Will RCM Vendors Survive the AI Revolution? What Waystar's Big Quarter Actually Tells Us About the Future of the Vendor Landscape

Waystar just posted 22% revenue growth in Q1 2026. Their CEO said AI-powered capabilities drove 40% of new bookings. Then he said the quiet part out loud: they are building toward automating a meaningful portion of the $100 billion in annual RCM labor pool.

That is a remarkable thing to say in an earnings call.

It might also just be excellent marketing. Both things can be true.

I get asked by PE firms regularly to explain where the RCM vendor landscape is headed, which vendors are going to survive, which ones are going to get eaten by Epic, and whether the bolt-on model is dead. These are not abstract questions. There is real money at stake in how you answer them.

I spent the last year writing RCM 2030: Strategy and Survival for Revenue Cycle Leaders and its companion guides trying to answer exactly these questions. What follows is my honest read of where things are going, updated for what Waystar's quarter and this week's news confirm.

Spoiler: the autonomous revenue cycle is not a software problem. It is a readiness problem. And readiness has a very specific human shape.

What does the future of RCM vendor consolidation look like?

When I look out to 2030, I do not see hospitals juggling a dozen bolt-ons to make the revenue cycle work. I see the market converging on revenue intelligence platforms: one environment where EHR, billing, and payer data actually talk to each other, and where AI is not a widget added on top, it is the engine underneath.

Why? Because the reconciliation bill hardens the push for interoperability and price transparency, and Project 2025 doubles down on cutting administrative duplication. Those two forces together make "one patient, one financial record, one set of rules" more than a talking point. They make it the least expensive way to operate. In a margin-thin decade, redundancy dies first.

The vendors that survive this consolidation are not necessarily the biggest ones. They are the ones who can answer one question in dollars and days in A/R: what do you do that Epic cannot do yet?

Because "yet" is doing a lot of work in that sentence.

Will Epic replace RCM bolt-on vendors by 2030?

The short answer is: some of them, yes. Not all of them, and probably not the ones you are most worried about.

Here is how I break it down after 13 years of building products designed to be embedded in Epic's marketplace.

Epic has made real progress on consumer-facing tools. They have rolled out AI assistants inside MyChart. They have Penny, an AI agent for billing and denials. They support two-way text messaging. These are steps in the right direction.

But the consumer side still feels thin compared to what is available on the clinical or staff-facing side. I have not found evidence that Epic allows hospitals to schedule outbound texts by the patient's time zone. That may sound minor. A patient on the West Coast getting a 9 a.m. appointment reminder from a New York hospital at 6 a.m. is not having a consumer-centric experience. The fact that something this basic is not built in tells me it is not a priority.

Most of Epic's high-profile AI announcements are aimed at staff efficiency: clinician note drafting, coding, back-end RCM work. That is exactly what I would do if I were leading product development there. Consumer-facing tools look reactive by comparison.

And on claims editing, Epic will not go where you might think. EHR vendors have editing engines, but those engines are hollow until you feed them rules. Epic, Oracle Cerner, and Meditech are not in the business of maintaining national edit libraries, and they do not want to be. Taking that on would mean carrying the liability if something fires wrong and millions in claims get stuck. That is not a risk they will put on their books.

By 2030, I expect EHRs will still take a "Switzerland" approach. They will keep building the pipes and workflows, but they will not own the rulesets. At most, they will offer tighter integration with clearinghouses or denial data feeds if regulators force the issue. But the core business model stays the same: hospitals will still need outside partners to fuel the editing engines, and EHR vendors will be content to keep it that way.

Which RCM vendor categories are most at risk by 2030?

This is the question PE firms ask me most often, and I give the same answer every time. Ask yourself one question about every vendor in your portfolio or your stack: is this function preventing the problem up front, or cleaning it up after the fact? If it is clean-up, plan to sunset it.

The categories most likely to consolidate or exit by 2030:

Standalone payment portals. These will collapse into integrated digital front doors tied directly to EHRs and patient access. If your patient portal is a bolt-on today, it is a dead vendor walking.

Point-solution estimation tools. Will merge into EHR scheduling and payer-facing APIs as interoperability mandates force real-time benefit data to the surface.

Denial management platforms built for reactive clean-up. As denial prevention moves upstream to scheduling, eligibility, and order entry, the market for downstream denial chase software shrinks significantly. The vendors who pivot to prevention survive. The ones who keep selling workqueue management do not.

Paper statement vendors. Paper becomes an opt-in channel, not the default. The vendors whose entire business model is print-and-mail have a clock on them.

Insurance discovery vendors running batch jobs. They lose relevance as eligibility becomes real-time and API-driven. The value of finding insurance coverage three days after a self-pay patient was billed disappears when you had the information at registration.

Contingency fee collection. In my opinion, contingency fees are a symptom of immature revenue cycles. Hospitals that are still paying them in 2030 will be behind. The smarter move is to build prevention and accountability into vendor contracts now and pivot to flat fees while there is still room to negotiate.

The categories that are more durable:

Cybersecurity and compliance platforms with real-time monitoring. The Change Healthcare attack took down the largest clearinghouse in the country and generated $1 million a day in losses for 60% of affected hospitals. Every API connection required by interoperability mandates is a new doorway. By 2030, CFOs will need continuous monitoring of every payer, vendor, clearinghouse, and payment partner they are connected to. That is not a commodity function.

Integrated financial experience platforms that unify estimation, payments, and assistance workflows. These survive because EHR vendors will not own the consumer-centricity problem. If hospitals want to deliver a world-class patient financial experience, they will either need to push Epic to evolve faster or bolt on external solutions that can meet patients where they are.

Analytics and benchmarking tools that can link financial, clinical, and quality metrics into one view. Especially tools with cross-client pattern recognition and payer intelligence that a single hospital cannot replicate internally. This is where vendors earn a durable seat at the table: data nobody else has.

Fraud and anomaly detection. True anomaly detection needs high computing power and pattern recognition across multiple clients. External vendors with broader datasets will always outperform single-hospital models here.

So what should PE firms make of Waystar's $100 billion claim?

Waystar is making a smart bet. The agentic AI pivot, the Iodine acquisition to extend into mid-cycle clinical intelligence, the framing of themselves as an "autonomous revenue cycle platform" rather than a payment software company: all of it is strategically coherent.

I said in RCM 2030 that the only RCM partners still standing by 2030 would be the ones who could answer for their value in actual dollars and actual days in A/R. Waystar is making exactly that case.

But here is what the earnings call did not say.

The autonomous revenue cycle is not a software problem. It is a readiness problem. Every health system that goes all-in on Waystar's agentic platform still has to answer the same questions. Is the data clean enough for the agent to work with? Do you have the people who can translate AI output into operational workflow changes? Is your denial prevention actually upstream or are you still chasing workqueues and calling it automation?

I have watched hospitals buy the most sophisticated RCM technology on the market and then run it on top of broken processes and dirty data. The technology delivers a fraction of what it promised, and everyone blames the vendor. Usually the vendor is not the problem.

Waystar's platform is only as autonomous as the revenue cycle underneath it. If the intake process is sloppy, the agent grabs what is there. Rush University Medical Center found this out the hard way this year: what looked like an AI hallucination was actually bad data and bad process that nobody had fixed because a human had been quietly working around it for years.

The agent does not work around anything. The agent grabs what is there.

What should CFOs and RCM leaders actually do right now?

Ask these questions of every vendor in your current stack before 2027:

What does this vendor do that Epic cannot do yet, and what is the realistic timeline for "yet" to arrive?

Is this vendor solving a root cause or keeping the peace between two internal silos that should have been fixed years ago?

Can this vendor answer for their value in net collections, first-pass yield, denial prevention rate, and patient-pay conversion? If they cannot, that is your answer.

Does their contract allow you to exit in one year? Jefferson Health's CIO said this week he will not sign anything longer than three years and insists on a one-year out clause. He expects to exercise it soon. That posture should be your posture.

And for PE firms evaluating RCM platforms or health system assets: vendor stack quality is not an IT question. It is a margin question. A health system running six reactive clean-up vendors and calling it an AI strategy is a health system with a margin problem waiting to declare itself. The governance quality, the contract terms, and the data infrastructure underneath the technology stack determine how fast bad decisions compound.

By 2030, the revenue cycle leaders who survive are not going to be the ones who bought the most tools. They are going to be the ones who built the organizational muscle to turn AI output into clean claims, fast adjudication, and cash.

That muscle is human. And the time to build it is now.