AI Took the Stage at HFMA. The Buyers Came for Evidence.

Right now, a few thousand healthcare finance leaders are walking the floor at HFMA’s annual conference in National Harbor. If you read the agenda, you would think the entire industry had already automated its revenue cycle. The reality on the ground is quieter. And more interesting.

The boardroom moment nobody planned for

For years, revenue cycle lived in the back office. Claims went out, money came in, and most CFOs only looked twice when cash slipped. That era is over. Black Book’s 2026 trends report put it plainly: RCM has become a board-visible control system. Denials, payer behaviour, prior authorisation, cash predictability — these are executive conversations now, not operational footnotes.

Here is what that shift actually means. The moment something reaches the boardroom, it stops being judged on activity and starts being judged on outcomes.

“We’re working denials” is no longer an answer. “We prevented them” is.

The AI gap

Walk the HFMA floor and every booth promises an AI-powered revenue cycle. Yet most provider organisations have not deployed one. Industry surveys this year show a majority of RCM teams still have not implemented AI or automation at all. A tiny single-digit fraction have it fully integrated.

So we have a strange split. Enthusiasm at the podium. Hesitation in the field. Black Book’s founder framed the buyer mood bluntly ahead of the conference: leaders are arriving with specific problems, not browsing. They want evidence, not slogans.

Why the hesitation is rational

There is a reason providers are cautious. A recent payer-backed study flagged that AI billing tools may be pushing healthcare costs up by billions — faster claim generation, more aggressive coding, an arms race between payer bots and provider bots.

If your AI just submits more claims faster, you have not fixed the revenue cycle. You have accelerated the fight — and the payer’s algorithm denies faster than yours can appeal. This is the trap small and mid-size practices keep falling into. They buy the demo, not the outcome.

What actually wins

The practices we work with do not win on the flashiest model. They win on fundamentals that AI should support, never replace:

• A clean front end. Eligibility, authorisation and benefit checks done right the first time kill denials before they exist.
• Prevention over management. Reworking a claim costs real money. Not generating the denial costs nothing.
• Payer intelligence. Knowing how each payer behaves — which codes they fight, how fast they pay — is worth more than any generic automation.
• Human oversight on every automated decision. Governed AI, not autonomous AI.

None of that fits on a conference banner. All of it shows up in your cash.

The 2030 view

By the end of this decade, every RCM vendor will claim AI. Having it will not be the differentiator. Proving it works will be — with auditable results a CFO can take to a board. The winners will not be the loudest. They will be the ones holding the evidence.

That is the standard we are building toward at Medalyze. Not slogans. Receipts.