Why Healthcare’s AI Arms Race May Be Solving the Wrong Problem—and What Independent Physicians Can Do About It
“The decisions we make today shape the future of medicine.” — John Whyte, MD, MPH, Chief Executive Officer, American Medical Association
The Patient Was Treated. The Paperwork Needed Resuscitation.
Imagine this.
You finish a demanding clinic day.
Twenty-seven patients. Three urgent calls. Two unexpected walk-ins. One patient who has already consulted Google and would like a second opinion from you.
You listen, examine, diagnose, treat, prescribe, document, and finally go home.
Medicine has done its job.
Three weeks later, your billing team delivers the news.
Claim denied.
Was the diagnosis wrong?
No.
Was the treatment inappropriate?
Not necessarily.
Did the physician forget how to practice medicine?
Fortunately, no.
The problem?
A missing modifier. An insurance eligibility mismatch. An expired authorization. Or documentation that didn’t satisfy a particular payer requirement.
The patient received care.
The physician performed the work.
The insurer received the claim.
And somehow, the practice received homework.
Welcome to American healthcare, where completing the medical work doesn’t necessarily mean you’ve finished explaining why you deserve to be paid.
This is an illustrative scenario, not a specific patient case. But the frustration will sound familiar to physicians and clinic owners.
Now comes the remarkable part.
Instead of asking why so many billing problems keep happening, the industry is investing heavily in artificial intelligence to fight over them.
AI to predict denials.
AI to prepare appeals.
AI to review claims.
AI to challenge payer decisions.
Soon, we may have increasingly sophisticated algorithms arguing over a missing modifier while the physician who performed the service is still waiting for reimbursement.
We have entered the age of artificial intelligence. Apparently, administrative common sense is still awaiting authorization.
Here’s the question that should make every physician, practice owner, payer executive, and healthcare technology investor uncomfortable:
What if healthcare’s biggest billing innovation isn’t another AI system that fixes mistakes—but a system that prevents avoidable mistakes from happening?
That question may be more important than the next software demonstration.
Because the real opportunity may not be making the billing war faster.
It may be reducing the number of unnecessary battles.
Healthcare’s Most Expensive Habit: Fixing What Shouldn’t Have Broken
Healthcare has become extraordinarily sophisticated at treating disease.
We use advanced imaging to detect subtle abnormalities.
We sequence genomes.
We perform robotic surgery.
We develop targeted cancer therapies.
We use AI to assist with clinical interpretation and documentation.
Yet an incorrect insurance identifier can still delay reimbursement for a legitimate medical service.
We have built a healthcare system capable of extraordinary clinical precision and remarkable administrative confusion.
Some of that complexity is unavoidable.
Healthcare coverage involves contracts, medical necessity, eligibility, benefits, fraud prevention, and regulatory requirements.
Payers have legitimate responsibilities.
Providers have legitimate obligations.
But complexity becomes a problem when necessary controls turn into repetitive administrative work that adds little value.
Consider the modern medical billing ecosystem.
We have companies that submit claims.
Companies that validate claims.
Companies that correct claims.
Companies that appeal claims.
Companies that analyze denials.
Companies that help recover unpaid reimbursement.
And now companies that use AI to automate these activities.
Many provide genuine value.
But the larger system deserves scrutiny.
When an industry becomes exceptionally good at repairing a recurring problem, we should ask whether enough attention is being paid to preventing it.
This isn’t an accusation that every participant benefits from inefficiency.
It’s a question about incentives and system design.
A complicated system can produce waste even when everyone involved is trying to do their job.
And technology can preserve that waste when it automates existing processes without questioning whether those processes are necessary.
The first opportunity for innovation may therefore be surprisingly unglamorous.
Get the information right before the claim goes out.
That sounds simple.
In practice, it requires rethinking how registration, eligibility, authorization, documentation, coding, charge capture, and claim submission work together.
The Real Diagnosis: Healthcare Has an Information Problem Wearing a Billing Costume
Physicians are trained to distinguish symptoms from causes.
A fever is not a diagnosis.
Chest pain is not a diagnosis.
An abnormal laboratory result is not a diagnosis.
We investigate before deciding how to treat.
Yet healthcare organizations sometimes approach billing differently.
The claim was denied?
Hire another denial specialist.
Collections are slow?
Purchase another dashboard.
Authorization is overwhelming?
Add another administrative employee.
Claims are repeatedly rejected?
Install another automation tool.
These interventions may help.
But they may also treat symptoms without correcting the underlying problem.
A denied claim is often the financial symptom of an earlier information failure.
The error may begin when a patient registers.
It may occur when insurance eligibility is checked.
It may involve an authorization that was never obtained.
It may result from incomplete documentation or incorrect coding.
Or it may arise because a payer applied a policy incorrectly.
These causes are different. They require different responses.
The important insight is that the problem often begins long before the denial appears.
Imagine a patient whose insurance card carries the branding of a familiar national company.
The front desk assumes the patient belongs to one commercial plan.
But the patient is enrolled in a different product administered by a related organization.
The claim goes to the wrong destination.
Weeks later, the practice discovers the problem.
Staff correct the information and begin the process again.
The medical service hasn’t changed.
The patient hasn’t changed.
Only the administrative representation was wrong.
A two-minute mistake has become a two-week investigation.
This is why I believe medical billing must be understood as an information-quality problem with financial consequences.
That doesn’t mean every denial can be prevented.
Some coverage disagreements are legitimate. Some payer decisions are incorrect. Some claims require clinical review.
But preventable errors deserve a different strategy.
Instead of waiting for failure, identify the conditions that produce it.
The cheapest denial to manage is the preventable denial that never happens.
The Statistics: The Numbers Are Serious. The Absurdity Is Almost Comical.
According to the American Medical Association’s 2025 Prior Authorization Physician Survey, released in 2026, the administrative burden on physicians remains substantial.
Consider this.
Physicians report completing an average of 40 prior authorization requests every week, consuming approximately 13 hours of combined physician and staff time.
That amounts to 676 hours annually—nearly 17 conventional 40-hour workweeks.
Not 17 weeks of physician labor alone, but a considerable amount of combined clinical and administrative capacity.
Even more concerning, 95% of physicians reported that prior authorization delays access to necessary patient care, while 94% said it contributes to physician burnout.
The survey also found that 79% reported patients abandoning recommended treatment because of authorization challenges, while 26% reported a serious adverse event associated with prior authorization involving a patient in their care.
These are physician-reported survey findings, not independently verified causal rates for every authorization request.
Nevertheless, they raise an important question.
We spend years training physicians to make complex medical decisions.
Then we create administrative processes that require them to repeatedly justify those decisions.
Medical school teaches physicians how to treat patients. The insurance system teaches them how to ask permission.
Payers reasonably argue that prior authorization can help manage utilization and ensure appropriate care.
That purpose matters.
But so do proportionality, transparency, clinical relevance, and timeliness.
A process intended to reduce unnecessary spending should also be evaluated for the costs it creates.
Staff time.
Delayed treatment.
Additional visits.
Repeated documentation.
Interrupted workflows.
And the cost of managing the process itself.
Here’s the uncomfortable question:
Are we using AI to eliminate unnecessary bureaucracy—or simply teaching bureaucracy to run at machine speed?
Payer Bots Versus Provider Bots: The World’s Most Expensive Argument
The medical billing industry is entering an AI arms race.
Payers are increasingly using automated systems to evaluate claims, identify irregularities, and apply coverage rules.
Providers are adopting AI to predict denials, review documentation, assist with coding, prepare appeals, and communicate with insurers.
Some of these technologies can reduce genuine administrative burden.
But consider where the competition could lead.
A payer’s AI denies a claim.
A provider’s AI prepares an appeal.
The payer’s AI reviews the appeal.
The provider’s AI drafts a rebuttal.
The payer requests more information.
The provider’s AI responds.
Eventually, two sophisticated systems may spend substantial computing resources arguing about a service performed months earlier.
Congratulations. We’ve automated the argument.
Have we improved healthcare?
Possibly.
Have we removed the underlying complexity?
Not necessarily.
This is the distinction that matters.
Automation makes a process faster. Innovation makes a process better.
The two are not synonymous.
An automated appeals system can create real value when a payer makes an incorrect decision.
A predictive claims tool can prevent avoidable submission errors.
An authorization assistant can reduce staff workload.
But if the underlying process remains fragmented, AI may simply help everyone navigate the fragmentation more efficiently.
There is a difference between building a faster ambulance and preventing the accident.
Healthcare needs both.
But we shouldn’t confuse one with the other.
The ultimate victory in the billing war isn’t defeating the other side’s algorithm. It’s making unnecessary conflict less common.
Three Expert Perspectives: What Medicine, Leadership, and Policy Can Teach Us
Medical billing is too complicated to solve from a single perspective.
Physicians understand clinical judgment.
Practice administrators understand operational reality.
Policymakers understand regulatory requirements.
Technology developers understand automation.
The challenge is bringing these perspectives together.
Expert 1: Atul Gawande, MD — Stop Relying on Heroes. Build Better Systems.
Surgeon and author Atul Gawande, MD, has written extensively about systems thinking, checklists, and reducing preventable failures.
His work offers a lesson that extends beyond the operating room.
Reliable performance shouldn’t depend entirely on exceptional individuals.
Every practice has someone who knows the billing system unusually well.
The veteran biller who remembers obscure payer rules.
The office manager who knows which insurer representative to contact.
The front-desk employee who recognizes problematic insurance plans immediately.
These people are invaluable.
But what happens when they retire?
What happens when they’re sick?
What happens when the practice expands?
If your billing system collapses when your best employee takes a vacation, you don’t have a reliable system. You have institutional knowledge with vacation restrictions.
The lesson isn’t to replace experienced staff.
It’s to preserve their expertise in repeatable workflows.
Document recurring rules.
Create verification checkpoints.
Maintain payer-specific guidance.
Establish clear exception procedures.
Expert insight: The best system doesn’t eliminate expertise. It makes expertise more dependable.
Expert 2: Bobby Mukkamala, MD — Administrative Promises Need Measurable Results.
AMA physician leader Bobby Mukkamala, MD, has emphasized the need for meaningful prior authorization reform.
His public commentary reflects continuing physician skepticism about insurer promises to simplify administrative requirements.
That skepticism raises an important question.
What does improvement actually look like?
An insurer announces a new electronic authorization portal.
A software vendor announces an AI agent.
A healthcare organization announces a digital transformation initiative.
Wonderful.
But did the patient receive care sooner?
Did the physician spend less time on paperwork?
Did the practice experience fewer avoidable denials?
Did administrative costs decline?
If those outcomes remain unknown, the announcement isn’t yet proof of success.
A digital form is still a form. An automated delay is still a delay.
Physicians should demand measurable evidence.
Show me the time saved.
Show me the errors prevented.
Show me the improvement in patient access.
Show me the actual effect on practice operations.
Expert insight: Judge administrative innovation by the burden it removes, not the technology it advertises.
Expert 3: John Whyte, MD, MPH — Physicians Must Help Shape the Future.
In its November 6, 2026 advocacy update, the American Medical Association highlighted physician-led care, prior authorization reform, AI governance, and practice efficiency.
AMA CEO John Whyte, MD, MPH, emphasized the importance of physicians helping shape healthcare’s future rather than merely reacting to change.
That principle is especially relevant to AI-powered medical billing.
Technology companies can build sophisticated software.
Payers can establish administrative requirements.
Regulators can define standards.
But physicians and practice staff understand how those systems affect real clinical workflows.
Their participation is essential.
A technology that looks impressive in a demonstration may create new burdens inside a busy clinic.
A system designed without physician input may overlook the distinction between clinical necessity and administrative convenience.
Expert insight: Physicians should be participants in designing healthcare technology, not merely the people expected to tolerate its consequences.
These expert perspectives are interpretations of published work and statements, not new interviews or endorsements of OnnX.
Together, they point toward one conclusion.
Better systems. Measurable outcomes. Physician-led innovation.
Recent News: The Paperwork War Is Entering a New Phase
The discussion about AI and medical billing is changing.
The industry is moving beyond promises of automation toward questions about implementation, accountability, interoperability, and actual outcomes.
Three developments are especially relevant this week.
November 3, 2026: Humana and athenahealth Expand Electronic Prior Authorization
Humana and athenahealth announced an expansion of electronic prior authorization capabilities across athenahealth’s nationwide network of more than 180,000 physicians and clinicians.
The initiative aims to bring more authorization functions directly into existing clinical workflows.
The companies describe the expansion as a way to reduce manual steps, simplify communication, and help patients access care more quickly.
This is an important development.
But the distinction between an announcement and demonstrated outcomes matters.
The expansion’s actual impact should be measured through authorization turnaround times, administrative workload, clinical delays, and user experience.
Putting prior authorization inside the EHR is progress only if it makes the physician’s work easier rather than moving the paperwork to another screen.
November 6, 2026: The AMA Highlights Physician-Led AI and Administrative Reform
The AMA’s newly published advocacy update identifies prior authorization reform, physician well-being, patient access, and the transformation of medicine through AI as major priorities.
This matters because the future of AI in healthcare isn’t solely a technical question.
It’s also a question of governance.
Who designs the workflow?
Who evaluates the results?
Who remains accountable?
And whose interests are prioritized?
If physicians aren’t involved in designing the future of healthcare technology, they may spend the next decade working around someone else’s assumptions.
The 2027 Regulatory Milestone: Electronic Authorization Requirements
The CMS Interoperability and Prior Authorization Final Rule establishes new requirements for certain affected payers.
Operational provisions generally began in 2026, while major API requirements begin in 2027.
For applicable non-drug services, certain affected payers must generally issue decisions within 72 hours for expedited requests and seven calendar days for standard requests, subject to the rule’s scope and exceptions.
The rule also advances electronic exchange of authorization requests, responses, and relevant documentation requirements.
This could reduce administrative fragmentation.
But an API cannot guarantee that the information exchanged is accurate.
A digital highway can move information faster. It can’t guarantee the information is worth transporting.
That is why interoperability and data quality must advance together.
Ten Practical Ways to Stop Feeding the Denial Machine
Physicians don’t need another lecture about working harder.
Most are already doing that.
What they need are practical ways to make administrative systems less demanding.
1. Treat Registration as Revenue Protection
The revenue cycle begins before the physician enters the examination room.
Verify patient identity, payer information, active coverage, and relevant benefits whenever feasible.
Don’t assume that a familiar insurance company logo identifies the correct plan or claims destination.
The first line of revenue protection may be the front desk, not the billing department.
2. Stop Treating Authorization Like a Surprise Birthday Party
Identify authorization requirements before services whenever clinically appropriate.
Track request status, authorization numbers, effective dates, approved services, and escalation deadlines.
Surprises are delightful at birthday parties.
They’re considerably less delightful when they involve an unpaid MRI.
3. Document Clearly, Not Endlessly
More documentation doesn’t automatically mean better documentation.
The objective is accurate, clinically appropriate information that supports the service performed.
AI documentation tools may help, but physicians should review and authenticate records as required.
A longer note isn’t necessarily a better note. Sometimes it’s just a bigger haystack hiding the same needle.
4. Validate Claims Before Submission
Check required identifiers, coding relationships, applicable modifiers, authorization information, and payer-specific requirements.
Catch preventable errors before they become denials.
Don’t make the payer your practice’s first quality-control checkpoint.
5. Separate Rejections From Denials
A claim rejected during initial validation is different from a claim denied after payer processing.
The causes and corrective actions may differ.
Treating every problem as an identical insurance issue makes root-cause analysis harder.
6. Investigate Recurring Errors
If the same authorization problem appears every month, ask why.
If eligibility errors cluster around particular plans, investigate registration.
If documentation problems repeat, review the clinical workflow.
A recurring error isn’t merely an incident. It’s feedback from a system that hasn’t learned.
7. Measure Staff Rework
Track how much time employees spend correcting avoidable mistakes.
Use the information to improve processes, not punish individuals.
The objective is to eliminate unnecessary work.
8. Review Accounts Receivable Intelligently
Monitor aging balances, filing deadlines, payer response times, and appeal requirements.
Older claims may become harder to collect, but there is no universal 90-day rule determining collectibility.
Prioritize based on claim circumstances, value, deadlines, and likelihood of recovery.
9. Test AI Before Expanding It
Start with one measurable workflow.
Establish baseline performance.
Introduce the technology.
Compare results.
Expand only when evidence supports expansion.
Don’t automate everything simply because you can. That’s how a small problem becomes a scalable problem.
10. Preserve Institutional Knowledge
Experienced billers and administrators understand recurring payer requirements, exceptions, and operational workarounds.
Document that knowledge appropriately.
Make it accessible to authorized staff.
Incorporate it into reliable workflows.
Your best employee’s memory shouldn’t be your practice’s only backup system.
The 90-Day Challenge: Diagnose Before You Automate
Suppose your practice commits to improving one part of its revenue cycle over 90 days.
Not replacing every application.
Not purchasing the most expensive AI platform.
Just identifying and reducing one recurring problem.
Days 1–15: Establish the Baseline
Collect recent information on claim rejection rates, initial denials, authorization delays, accounts receivable, and staff rework.
Use consistent definitions.
Identify which problems occur most frequently and create the greatest burden.
Days 16–30: Follow the Error Backward
Select a representative sample of problematic claims.
Trace each through registration, authorization, documentation, coding, and submission.
Determine where the failure first occurred.
Separate preventable errors from legitimate coverage disputes.
Days 31–60: Repair One Workflow
Choose one high-impact problem.
Introduce a targeted improvement.
It may involve eligibility verification, authorization tracking, documentation standards, or pre-submission validation.
Don’t redesign everything simultaneously.
Days 61–75: Measure the Difference
Compare results with the baseline.
Did errors decline?
Did staff time improve?
Did reimbursement become more predictable?
Were new problems introduced?
Days 76–90: Preserve What Works
Document successful changes.
Revise unsuccessful ones.
Train staff.
Assign responsibility for continued monitoring.
Then select the next problem.
The goal isn’t a perfect practice in 90 days. It’s a practice that stops making the same preventable mistake indefinitely.
The Metrics That Matter: Stop Celebrating the Rescue Without Investigating the Fire
Imagine a medical billing company announces:
“We recovered $500,000 in denied claims.”
Impressive.
Recovering legitimate reimbursement matters.
But I would ask another question.
Why were those claims denied?
If the payer made incorrect decisions, successful appeals are valuable.
If many denials resulted from avoidable administrative mistakes, recovery tells only part of the story.
Imagine a fire department celebrating the number of fires it extinguishes without investigating why the same buildings keep catching fire.
That would be a strange approach to prevention.
Yet healthcare sometimes celebrates dollars recovered without giving equal attention to errors prevented.
Money recovered is visible. Mistakes prevented are invisible. Both deserve measurement.
For independent practices, several metrics are particularly useful.
First-pass acceptance rate measures how often claims are accepted for processing without correction.
Initial denial rate measures how frequently processed claims are denied.
Preventable denial rate identifies denials attributable to controllable administrative failures.
Days in accounts receivable helps assess how long reimbursement remains outstanding.
Authorization turnaround time reveals delays in administrative decisions.
Staff rework hours capture the labor spent correcting avoidable errors.
Net collection rate measures how effectively legitimate reimbursement is collected.
These metrics should be evaluated together.
A practice might reduce denials by delaying claim submission.
It might improve collections while increasing overtime.
It might lower billing expenses while introducing new compliance risks.
The goal is balanced improvement.
A dashboard that only counts failures is a scoreboard. A dashboard that explains failures becomes a management tool.
What Does a Small Error Really Cost?
Consider a hypothetical independent practice submitting 2,000 claims monthly.
Suppose 10% require preventable rework.
That’s 200 claims needing additional attention.
At 20 minutes per claim, staff spend approximately 67 hours monthly correcting errors.
At an illustrative fully loaded labor cost of $35 per hour, that’s roughly $2,333 monthly, or $28,000 annually.
These are hypothetical assumptions, not national averages.
They exclude delayed cash flow, external vendor costs, physician interruptions, and potential lost reimbursement.
But they demonstrate an important principle.
A small mistake repeated often enough becomes an expensive operating habit.
Seven Pitfalls: How to Make Medical Billing Worse With Better Technology
Pitfall 1: Automating Bad Information
AI can’t reliably solve a problem when the inputs are inaccurate.
Garbage in, garbage out was the old warning. AI introduces a new possibility: garbage in, beautifully formatted garbage out.
Pitfall 2: Confusing Speed With Improvement
Processing 10,000 claims quickly is impressive.
Processing 10,000 incorrect claims quickly isn’t.
Measure accuracy, downstream consequences, and staff rework—not just transaction volume.
Pitfall 3: Assuming Every Denial Is Unjustified
Some denials arise from legitimate coverage restrictions or documentation deficiencies.
Others reflect payer errors.
A responsible system must distinguish among them.
Pitfall 4: Replacing One Middleman With Five Subscriptions
A practice may eliminate an outside billing service only to purchase multiple disconnected software products.
Each introduces contracts, interfaces, expenses, and operational dependencies.
Simplification should be measured by total burden, not vendor count alone.
Pitfall 5: Ignoring Experienced Staff
Front-desk employees, billers, coders, and administrators often understand the system’s weaknesses better than anyone.
Include them in technology selection and workflow redesign.
Pitfall 6: Treating AI as a Compliance Shield
AI doesn’t eliminate legal obligations.
Unsupported coding remains unsupported.
Inaccurate documentation remains inaccurate.
Protected patient information still requires safeguards.
Pitfall 7: Buying Technology Before Defining the Problem
A vendor promises a 40% reduction in denials.
Ask compared with what, over which period, across which claims, and under what conditions.
An impressive percentage without a meaningful denominator is marketing, not measurement.
Myth Busters: Seven Beliefs Worth Retiring
Myth 1: More AI automatically means better billing.
Reality: AI can improve workflows, but it can also amplify errors and introduce new risks.
Myth 2: A denied claim must have been billed incorrectly.
Reality: Denials may reflect coverage restrictions, payer errors, missing information, or medical necessity disputes.
Myth 3: Outsourcing billing eliminates physician responsibility.
Reality: Outsourcing operational tasks doesn’t automatically eliminate compliance and oversight obligations.
Myth 4: More documentation always improves reimbursement.
Reality: Accurate, relevant documentation matters more than unnecessary volume.
Myth 5: Experienced billers will become obsolete.
Reality: Automation may reduce repetitive work while increasing the value of expertise in complex exceptions, compliance, and oversight.
Myth 6: A clean claim guarantees payment.
Reality: A claim can pass initial validation and still be denied after adjudication.
Myth 7: The smartest denial bot will win the future.
Reality: The greater opportunity may lie in reducing preventable denials before they occur.
The best denial-management system may eventually be the one that has less denial management to perform.
Legal Implications: The Algorithm Made the Mistake. Guess Who Still Has the Problem?
AI can recommend codes, organize records, draft appeals, and identify potential errors.
But it doesn’t automatically assume the legal responsibilities of the physician or medical practice.
You can outsource a task. You can’t automatically outsource accountability.
HIPAA and Patient Privacy
AI billing systems may process protected health information.
Practices must evaluate applicable HIPAA requirements, security safeguards, permitted data uses, access controls, and business associate agreements.
A vendor’s security promise is not a substitute for appropriate due diligence.
False Claims Act Considerations
AI-assisted billing must not generate unsupported charges, fabricate services, or systematically recommend inappropriate codes.
Federal and state fraud-and-abuse requirements remain relevant.
An incorrect claim doesn’t become correct because a computer submitted it.
Documentation Integrity
AI-generated notes should accurately reflect the clinical encounter.
Physicians should review and authenticate documentation as required.
Technology should not invent medical necessity or alter clinical facts to satisfy reimbursement requirements.
Data Ownership and Vendor Contracts
Independent practices should examine data portability, retention, access rights, subcontractor arrangements, auditability, security obligations, and termination provisions.
A clinic shouldn’t have to surrender control of its information to obtain administrative efficiency.
Physician independence is difficult to preserve when the practice can’t control its own data.
Ethical Considerations: When the Algorithm Loves Money More Than Medicine
Every medical claim represents something more important than a financial transaction.
A patient received care.
A physician made a clinical decision.
A healthcare organization committed resources.
That makes billing fundamentally different from an ordinary commercial process.
Imagine an AI system designed exclusively to maximize reimbursement.
It might recommend aggressive coding.
Now imagine a payer AI designed exclusively to minimize expenditures.
It might encourage excessive scrutiny or inappropriate denials.
Both systems could perform well against their narrow financial objectives.
Neither would necessarily serve patients.
An algorithm can optimize a number without understanding why the number matters.
Responsible healthcare AI should emphasize accuracy, fairness, transparency, privacy, clinical integrity, and human oversight.
Financial efficiency matters.
But it can’t be the only objective.
The appropriate question isn’t merely whether AI can improve revenue.
It’s whether AI can improve administrative efficiency while preserving accurate billing, appropriate care, and patient trust.
When financial optimization and clinical integrity conflict, technology shouldn’t quietly make the choice.
The OnnX Perspective: What If We Stopped Building Better Firefighters and Started Preventing Fires?
As a physician-entrepreneur and founder of OnnX, an AI-powered medical billing SaaS initiative focused on small and medium-sized clinics, I approach this problem from a different starting point.
Much of medical billing innovation focuses on what happens after something goes wrong.
A claim is rejected.
An authorization is missing.
A payment is delayed.
A denial arrives.
Then the machinery begins.
Investigate.
Correct.
Appeal.
Resubmit.
Wait.
Repeat.
My interest is in the question that comes before all of that.
Why did the problem happen, and could it have been prevented?
This principle informs the development direction of OnnX 2.0.
Rather than treating billing as an isolated financial process, the proposed approach emphasizes information quality throughout the patient encounter.
Eligibility.
Authorization.
Documentation.
Coding.
Charge capture.
Claim readiness.
Workflow visibility.
These activities are connected.
When information breaks down in one area, the consequences often appear elsewhere.
A practice experiencing recurring authorization denials may not need a more sophisticated appeals engine as its first intervention.
It may need better authorization tracking.
A practice experiencing frequent eligibility errors may need improved registration verification.
A practice experiencing documentation-related denials may need clearer documentation workflows rather than longer notes.
The right solution depends on the cause, not merely the symptom.
Preserve What Experienced People Already Know
One overlooked asset in medical billing is institutional knowledge.
Experienced billing professionals understand payer behavior, recurring exceptions, contractual complexity, and practical workflows.
That knowledge should not disappear when software changes or employees leave.
It should be preserved appropriately, documented, and incorporated into future systems while respecting privacy, security, and proprietary rights.
A new platform shouldn’t require a practice to forget everything its people have already learned.
Make Affordability Part of Innovation
Independent clinics cannot operate like major hospital systems.
They often lack large IT departments, specialized implementation teams, and substantial technology budgets.
A solution that requires expensive customization may be unsuitable for a small practice.
The goal should be practical improvement at a sustainable cost.
That may involve selective automation, existing integrations, and carefully governed low-code or no-code components where appropriate.
Security and compliance depend on the actual architecture and implementation, not the development label.
A brilliant billion-dollar solution is useless to a clinic that can’t afford the subscription.
Shift From Denial Management to Denial Prevention
The direction I’m exploring for OnnX 2.0 emphasizes upstream information quality, error prevention, workflow visibility, and measurable operational improvement.
These are development priorities, not claims of already validated commercial performance.
The objective isn’t to eliminate experienced billers or necessary professional oversight.
It’s to reduce unnecessary intermediaries, repetitive work, and avoidable information loss.
Less friction. Better information. More predictable reimbursement. More time for medicine.
Future Outlook: Will Healthcare’s Next Breakthrough Be Less AI?
That sounds like a strange question in an industry racing to adopt artificial intelligence.
But consider what genuine progress might look like.
Fewer manual eligibility checks.
Fewer missing authorizations.
Fewer documentation corrections.
Fewer preventable claim rejections.
Fewer unnecessary appeals.
Less duplicate data entry.
Less time searching for information that should already be available.
None of these outcomes requires celebrating how many AI agents were deployed.
They require measuring how much unnecessary work disappeared.
Over the next several years, healthcare is likely to see more automated claims processing, electronic authorization, predictive analytics, and integrated clinical-financial workflows.
Those developments could be valuable.
But their greatest benefit may come from combining technological capability with better process design.
The future doesn’t necessarily belong to the company with the most AI. It may belong to the company that makes the most unnecessary work disappear.
For independent practices, reliability matters more than spectacle.
A physician doesn’t need another impressive dashboard that takes an hour to interpret.
A clinic owner doesn’t need five disconnected applications to solve one administrative problem.
A biller doesn’t need an AI assistant that creates more exceptions than it resolves.
What practices need is accurate information, predictable workflows, appropriate reimbursement, and fewer unnecessary interruptions.
Perhaps the next great healthcare innovation won’t look spectacular.
Perhaps it will be a medical claim that moves through the system correctly the first time.
Sometimes the most revolutionary technology is the one that makes ordinary things work as they should.
Frequently Asked Questions
Can AI eliminate medical billing denials?
No. Some denials involve legitimate coverage restrictions, medical necessity disputes, or contractual requirements. AI may help reduce certain preventable errors, but appropriate oversight and appeals remain necessary.
What is the difference between denial prevention and denial management?
Denial prevention identifies and corrects avoidable problems before they produce denials. Denial management investigates and responds to denials after they occur.
What should a small clinic automate first?
Start with the most frequent, measurable, and preventable operational problem. That might be eligibility verification, authorization tracking, claim validation, or denial classification.
Can AI replace medical billing companies?
AI may automate repetitive tasks, but complex coding, compliance, payer disputes, and exception management still require professional expertise.
Is AI medical billing software automatically HIPAA-compliant?
No. Compliance depends on the product, implementation, contracts, safeguards, data flows, and intended use.
How can a clinic reduce its denial rate?
Classify denials by root cause, identify recurring preventable errors, improve the responsible workflows, and measure results.
Does a lower denial rate always mean better performance?
No. A lower rate may reflect changes in claim mix, delayed submissions, or other factors. Evaluate it alongside collections, A/R, acceptance rates, and staff burden.
Why is data quality important for AI?
AI systems depend on their inputs. Incomplete or inaccurate information can produce unreliable recommendations, even when the underlying technology is sophisticated.
What is the biggest mistake when purchasing AI billing technology?
Buying the product before identifying the problem.
A software demonstration is not an operational diagnosis.
What should physicians know about CMS’s 2027 prior authorization requirements?
Certain affected payers must implement specified electronic prior authorization interfaces beginning in 2027. Requirements vary by payer and service, and electronic processing does not guarantee automatic approval.
What makes OnnX’s proposed approach different?
The proposed OnnX 2.0 direction emphasizes upstream information quality, error prevention, workflow visibility, and measurable improvement, rather than relying exclusively on downstream denial recovery.
These remain development goals requiring validation.
Where should an independent practice begin?
Start with one question.
Which administrative mistake does our practice keep paying to correct?
Find it.
Understand it.
Measure it.
Fix the process that creates it.
Then repeat.
Final Thoughts: Stop Building Smarter Mistakes
Healthcare has achieved remarkable clinical progress.
We diagnose diseases earlier.
We perform increasingly complex procedures.
We develop treatments that once seemed impossible.
Yet physicians still operate within administrative systems that can turn minor information errors into prolonged financial disputes.
Artificial intelligence can help.
But it cannot substitute for better thinking.
If we continue applying advanced technology to poorly designed processes, we may simply become more efficient at repeating the same failures.
The next phase of medical billing innovation should focus on information quality, prevention, accountability, and measurable improvement.
Because a denied claim isn’t always the original problem.
Sometimes it’s the final symptom of a problem that began long before the patient saw the physician.
And perhaps that’s where the solution should begin.
The most expensive billing mistake isn’t necessarily the one that costs the most to correct. It’s the one we keep paying to make.
The smartest AI isn’t always the one that wins the argument. Sometimes it’s the one that prevents the argument from becoming necessary.
Healthcare doesn’t need an infinite army of bots fighting over broken information. It needs better information before the battle begins.
Get Involved: The Question Every Physician Should Ask
If your practice could prevent half its avoidable billing errors tomorrow, would you invest in another denial-management tool—or fix the processes that create those errors?
Share your perspective in the comments. Which administrative problem creates the greatest burden in your practice: eligibility, prior authorization, documentation, coding, denials, or delayed reimbursement?
Repost this article if you believe physicians and clinic owners should help shape the next generation of healthcare technology rather than simply adapt to systems designed without them.
Less friction. Better information. Stronger independent practices.
References, Research & Related Reading
The following resources provide supporting evidence, regulatory context, and additional perspectives on medical billing, artificial intelligence, prior authorization, and denial prevention.
1. American Medical Association — 2025 Prior Authorization Physician Survey (Released 2026)
Primary research supporting the article’s statistics on prior authorization workload, treatment delays, physician burnout, and patient-care concerns.
2. American Medical Association — 8 Ways the AMA Is Shaping Medicine’s Future (November 6, 2026)
Examines physician-led healthcare innovation, AI governance, administrative reform, and the importance of involving physicians in shaping healthcare’s future.
3. Humana and athenahealth — Nationwide Expansion of Electronic Prior Authorization (November 3, 2026)
Describes the expansion of electronic prior authorization capabilities and the movement toward more integrated clinical and administrative workflows.
4. Centers for Medicare & Medicaid Services — Interoperability and Prior Authorization Final Rule
Official federal guidance on electronic prior authorization, decision timelines, interoperability requirements, and implementation milestones.
A practical physician-focused resource examining authorization delays, administrative costs, patient consequences, and reform opportunities.
6. Your Billing Denial Has a Plot Twist: The Problem May Have Started Before the Claim Existed
Explores how upstream failures in eligibility, authorization, documentation, and information exchange become downstream billing problems. Recommended companion reading for Section
7. The Cheapest Claim Is the One You Never Have to Fix
Examines the financial and operational value of denial prevention, reducing staff rework, and measuring revenue cycle improvement. Recommended companion reading for Section 10.
8. The Denial Isn’t the Problem. It’s the Clue.
Discusses claims complexity, recurring administrative failures, and why revenue cycle improvement should begin with identifying root causes rather than repeatedly correcting symptoms.
About the Author
Dr. Daniel Cham, MD, is a physician, medical consultant, healthcare entrepreneur, and founder of OnnX, an AI-powered medical billing SaaS initiative focused on reducing administrative complexity for independent medical practices.
His professional interests span medical technology, healthcare management, revenue cycle innovation, artificial intelligence, and physician entrepreneurship.
Through his writing and educational work, he examines the systems shaping modern healthcare and explores practical ways to improve operational efficiency, preserve clinical judgment, strengthen physician independence, and support better patient care.
Continue the Conversation: Challenge Assumptions. Build Better Systems.
Healthcare advances when professionals ask uncomfortable questions, share practical experience, and challenge processes that no longer serve their intended purpose.
Explore additional perspectives on medicine, healthcare technology, entrepreneurship, and practice management.
Connect: LinkedIn | Website | YouTube| Spotify
Free educational resources: Visit the Featured section of my LinkedIn profile for available practical guides and downloads. Direct-download resources require no signup.
Curiosity challenges assumptions. Better questions reveal better solutions. Better solutions move healthcare forward.
Disclaimer
This article provides general educational information and professional commentary on healthcare administration, medical billing, artificial intelligence, and revenue cycle management. It does not constitute individualized medical, legal, financial, coding, or regulatory advice. Readers should consult appropriately qualified professionals and applicable authoritative guidance before making decisions involving patient care, practice operations, technology implementation, or compliance.
#MedicalBilling #HealthcareAI #RevenueCycleManagement #DenialPrevention #PriorAuthorization #PhysicianEntrepreneur #IndependentPhysicians #HealthcareInnovation #MedicalPracticeManagement #HealthcareLeadership #ArtificialIntelligence #DigitalHealth #PhysicianBurnout #HealthcareFinance #OnnX
Leave a Reply