AI, Expertise, and the Future of Surgical Revenue Cycle
Author:
PV Shrikanth, Chief Information & Technology Officer, nimble solutions
Quickly browse this article:
AI adoption is outpacing governance → increasing audit risk
Rapid use of AI in coding is triggering payer scrutiny, with billions tied to unsupported or poorly validated claims.
Human validation is required—not optional
Even leading research confirms AI must be paired with expert oversight to ensure accuracy, compliance, and defensibility.
Surgical RCM complexity demands specialized AI
Coding, payer rules, and documentation vary widely by specialty—generic AI models miss critical nuances that impact reimbursement.
AI delivers the most value when embedded across workflows
From transcription to coding to AR prioritization, AI improves speed and accuracy when it enhances—not replaces—human decision-making.
The urgency is real: payers, regulations, and talent are shifting fast
Payer AI, 2027 CMS auditability rules, and workforce changes are forcing organizations to adopt governed, audit-ready AI now.
Artificial Intelligence (AI) has quickly become one of the most talked-about topics in healthcare, and revenue cycle management is no exception.
Nearly every week, I find myself having conversations with clients, partners, and industry peers about AI. The questions are usually similar:
How are you using it?
How much of the revenue cycle can AI automate?
And what does it actually mean for organizations like ours?
The excitement is warranted. At this year’s HIMSS conference, over 80 companies showcased AI solutions for revenue cycle management. Analysts project that the AI-powered RCM market will exceed $150 billion by 2030. And every vendor, from billion-dollar public companies to last month’s startup, is promising that artificial intelligence will transform how providers get paid.
But when it comes to surgical revenue cycle management, the conversation needs more than enthusiasm. It needs honesty.
Because in the rush to adopt AI, the industry is learning an uncomfortable lesson about what happens when technology outpaces governance.
Where AI in Revenue Cycle Is Breaking Down
Earlier this year, Blue Cross Blue Shield published an analysis* that should concern anyone deploying AI in revenue cycle. They examined tens of thousands of claims and found that hospitals using AI-assisted coding without strong governance showed a sharp rise in billing patterns that triggered payer scrutiny, with an estimated cost gap exceeding $2.3 billion. The pattern was clearest at organizations that publicly disclosed AI adoption without demonstrating how that AI was validated.
The takeaway is straightforward: payers are now specifically looking at AI-coded claims. And starting January 2027, new CMS interoperability rules will require standardized APIs that make AI-driven coding and documentation decisions fully auditable.
Meanwhile, McKinsey’s landmark study, “Agentic AI and the Race to a Touchless Revenue Cycle**” on AI in the revenue cycle, published just this January, concluded with a line that should be framed on every RCM leader’s wall:
“A human will always remain in the loop.”
That isn’t nimble marketing. That’s the world’s most respected consulting firm looking at every AI deployment in healthcare and concluding that human expertise is not optional. It’s required.
At nimble, we didn’t need McKinsey to tell us this. It’s been our operating philosophy since day one. But it’s validating to see the industry catching up to what we’ve always believed: AI should enhance expertise, not replace it.
How nimble Approaches AI, Differently
We embed AI across our core service lines: transcription, coding, accounts receivable, and analytics. But here’s what makes our approach different from the vendors making headlines:
Every AI-assisted output passes through expert review and validation by our experienced teams. Not as a checkbox. As the design.
We built our AI systems with a multi-layer architecture specifically because surgical revenue cycle demands it. In coding, for example, our AI operates through four distinct layers:
- First, it generates. The AI reads the clinical documentation and proposes codes (ICD-10, CPT) based on patterns it has learned from hundreds of thousands of charts coded by nimble’s own surgical coding experts.
- It then validates its own reasoning. The system produces a transparent chain-of-thought explanation for every code it proposes. Not just “here’s a code,” but “here’s why, based on this specific finding in the documentation.” Every decision is auditable.
- Finally, it challenges itself. A second validation layer reviews the initial output and flags inconsistencies, questionable assignments, or areas where the documentation doesn’t fully support the proposed code.
Only then does it reach the human expert. The coder reviews the AI’s work, approves or corrects it, and, critically, those corrections feed back into the model to make it smarter over time.
This isn’t AI for the sake of AI. It’s a system designed from the ground up to withstand the exact kind of payer audits that Blue Cross is now conducting. When a claim is reviewed, there’s a defensible, transparent reasoning chain behind every code, with a human expert’s signature on the final decision.
In specialties like ophthalmology, this approach consistently delivers accuracy exceeding 95% on cataract coding, which directly translates to fewer denials and faster payments. Across our revenue cycle operations, our systems process hundreds of thousands of transactions spanning hundreds of facilities.
Why Surgical Revenue Cycle Requires Expertise
Anyone who works in surgical RCM understands that the process is far more nuanced than many people assume. Coding and reimbursement decisions depend on a complex web of factors, which can include:
- Specialty-specific coding requirements that vary between ophthalmology, orthopedics, GI, pain management, and dozens of other surgical disciplines
- Local Coverage Determinations (LCDs) and National Coverage Determinations (NCDs) that differ by geography and payer
- NCCI edits and Medically Unlikely Edits (MUEs) that drive bundling and frequency rules
- Payer policy variations that change quarterly, sometimes monthly
- Physician documentation quality that ranges from highly structured to nearly illegible
- Contract-specific reimbursement rules unique to each client-payer relationship
Even small differences in clinical documentation can significantly impact how a case should be coded or reimbursed. A generic AI model trained on hospital emergency department data doesn’t understand the difference between a cataract extraction with IOL insertion and a complex combined procedure requiring modifier-specific coding logic.
nimble’s AI was built for surgical RCM. Trained on surgical data. Validated by surgical coding experts. And applied to the specialties and payer environments our clients actually operate in.
That specificity matters. Denial rates across the industry have risen from 30% to 41% of providers reporting rates above 10% in just three years. In that environment, the difference between a generic AI suggestion and a specialty-validated AI recommendation isn’t academic. It’s revenue.
Where AI Creates Real Value
Transcription: From Manual Typing to Clinical Validation
Speech recognition technology has improved dramatically, but clinical documentation still requires careful validation to ensure accuracy. Our transcription workflow combines AI-powered speech-to-text capabilities with structured formatting tools designed specifically for surgical specialties.
What this means in practice: instead of a transcriptionist spending their time typing what a physician dictated, the AI produces a structured draft, and the transcriptionist’s role becomes clinical quality validation. They’re checking that the AI captured the medical meaning correctly, not retyping words.
That’s a more valuable role, not a lesser one. And our multi-layer quality assurance process validates every transcript against verbatim clinical standards before it reaches the provider. The result: transcription accuracy above 95%, with cases moving into billing faster, accelerating the overall revenue cycle timeline.
Coding: Catching Issues Before Claims Go Out
Our AI analyzes clinical documentation as it enters the workflow, identifying potential gaps, inconsistencies, or areas where additional detail may be needed to support accurate coding. This proactive approach helps ensure claims are accurate and complete before submission, reducing denial risk and accelerating reimbursement timelines.
But the system does more than flag problems. It proposes solutions, with transparent reasoning, and then waits for the expert to decide. The coder isn’t checking boxes. They’re exercising clinical judgment on an AI-prepared foundation, which means they can handle more volume at higher accuracy than either the human or the AI could achieve alone.
Improving Clinical Documentation
If documentation is missing key elements required for coding or reimbursement, our systems flag those issues early in the process, before they become denials. Our teams work with providers to ensure documentation supports both clinical accuracy and appropriate reimbursement.
The goal isn’t automation. The goal is getting the documentation right the first time.
Revenue Cycle Intelligence: Smarter Prioritization, Better Outcomes
Our analytics capabilities allow us to analyze large volumes of claims and payer activity across hundreds of facilities to surface patterns that individual teams might not see.
These insights identify emerging denial patterns, potential underpayment trends, and claims that should be prioritized for follow-up based on historical success rates and financial impact, not just position in a queue.
When paired with experienced revenue cycle professionals, this intelligence transforms how work gets done. Instead of an AR representative starting a worklist top-to-bottom and hoping they pick the right accounts, the AI tells them: “Start here. This account has the highest probability of recovery and the largest financial impact.”
Same effort. Better results.
What This Means for the People Who Do This Work
I want to address something directly, because I know it’s on people’s minds.
When you hear “AI in revenue cycle,” it’s natural to wonder: does this mean fewer jobs?
Here’s our answer: no. And it’s not just a talking point. It’s a design decision.
nimble’s AI was built to change what people work on, not whether they work.
The coders who learn to work with AI will be the most valuable coders in healthcare. The transcriptionists who validate AI output will be the quality gatekeepers that every surgical practice needs. The AR representatives working with intelligent worklists will collect more per dollar of effort than anyone in the industry.
We believe in a more expert-focused team, not a smaller one.
Protecting Your Data Is Non-Negotiable
Another question I hear frequently when discussing AI is about data protection, and rightfully so.
Many AI platforms in the market rely on shared datasets or third-party model training, which means your clinical and financial data could be used to train models that benefit your competitors. Some vendors route data through offshore processing centers with limited transparency into how that data is handled.
nimble has taken a fundamentally different approach. Our AI operates within our own governed infrastructure with strict security controls and per-client data isolation. Your data trains models for your benefit. It doesn’t leave our environment, and it isn’t shared across clients.
Simply put: your data remains yours. In a market where some competitors are processing claims data through third-party operations in other countries, we believe that matters.
Why This Matters Now
This isn’t a “someday” conversation. Three forces are converging that make AI governance urgent:
- Payer AI is accelerating. Payers have built, and continue to improve, their own AI systems to scrutinize claims, detect patterns, and issue denials. The information asymmetry between payers and providers is widening. Providers need AI that can match payer AI, with the governance to defend every decision.
- Regulation is tightening. CMS interoperability rules taking effect in 2027 will make AI-driven coding and documentation decisions auditable through standardized APIs. Organizations deploying AI today need to ensure their systems can withstand tomorrow’s regulatory scrutiny.
- The talent landscape is shifting. More than 30% of providers now prioritize AI and automation across seven or more revenue cycle use cases, up from four or five just two years ago. The organizations that invest in AI-augmented workflows now will attract and retain the best talent. The ones that don’t will struggle to compete for increasingly scarce revenue cycle expertise.
Technology + Expertise: The nimble Approach
AI will continue to evolve rapidly, in months, not years. But its effectiveness in surgical revenue cycle management will always depend on how thoughtfully it’s applied.
Success still depends on understanding payer behavior, regulatory requirements, documentation standards, and the countless nuances that shape reimbursement outcomes. Technology can accelerate that work. But it cannot replace the expertise required to do it well.
At nimble, our approach rests on four principles:
- Surgical Specialization. Our AI is trained on surgical data, not hospital data repackaged for outpatient. Cataract coding is different from E&M coding. Orthopedic transcription is different from primary care. We built AI for the surgical world because that’s the world our clients operate in.
- Governance First. Every AI output is validated by a human expert. Our multi-layer architecture (generate, validate, critique, human decision) is designed to withstand payer audits, not just pass demonstrations.
- End-to-End Intelligence. AI touches every step of the revenue cycle, from transcription to coding to claims to analytics. One integrated approach. One accountability chain. One partner who owns the outcome.
- Human-Elevated. AI makes our teams more accurate, more efficient, and more focused on the high-value work that drives results. Technology handles the volume. People handle the judgment.
The result is higher quality, faster workflows, and better financial visibility for the surgical organizations we serve.
Because ultimately, the goal isn’t adopting technology for its own sake. The goal is helping our clients maximize the financial performance of their surgical operations, with the confidence that every AI-assisted decision can be explained, defended, and trusted.
And when the right technology works alongside the right expertise, that’s exactly what becomes possible.
PV Shrikanth serves as Chief Information & Technology Officer at nimble solutions, providing enterprise-wide technology leadership across infrastructure, security, AI, and data — with overall accountability for the company’s technology strategy and execution.
References:
* New BCBSA Research Suggests AI in Hospital Billing is Leading to Higher Health Care Costs (BlueCross BlueShield article dated March 5, 2026)
** Agentic AI and the race to a touchless revenue cycle (McKinsey & Company’s Article dated January 9, 2026; Authored by Michael Peterson and Sanjiv Baxi with Chloe Chan and John Mollica.)
Rising Coding Intensity and Its Impact on Health Care Affordability (BlueCross BlueShield Association’s Issue Brief – March 2026; Authored by: David Wennberg, MD; Francois Fressin, PhD; Ryan Everist, MBA, MHI; Elizabeth Horng; Elise Nelson, MS; Sara Clemente, MS)
State of AI trust in 2026: Shifting to the agentic era (McKinsey & Company’s dated March 25, 2026; Authored by: Gabriel Morgan Asaftei, Roger Roberts, Abby Sticha, Cécile Prinsen)
