Triage by reversibility. Denials that expire or age badly - prior auth, timely filing, records requests with deadlines - should be worked before static ones.
- Draft first responses the same day. An appeal letter assembled from templates and attached records within 24 hours converts far better than a thorough one assembled in two weeks.
- Track the payer's pattern. If a payer's denial mix shifts toward medical necessity or bundling edits, that is policy change showing up in your queue. Feeding it back to front-end edits prevents the next batch.
Kodiak's own read on the market points the same direction: providers are looking at technology to draft appeals and automate routine tasks so their people can focus on follow-up. When the industry's benchmarking body and the health systems losing to seconds-fast denials converge on the same answer, the direction is settled.
The practical starting point is unglamorous: measure how long a denial sits before a human first touches it. That number, more than denial rate or AR days, tells you whether your operation can survive machine-speed adjudication. Then shrink it - daily portal sweeps, same-day triage, first responses drafted within 24 hours, tracked weekly by payer so the worst offenders name themselves. What gets measured weekly gets managed eventually; AR days at quarter's end is too late to act on.
The takeaway
The 11.65 percent denial rate is not a number your team can negotiate with. It is the market clearing price of manual follow-up in an automated payer environment. Organizations that close the speed gap - same-day visibility, same-day first response - keep the revenue. The rest fund their payers' efficiency gains one denied claim at a time.