Health system leaders don’t need another conversation about AI potential. They’ve seen over a hundred demos. They’ve heard the promises. They’ve launched the pilots.
The problem is that too many pilots never become production systems, and even if they do, too many production systems never deliver measurable value. I’ve seen too many health systems shift from innovation to a pile of pilots and leaders asking “whatever happened to [insert initiative here]?”
In today’s environment, AI cannot simply be another line item justified by possibility. It has to be tied to performance. Because health systems aren’t shopping for AI in theory — they’re evaluating it under real revenue pressure as payer friction, denial complexity, staffing constraints, and cash flow volatility reshape the revenue cycle. In recent research conducted among 81 C-suite RCM executives, though 30% of respondents said their health system was investing in AI at scale, 50% said unclear ROI was a barrier preventing AI adoption. “Promising” is no longer enough, if it ever was. It’s time for health systems to adopt AI that works inside real workflows and produces financial impact a CFO or CRO can actually measure.
The AI mandate is here. The ROI case is still missing.
Too often, organizations start with the technology before they have defined the financial or operational problem they expect it to solve. That’s how pilot pileup happens.
A pilot gets launched. Another solution gets added. A new dashboard appears. The innovation roadmap gets longer. But when finance asks what changed after deployment — what revenue was captured, what work was reduced, what accuracy improved, what value was realized — the answer is too often incomplete.
That gap is showing up across healthcare. McKinsey found that while 82% of healthcare leaders expect a positive return from AI, only 45% have reported quantified returns. Gartner has also predicted that, through 2026, organizations will abandon 60% of AI projects that are not supported by AI-ready data.
For CFOs and CROs, that should be the line in the sand. AI shouldn’t be measured by whether it was adopted. It should be measured by whether it changed the business.
Forecasted ROI is not enough
The most important question to ask an AI vendor is simple: show me the realized number.
Not the forecast. Not the projection. Not the model built for the sales process. The actual result from live clients, in real workflows, after implementation.
That question changes the conversation quickly.
If a vendor cannot show realized value, the health system is being asked to take on the risk. If the product does not fit the workflow, the health system absorbs the disruption. If the recommendations are not trusted, the team absorbs the rework. If the ROI does not materialize, finance absorbs the miss.
Doing AI right means starting with a problem that is specific, measurable, and financially material. In revenue cycle, that may mean asking where earned revenue is being lost today, what teams can realistically validate, and how quickly a solution can begin producing value without creating operational drag.
It also means being honest about what good looks like. Good AI does not just identify more — it identifies what matters. It gives teams evidence they can trust. It fits into the way work actually gets done. And it produces outcomes that can be measured after go-live.
High standards are not optional in healthcare AI
At SmarterDx, we will never sell a solution that isn’t health system-tested and vetted. We will keep innovating, testing, and pushing what clinical AI can do next, but we will only sell what works.
That standard matters because healthcare AI is different. A revenue cycle recommendation is not useful simply because a model generated it. It has to be clinically supported — insights that teams can validate, defend, and tie to realized revenue.
That’s why the frontier clinical AI that is behind every SmarterDx solution analyzes the complete patient story. The patient story is not contained in the discharge summary alone. It lives across notes, labs, vitals, orders, imaging, codes, and longitudinal changes. SmarterDx analyzes the full patient record to identify missed, incomplete, or incorrect diagnoses across the revenue cycle.
It is also why the data foundation matters. SmarterDx is trained on 17M+ clinically validated encounters and more than 10,000 unique diagnoses and procedures, with expert-labeled clinical truth embedded directly into the model.
The goal is to surface the signal in the noise: clinically supported findings so teams can spend their time using their clinical expertise instead of manual, repetitive, and often impossible tasks.
That is the difference between AI that promises and AI that creates value.
The only ROI that matters is realized ROI
The standard should be concrete ROI, not estimates or vague promises.
Today, SmarterPrebill delivers an average of $3.5M in annual realized net new revenue per 10,000 DRG discharges for our health system customers. That is a customer average, not a projection.
Just as important, the value has continued to grow. As the clinical AI learns from more data, more expert feedback, and more cases, the return compounds. What started as $1.5M per 10,000 discharges in early 2024 grew to $2.5M in 2025 and $3.5M today.
For a CFO or CRO, that is the kind of AI story that matters. Not because the number is interesting on its own, but because it is realized, repeatable, and tied to a problem health systems already know they have: earned revenue that is missed because the full complexity of care is not always captured before final billing.
Adoption is not the achievement. Impact is.
The next era of healthcare AI will not be won by the vendor with the flashiest demo or the broadest promise. It will be won by the teams that can answer the questions executives are already asking:
- Does it work in our workflows?
- Can our teams trust it?
- How quickly can it create value?
- What does it require from our people?
- What realized ROI have clients actually seen?
Those questions are not barriers to innovation. They’re what make innovation durable.
Healthcare AI has enormous potential. But potential alone doesn’t protect margins, reduce burden, strengthen workflows, or capture the full value of the care already delivered.
Proof does.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.