Shots:
- Did you know that nearly 15–20% of healthcare claims in the U.S. are denied on first submission, creating billions of dollars in administrative costs and forcing providers into complex, time-consuming appeals processes that strain revenue cycles, clinical staff, and patient care continuity?
- Authsnap is an AI healthcare technology company helps providers manage insurance denials and prior authorizations. Its platform analyzes clinical documentation, aligns it with payer criteria, and generates structured appeal documentation to support revenue recovery while reducing administrative workload for healthcare teams.
- PharmaShots welcomes Dr. Wael Khouli, Co-Founder & Chief Medical Officer, Authsnap, for an insightful dialogue on how AI-powered clinical intelligence and human expertise are transforming denial management and improving access to medically necessary care.
Saurabh: Let’s start at the ground level; denied claims! For someone outside revenue cycle, why is the denial and appeals process such a massive pain point in healthcare today?
Wael: The U.S. healthcare system spends an estimated $19.7 billion annually just processing denied claims — and that figure doesn’t account for the revenue that’s never recovered. Denied claims are not administrative inconveniences. They are delayed revenue, delayed care validation, and compounding operational drag.
In the U.S., 15–20% of claims are denied on first submission, with inpatient denial rates climbing year-over-year as payers refine their utilization management criteria. Every denial triggers a manual, labor-intensive process requiring clinical review, payer policy interpretation, documentation assembly, and appeal letter drafting — often on a tight filing deadline.
Most health systems are still managing this through spreadsheets, email threads, and fragmented workflows built for a simpler era. Meanwhile, payer criteria are growing more complex, medical necessity standards shift by region and contract, and the clinical staff expected to navigate all of this are stretched thin.
The burden lands differently across an organization. For CEOs and CFOs, denials represent margin leakage and cash flow volatility in an already narrow-margin environment. For revenue cycle leaders, they create unpredictable backlogs and staffing pressure. For clinicians, they mean hours spent re-justifying care that was already delivered and clinically appropriate. For patients, they introduce financial uncertainty and can delay or interrupt medically necessary care.
The pain point is not just volume. It is the intersection of variability, regulatory complexity, and the fact that effective appeals require clinical intelligence — in addition to clerical labor. That gap between what the process demands and what most teams can realistically provide is where revenue quietly disappears.
Saurabh: Appeals often fall through the cracks even when they’re valid. What usually stops providers from appealing more claims?
Wael: Several forces converge to create what I call the “appeals gap” — the space between claims that deserve to be overturned and claims that actually get fought.
The most visible barrier is capacity. Most hospitals triage denials by dollar value, which means lower-reimbursement claims — even valid ones — are routinely deprioritized. The calculus is straightforward: if appealing a $400 denial takes three hours of clinical staff time, the math doesn’t work. So the claim gets written off. Multiply that across hundreds of denials a month and you begin to understand why 30% or more of denied claims are never appealed — not because they lack merit, but because the system lacks capacity.
Then there’s the deadline problem. Appeals have timely filing windows — often 30 to 90 days depending on the payer and contract. When teams are overwhelmed, cases age. By the time someone gets to a denial, the window may have closed. That’s revenue that’s gone permanently, not just delayed.
But the factor that gets talked about least is cognitive fatigue. Effective appeals require careful, detailed reading of clinical documentation, cross-referencing payer-specific criteria, and constructing a logical, evidence-based argument. That’s cognitively demanding work. By the end of a shift, even experienced nurses and physician advisors are depleted. The quality and consistency of appeals drafted at 4pm on a Friday looks nothing like those drafted fresh on a Monday morning.
The result is a system where recovery depends less on the merit of the claim and more on timing, staffing levels, and who happened to pick it up. That’s not a revenue cycle problem. That’s a structural design flaw — and it’s exactly the gap that intelligent automation is built to close.
Saurabh: Authsnap uses AI to generate appeal documentation. In simple terms, how does your platform “understand” a claim well enough to build a strong appeal?
Wael: We approach this as a structured intelligence problem, not a text generation problem. That distinction matters.
Most AI writing tools take a document and produce a summary. That’s not what appeals require. A successful appeal needs to identify the specific clinical facts that satisfy a payer’s medical necessity criteria, address the stated reason for denial directly, cite relevant clinical guidelines, and frame the argument in language that resonates with a medical director reviewing hundreds of cases. Getting any one of those elements wrong — or missing a documentation gap — can sink an otherwise valid appeal.
Our platform is built around how clinicians actually think. When a physician evaluates a patient, they move through a structured reasoning process — presenting symptoms, differential diagnosis, supporting evidence, clinical guidelines, and a treatment rationale. Authsnap follows that same logic. It reads the clinical record the way a trained physician would, identifying what the documentation establishes, what it implies, and what it proves. That clinical reasoning layer comes first.
From there, the platform translates that reasoning into payer-aligned language. It maps the clinical findings against payer-specific policy criteria, identifies where the documentation supports medical necessity, and surfaces gaps that need to be addressed before submission. The result is an appeal that doesn’t just describe what happened — it makes the clinical case in the terms insurers are required to evaluate.
The phrase “clinician-authored logic” is important to us. It means that before any automation runs, physician advisors and appeals specialists have defined the clinical reasoning framework: what criteria the payer is applying, what evidence is most persuasive, and what documentation patterns are most likely to result in an overturn. The AI executes that framework at scale.
The output reads like it was written by an experienced physician advisor who had time to do it right — because effectively, it was. That consistency is what we mean when we say we reduce manual workload by up to 95%. We are not cutting corners. We are eliminating the repetitive, extractive labor so clinical expertise can be applied where it actually matters.
We do not replace clinical judgment. We operationalize it.
Saurabh: AI in healthcare raises questions of trust. How do you ensure appeal letters are clinically accurate and compliant?
Wael: It’s a fair and important question — and one we take seriously. Healthcare AI that operates without accountability isn’t just a liability risk. It’s a patient safety issue. So we’ve built trust into the architecture of the platform, not as an afterthought.
The first layer is clinical foundation. As I described earlier, our platform reasons through cases the way a clinician would — following established clinical guidelines, evidence-based criteria, and the same diagnostic logic a physician advisor would apply. That framework is authored and validated by our clinical leadership before any automation is deployed. The AI doesn’t invent an argument. It executes a clinically sound one.
The second layer is human oversight. We operate on a human-in-the-loop model, which means AI-generated drafts are reviewed by clinical experts — particularly during onboarding and for high-complexity or high-dollar cases. Our reviewers are not rubber-stamping output. They are actively validating that the clinical reasoning is accurate, the payer criteria are correctly applied, and the documentation supports the argument being made. Over time, that review process also feeds back into the model, continuously improving accuracy and alignment.
The third layer is governance and security. We operate within Microsoft Azure’s enterprise-grade cloud infrastructure — a platform that health system IT and legal teams already know, trust, and have established security frameworks around. Azure’s architecture is fully HIPAA-compliant and satisfies SOC2 requirements, which means our clients don’t have to take our word for it. The compliance foundation is built into the environment we operate in, and we layer our own strict data handling protocols on top of that.
Underlying all three layers is explainability. Revenue cycle leaders and clinical reviewers can see exactly which clinical elements were identified, how they map to payer criteria, and why the appeal was structured the way it was. There are no black boxes. Every output is traceable.
Ultimately, trust is not a feature we added. It is the design philosophy the platform was built around. In a domain where accuracy has direct financial and clinical consequences, anything less is not acceptable.
Saurabh: Your model combines AI with human claims expertise. Where does the human layer make the biggest difference?
Wael: Strategy and nuance — and those two things are harder to automate than most people assume.
Payer behavior is not uniform. The same diagnosis, the same procedure, and the same documentation can produce different outcomes depending on the payer, the regional market, the specific contract terms, and even the medical director reviewing the case. Experienced appeals specialists develop an intuition for that variability over years of practice. They know which payers respond to which arguments, when a peer-to-peer review is worth requesting, and when additional documentation will move the needle versus when it won’t.
That judgment is irreplaceable. And it’s where the human layer makes the biggest difference.
Concretely, human expertise drives three things in our model. First, escalation decisions — identifying which denials warrant a peer-to-peer conversation between our clinical team and the payer’s medical director, rather than a written appeal. That distinction alone can significantly change overturn probability on complex cases. Second, policy interpretation — payer criteria are written in language that requires clinical and contractual expertise to parse correctly. Humans catch nuances that pattern-matching alone would miss. Third, continuous improvement — our clinical team reviews outcomes, identifies what’s working, and refines the appeal logic accordingly. The AI gets better because humans are paying attention.
AI handles what it does best: structured extraction, documentation analysis, policy mapping, and consistent drafting at scale. Humans handle what they do best: judgment, relationships, escalation, and strategic decision-making.
This hybrid model is not a transitional phase until AI matures enough to operate independently. It is the right architecture for a domain where the stakes — financial and clinical — demand accountability at every step. And it is why we have seen recovery rates up to 70% in targeted denial categories. That outcome requires both.
Saurabh: From a user perspective, what does adoption look like?
Wael: One of the most common concerns we hear from health system leaders is: “We already have too many systems. How does this fit into what we have?” It’s the right question, and the answer shapes how we approach every implementation.
Authsnap is designed to integrate into existing workflows rather than replace them — but we also recognize that every organization has different security postures and IT priorities. That’s why we built the platform to operate in both modes. Our default onboarding approach requires no EMR integration at all. Clinical documentation is submitted as a PDF, which the platform ingests, analyzes, and uses to build the appeal. This keeps the barrier to entry low and eliminates any concern about direct system access during early implementation. For organizations that want deeper integration, we support EMR connectivity — but we let clients drive that decision on their timeline.
We offer two deployment models depending on organizational size and maturity. For smaller practices and specialty clinics, we operate as a managed service — our team handles the appeals process end-to-end, with clinical oversight built in. For larger health systems, the platform is embedded directly into existing RCM operations, giving internal teams a powerful tool while preserving their existing governance structures.
What organizations typically experience in the first 90 days is a meaningful shift in operational rhythm. Manual chart review time drops significantly as the platform handles extraction and documentation analysis. Appeal turnaround shortens because drafting is no longer the bottleneck. Fewer cases age past timely filing deadlines because the pipeline is visible and moving. And documentation quality becomes more consistent — not dependent on who happened to pick up the case that day.
Our early clients include specialty providers and clinics managing high-volume, high-denial categories — organizations where the financial and operational pressure was acute enough that the status quo was no longer acceptable. What drew them to Authsnap was not just the technology. It was the combination of clinical credibility, flexible implementation, and a model that met them where they were.
Adoption, done right, is not just a software rollout. It is the beginning of a shift from reactive denial management to structured revenue integrity.
Saurabh: Beyond revenue recovery, what changes do organizations notice?
Wael: Revenue recovery gets the headlines, but the organizations we work with often tell us the operational changes are just as meaningful — sometimes more so.
The most immediate effect is on the people doing the work. Appeals teams carry a particular kind of stress — a constant awareness of backlogs, aging cases, and deadlines that can’t be recovered once missed. When the platform absorbs the drafting burden, that pressure lifts. Clinical reviewers shift from spending their day writing letters to making strategic decisions about escalation, documentation strategy, and payer relationships. That’s a fundamentally different — and more sustainable — way to work. The morale impact is real, and in an environment where experienced revenue cycle staff are difficult to recruit and retain, that matters.
Beyond morale, there is a capacity dividend. Staff who were previously consumed by appeals drafting now have bandwidth for work they simply couldn’t get to before. One of our clients empowers their medical assistants — the same MAs who support physicians in the office day-to-day — to actively participate in the appeals process for the patients they already know and care for. That’s not a workaround. That’s a smart deployment of existing staff in a way that directly benefits the physicians they support and the patients whose care is on the line. It creates a tight feedback loop between the clinical team and the reimbursement process that most practices never achieve.
At the CFO level, the conversation shifts from variance explanation to revenue predictability. Denial write-offs have historically been treated as an accepted cost of doing business — a line item that fluctuates and gets explained after the fact. When denial management becomes systematic, net revenue becomes more forecastable. That changes how finance teams model collections, how boards evaluate revenue cycle performance, and how organizations make staffing and investment decisions.
The effect we find most strategically significant, though, is what happens upstream. As the platform processes denials and identifies patterns — which documentation gaps trigger which payer denials, which clinical language creates ambiguity, which order sets consistently produce medical necessity problems — that intelligence feeds back into clinical workflows. Documentation practices improve at the point of care. Prior authorization requests become more precise. Repeat denials in the same categories start to decline.
That feedback loop transforms denial management from a back-end revenue recovery function into a front-end quality improvement mechanism. It shifts the entire posture of the organization — from reactive firefighting to structured revenue integrity.
Saurabh: Do you see a connection between better appeals management and improved patient access?
Wael: The connection is direct, and it’s one we think about constantly — because it’s easy for this work to get framed purely as a financial problem when the human consequences are just as real.
When a claim is denied and left unappealed, one of two things typically happens. Either the provider absorbs the loss, which erodes the resources available for clinical care, or the bill gets transferred to the patient — often before the appeals process has run its course. That creates immediate financial stress for families who believed their insurance would cover their care. But beyond the bill, denied claims can interrupt ongoing treatment. A patient mid-course on a biologic therapy, a post-surgical rehabilitation plan, or a mental health treatment protocol may face gaps in care simply because the administrative process broke down. That is a clinical harm, not just a financial inconvenience.
Prior authorization denials carry their own access consequences that are even more direct. When a prior auth is denied, the treatment doesn’t just get delayed — it may not happen at all. Patients are turned away at the point of care, sometimes for treatments their physician has determined are medically necessary and time-sensitive. The administrative barrier becomes a clinical barrier.
There is also a subtler, and in some ways more troubling, effect on physician behavior. Physicians are rational actors. When they repeatedly experience denials for a particular service, procedure, or medication, they begin to order it less — not because the clinical indication has changed, but because the administrative friction has become too costly. This is called practice pattern modification, and it represents a quiet erosion of evidence-based care. Patients never know they didn’t receive something their physician might otherwise have recommended. It is one of the most invisible barriers to access in the system.
There is also an equity dimension to this conversation that the industry is beginning to reckon with. Denial rates are not uniform across patient populations. Patients in certain geographic markets, on certain plan types, or receiving care for certain conditions face disproportionately higher denial rates. When providers lack the capacity to appeal systematically, those disparities compound. Efficient appeals management is, in a meaningful sense, a health equity issue.
And at the highest level, there is a sustainability crisis unfolding in American healthcare that denial burden is directly contributing to. Financial pressure on providers and health systems has reached a breaking point in many markets. We have seen a steady rise in hospital closures — disproportionately in rural and underserved communities — where operating margins are thinnest and the administrative burden of managing denials is hardest to absorb. When a rural hospital closes, the access consequences for that community can be devastating and permanent. Reducing the revenue leakage caused by unappealed and underfunded denials is not just a financial optimization. It is part of keeping the care infrastructure intact.
Our mission is simple: every claim deserves a fair review. Denials should not be barriers to medically necessary care. The work we do at Authsnap is revenue cycle work on the surface. But underneath it, it is about making sure that when a physician makes a clinical decision to provide care, that decision is respected — by the payer, by the process, and ultimately by the system.
Saurabh: As you scale, how do you tailor the solution for different provider types?
Wael: The denial ecosystem looks very different depending on where you sit — and a solution that works for a 500-bed academic medical center will not map cleanly onto a specialty neurology practice or a community hospital running on thin margins. We built Authsnap to be adaptable across that spectrum.
Large health systems are primarily focused on high-dollar inpatient denials — complex cases where a single overturn can recover tens of thousands of dollars. At that scale, the priorities are enterprise integration, governance controls, and workflow that fits within existing RCM infrastructure. These organizations often have dedicated appeals teams; what they need is technology that makes those teams dramatically more efficient and consistent.
Mid-sized hospitals occupy a particularly difficult position. They face denial volumes that are too large to manage manually but often lack the staffing depth of larger systems. For them, the conversation is about ROI and capacity — demonstrating that the platform pays for itself quickly while reducing dependence on hard-to-fill clinical roles.
Specialty clinics present a different challenge entirely. The denial burden in specialty care is often concentrated in a few high-stakes categories — biologics, infusion therapies, complex surgical procedures, and prior authorizations for treatments that payers scrutinize heavily. A neurology practice managing patients on high-cost migraine therapies, for example, may find that a significant portion of their revenue is tied up in a small number of denial categories that require sophisticated, clinically nuanced appeals. That’s exactly the environment where Authsnap’s clinical reasoning framework delivers the most impact — because the appeals that matter most are also the ones that require the most expertise to get right.
Our approach across all of these segments is modular. We identify the highest-impact denial categories first, configure payer-specific logic for that organization’s market and contracts, and expand from there. We don’t ask organizations to boil the ocean on day one. We show value quickly, build trust, and scale.
We are also growing through channel partnerships with RCM firms that already serve hospitals and health systems nationwide. That strategy allows us to extend Authsnap’s reach without requiring every client to find us directly. Our partners bring the existing relationships and implementation infrastructure; we bring the clinical intelligence layer that makes their service offering meaningfully stronger.
The technology foundation remains constant across all of these deployments. The workflow configuration adapts to the organization’s operational maturity, size, and specialty mix. That flexibility is not an accident — it is how we built the platform from the beginning, because we knew the market we were entering was not one-size-fits-all.
Saurabh: Looking ahead, will AI help prevent denials before they happen?
Wael: Prevention is the next frontier — and honestly, it’s where the most transformative value lies. Appealing a denied claim is expensive, time-consuming, and never guaranteed. Preventing the denial from happening in the first place eliminates the cost entirely and keeps care on track for the patient from the start. The economics are not even close.
The foundation for prevention is already being built through our appeals work. Every denial we process, every documentation gap we identify, every payer trigger we map — that is intelligence about what the system requires to approve care. Appeals data is predictive data. The patterns that cause denials on the back end are the same patterns that can be intercepted on the front end, at the point of order entry, documentation, or prior authorization.
In practical terms, this means a physician placing an order for a high-cost biologic or a complex surgical procedure would receive a real-time prompt — not a bureaucratic checkbox, but a clinically intelligent signal — indicating what documentation is needed to satisfy medical necessity criteria before the claim is ever submitted. It means prior authorization requests that are built around the specific clinical language and evidence criteria that a given payer requires, rather than generic templates that leave gaps for reviewers to exploit. It means medical necessity scoring that gives clinical and revenue cycle teams a clear view of approval probability before a service is rendered.
Prior authorization is an area where we see particularly significant opportunity — both upstream and downstream. On the front end, we can help providers submit stronger, more complete prior authorization requests from the start, built around the specific clinical language and evidence criteria that each payer requires. That reduces the initial denial rate before it becomes a problem. On the back end, when prior authorizations are denied, our appeals platform is specifically equipped to fight those denials with the same clinical rigor we apply to claim appeals. We are not just an appeals tool that happens to touch prior auth. We see prior authorization improvement — from submission quality through denial reversal — as a core service area.
This vision connects directly to everything we discussed earlier in this conversation. If physicians are modifying their practice patterns because denials have made certain services feel not worth ordering, prevention changes that calculus. If rural hospitals are closing because administrative burden is eroding margins, preventing denials upstream restores revenue that was never at risk in the first place. If patients are experiencing care interruptions because prior authorizations are being denied, real-time documentation support means those authorizations are approved the first time.
We are positioning Authsnap to evolve from reactive appeals automation to predictive denial prevention — and we see that as a natural progression, not a pivot. The clinical intelligence we are building today is the same intelligence that will power prevention tomorrow. The feedback loops are already forming.
The long-term vision is not just faster appeals or even better prevention. It is a healthcare system where the administrative infrastructure is finally intelligent enough to get out of the way of clinical care — where reimbursement is predictable, documentation burden is minimal, and the energy of physicians, nurses, and care teams is directed at patients rather than paperwork.
Healthcare is demanding intelligent infrastructure. Our role is to build it responsibly, clinically, and with patients at the center. That has been true from the beginning, and it will remain true as we scale.
About Dr. Wael Khouli
Dr. Wael Khouli is a physician executive and healthcare innovation leader with deep experience at the intersection of clinical medicine, hospital operations, and health technology. He has served in senior leadership roles, including Chief Medical Officer, Medical Director of Case Management, and quality and medical staff leadership, where he focused on improving care delivery, utilization management, and system performance.
He is also a co-founder and Chief Medical Officer of Authsnap, an AI-enabled healthcare technology company, where he works at the intersection of clinical operations and applied artificial intelligence to address prior authorization and insurance denials and improve patient access to medically necessary care. His work centers on aligning clinical insight with operational strategy and emerging technologies to reduce administrative burden, improve outcomes, and help healthcare organizations adapt to a rapidly evolving, technology-driven environment.
Related Post: Bridging Innovation and Clinical Reality: Shalabh Gupta, CEO of Unicycive Therapeutics, on the Future of Biotech
