Only 27% of health system revenue cycle teams run AI at scale across multiple functions in 2026, according to HFMA's survey of 95 finance leaders published April 23, 2026. Another 53% are still piloting AI in select areas, which means most of the industry has not moved past the test phase. For independent practices and the billing companies that serve them, those numbers matter directly: they set the baseline for what a modern revenue cycle vendor should already be doing.

Three separate 2026 surveys, from HFMA, FinThrive, and Oliver Wyman, put numbers on how fast AI is actually moving through claims, coding, denials, and prior authorization work. The three do not agree on a single adoption rate, because they ask different questions of different audiences. Read together, they show where the real activity sits and where it is still mostly talk.

Where deployment actually stands

HFMA's 95-person survey found 27% of respondents deploying AI at scale across multiple revenue cycle functions, with 53% running pilots in select areas. Workforce readiness lagged behind the tools: only 7% of respondents described their teams as very prepared for the skills AI work demands, while 44% called their teams somewhat prepared.

Vendor relationships showed similar strain. 37% of HFMA's respondents described their vendor relationships as functional but increasingly complex, and 19% called them fragmented and difficult to manage. McKinsey and Company research cited in the same HFMA report put a number on the payoff for getting past that complexity: AI-enabled workflow redesign can cut cost-to-collect by 30% to 60%.

What FinThrive's leaders are funding

FinThrive's third annual Transformative Trends Report, based on a survey of 100 hospital financial and revenue cycle leaders fielded by the research firm Hanover between July 18 and August 8, 2025, found 76% of respondents naming automation their top initiative for 2026. 56% called AI and automation their single most significant investment area.

The report breaks out where that AI spending is landing inside the revenue cycle. 73% of FinThrive's respondents are deploying AI in prior authorization work, 67% in denials and underpayment management, and 60% in clinical documentation and coding. Patient experience ranked as the top organizational goal for 71% of leaders, ahead of increasing revenue at 58% and reducing costs at 36%.

Vendor consolidation and a security push

Both HFMA and FinThrive found a matching move toward fewer, more integrated vendors. In FinThrive's survey, over 70% of leaders expect to reduce their reliance on third-party RCM vendors, and close to 60% plan to consolidate their RCM vendor list within three years.

Cybersecurity is a direct driver of that shift. FinThrive's report counted 471 healthcare hacking incidents reported through December 12, 2025, affecting 42 million Americans, and found that 85% of surveyed leaders changed their strategic RCM technology investments because of cybersecurity or clearinghouse disruptions. Among leaders shifting technology strategy, 33% are moving toward hybrid point solutions paired with platform-ready options, 25% are planning full platform adoption, and 22% are keeping point solutions with backups for critical areas.

The bigger picture from Oliver Wyman

Oliver Wyman's May 2026 survey, covering more than 200 decision-makers and 90 end users at provider organizations, found a wider range: 20% to 40% of organizations report broad or enterprise-wide use of AI-enabled tools, while 63% have integrated some form of AI-powered automation into revenue cycle workflows.

The trend line moved fast. 80% of organizations in Oliver Wyman's survey are now exploring, piloting, or implementing generative AI for revenue cycle work, up 38 percentage points from where the same question stood less than two years earlier. 92% of respondents agreed there are no-regret AI investments worth making now, and 70% to 90% of decision-makers expect to increase AI-enabled RCM spending over the next three years. On performance, some coding applications reached accuracy of 90% or higher in specific clinical domains, and complex-case coding time fell by close to 46% where AI tools were in use.

SurveySampleKey adoption figure
HFMA, April 202695 revenue cycle finance leaders27% run AI at scale; 53% still pilot it
FinThrive, January 2026100 hospital RCM leaders (Hanover fieldwork, Jul-Aug 2025)76% name automation their top 2026 initiative
Oliver Wyman, May 2026200+ decision-makers, 90 end users80% exploring, piloting, or implementing generative AI

What this means for practices and billing companies

These three surveys drew mostly from hospitals and health systems, not independent practices directly. But the billing companies and RCM vendors that serve independent practices largely run on the same underlying platforms these leaders are describing, so the adoption gap applies to them too. A vendor that says it uses AI without specifying which of the 27% or the 53% it falls into is describing a pilot, not a production system.

The practical questions for a practice vetting a billing partner follow directly from the survey data: which specific function runs AI today, denials, prior authorization, or coding, since FinThrive found adoption concentrated unevenly across those three areas. Ask for the cost-to-collect number before and after, since McKinsey's cited 30% to 60% range is a target, not a guarantee. And ask about vendor consolidation plans, since close to 60% of FinThrive's respondents expect to narrow their own vendor list within three years, a shift that could affect which billing platform a practice is tied to.

Sources

Talk to us

Fifteen minutes about your practice and the systems you run.

Book a demo