McKinsey surveyed 215 US healthcare leaders in 2025 on the state of revenue cycle. The results landed in April 2026 and they're not good. Payers are driving A/R longer, clinical denial write-offs are compounding, cost to collect is up for nearly half of respondents, and AI expectations — once sky-high — are quietly coming back to earth. Here's the full breakdown.
McKinsey has been running this survey since 2023. This is the third installment, and it's the most pointed. The headline framing — "RCM at a strategic turning point" — isn't hype. The data shows a function that has been treated as a back-office cost center for two decades now facing structural headwinds that incremental fixes won't solve.
Expect out-of-pocket share of payments to increase, compounding self-pay collection burden
Report at least 10% of A/R lengthening is attributable to payer-caused delays
Say cost to collect has been rising — across all organization sizes
Plan to change their outsourcing vendor strategy within the next two years
The survey identifies two compounding forces: regulatory and payer. Neither is new, but the combination is more severe than at any point in the past decade.
On the regulatory side, more than 65% of respondents anticipate negative financial impact from H.R.1 alone. Stack on the No Surprises Act, hospital price transparency rules, and the accelerating shift in Medicaid — and the administrative overhead attached to every claim is growing quarter over quarter with no relief in sight. 76% expect self-pay bad debt and uncompensated care to increase. 56% expect A/R days to rise.
On the payer side, the numbers are more specific and more damaging. Clinical denial write-offs averaged 2.63% of net patient service revenue (NPSR) across respondents. That's not a rounding error — at a $500M hospital, that's $13M per year written off to clinical denials alone. And it's not getting better: 50% of respondents expect denial rates to increase.
Average written off to clinical denials — the primary driver of reimbursement deterioration, per McKinsey's 2025 survey. At a $500M health system, that's $13M annually, gone.
The payer escalation data is equally telling. Escalated reimbursement issues took an average of 36 days to resolve — and only 43% of those escalations resulted in a favorable outcome for the care delivery organization. That means your team is spending over a month on disputes that fail more than half the time. The 30% most commonly cited obstacle: communication and coordination gaps with payers.
This is the most important shift in the 2025 data, and it signals a maturity inflection in the market.
AI and advanced technologies are now a priority for 51% of respondents — up from 33% in the 2024 survey. That's a meaningful jump in interest. But the expectation data tells a different story: leaders are getting more realistic about what AI can actually deliver and how fast.
In 2024, roughly half of respondents believed autonomous coding could handle over 35% of outpatient coding volumes. In 2025, only a third believe 30% of outpatient coding will be automated. That's not a loss of faith — it's a calibration. The leaders who've tried to implement AI in RCM workflows are now setting more pragmatic timelines.
Only 8% of 2025 respondents expect very high ROI from automation over five years. 42% expect high ROI at five years. But over ten years? Those numbers jump to 18% and 49%. RCM leaders increasingly believe automation's biggest payoff is a decade out, not 18 months. That changes how you build the business case internally.
The barriers to full automation were assessed across seven dimensions. No single barrier exceeded 50% of responses, but four cleared 40%: financial constraints, interoperability challenges, staff training, and change management. The implication: nobody's failing because the technology doesn't work. They're failing because implementation is operationally hard and organizational change is slow.
The top two functions where organizations are prioritizing AI investment:
Those two categories aren't a coincidence. They're the most direct levers on yield. Denials are where revenue disappears. Coding accuracy is where it gets left on the table. AI that actually moves those numbers is worth buying. AI that doesn't — isn't.
One of the more counterintuitive findings: despite all the AI investment and automation interest, only 6% of respondents plan to reduce outsourcing because of technology. The rest are maintaining or expanding it.
60% of respondents expect to change their outsourcing approach over the next two years. Of that group, three-quarters plan to expand outsourcing — not pull it in-house. The functions they're outsourcing to vendors are the exact same ones they're trying to automate: denials, A/R follow-up, coding, and eligibility/authorization.
The market interpretation: AI and outsourcing aren't substitutes. The winning model is integrated platforms that do both — vendors that layer automation onto the workflows they already own. That's why the Ensemble deal, the Procode roll-up, the Candid Health raise, and a dozen other consolidation plays are happening simultaneously. Scale platforms that combine service capacity with AI are the product the market is buying.
Organizations that exceeded their RCM goals in the prior 12 months preferred expanding with existing vendors (59% vs. 41% for new vendors). Underperformers did the opposite — 68% preferred switching to new vendors. The lesson: if you're performing, stay and deepen. If you're switching vendors every cycle hoping for improvement, you're probably the common denominator.
McKinsey's most actionable finding comes from how organizations structure denial prevention. The performance gap is stark:
A 10-point swing in appeal success rate from process standardization alone. That's not a technology purchase — that's a governance decision. Most organizations are still fighting denials downstream at the appeal stage when the fix belongs upstream: medical necessity confirmation before service delivery, clinical validation embedded into documentation workflows, standardized submission processes that reduce clean-claim errors at origin.
The survey is direct: "Many organizations still underinvest in denial prevention and instead rely primarily on downstream appeal management." That's the most important sentence in the report for any operator reading it. If your denial strategy is primarily appeals-based, you're paying for failure twice.
Strip the consulting language and the core message is four things:
Run your own denial write-off as a percentage of NPSR. If it's above the 2.63% McKinsey benchmark, you have a documented gap versus your peer set. That's your starting point for a CFO conversation about denial prevention investment — you're not speculating, you're citing a published number.
McKinsey's third annual RCM survey is the clearest data yet that the era of treating revenue cycle as a back-office function is over. The organizations still managing RCM reactively — fighting denials after the fact, evaluating AI on 18-month promises, cycling through vendors without fixing upstream processes — are falling behind the benchmark in every metric that matters.
The survey isn't pessimistic. It's specific. And specific problems have specific solutions. The question is whether RCM leaders have the organizational capital to implement them before the margin compression makes the choices for them.
Daily coverage of payor policy, denials, and vendor moves — written for people who live in revenue cycle.