AI tools, scored for your job
Learn AI in 30 days
Home AI Directory Career Paths AI News
Home AI News Pro Services
📰 Pro Services

AI Prompts for Medical Billers & Coders: 25 to Save (2026)

25 copy-paste AI prompts for medical billers and coders to draft denial appeals, check coding accuracy, and write patient communications as of September 20

September 5, 2026· 19 min read

The short answer

These 25 copy-paste AI prompts help medical billers and coders draft denial appeals, analyze documentation for coding accuracy, request prior authorizations, and communicate with patients. Always de-identify patient data before using public AI tools like ChatGPT, as they are not HIPAA-compliant by default.

Verified against live pricing pages·30 Aug 2026·How we test

Large language models (LLMs) like ChatGPT and Claude are powerful assistants for administrative work, but most public prompt libraries are built for physicians or generic office tasks. As a medical biller or coder, your needs are more specific, tied to a vocabulary of CPT codes, CARC/RARC denial reasons, and payer-specific policies. Using the right prompts can save hours of manual work each week.

At ZEKAI, we review AI tools independently. This guide provides 25 practical, copy-paste prompts written specifically for the workflows of professional billers and coders. We’ll cover denial appeals, coding validation, patient communication, and more—all with a critical focus on compliance. For a deeper look at the software landscape, see our hub for AI in medical billing and coding.

Before You Paste: The #1 Rule for AI and PHI

Before using any prompt from this or any other guide, you must understand the compliance risks. Public-facing AI models like the free versions of ChatGPT, Claude, and Gemini are not HIPAA-compliant out-of-the-box. Any Protected Health Information (PHI) you paste into them can be used to train the model and is not secured by a Business Associate Agreement (BAA).

De-identification is mandatory.

Before pasting any clinical note, explanation of benefits, or patient message into a public LLM, you must remove all 18 identifiers defined by HIPAA.

De-identified data includes removing:

Replace names with placeholders like [PATIENT], [PROVIDER], or [FACILITY]. This protects patient privacy and your organization’s legal standing. For routine work with PHI, use a HIPAA-compliant AI platform that provides a BAA.

Denial Management & Appeal Prompts

Denied claims are a primary source of lost revenue, yet up to 60% of them are never resubmitted. AI can’t automatically win an appeal, but it can draft a structured, evidence-based letter in seconds, turning a 30-minute task into a 5-minute review.

The key is to give the AI the right context: the original claim details, the payer’s exact denial reason (CARC/RARC codes), and the specific policy or clinical documentation that refutes the denial.

Prompt 01 Draft an Appeal for Medical Necessity (CARC 50)
[Paste the relevant, de-identified section of the clinical note here. For example: "Patient presented with [symptoms]. Prior conservative treatments including [treatment 1] and [treatment 2] over [duration] failed to provide relief. The provider, Dr. [PROVIDER], determined [service] was the necessary next step to diagnose/treat the condition."]
Goal: Write a formal appeal letter to [Payer Name]. The letter must argue that the service was medically necessary according to the patient’s condition and standard clinical guidelines.
Prompt 02 Analyze Denial Pattern from EOBs
[Paste the text from multiple EOBs here. Ensure all PHI is removed.]
Tested on Claude, ChatGPT and Gemini
Prompt 03 Appeal a Timely Filing Denial (CARC 29)
Act as a medical billing specialist. A claim was denied for timely filing, but we have proof of original, timely submission.
**Context:**
*   **Claim Number:** [Claim #]
*   **Date of Service:** [YYYY-MM-DD]
*   **Payer's Timely Filing Limit:** [Number] days
*   **Date of Original Submission:** [YYYY-MM-DD]
*   **Proof of Submission:** We have a clearinghouse confirmation report or payer acceptance report from the original submission date. The report number is [Report #].
**Instructions:**
Write a formal letter that:
1.  Clearly states we are appealing a denial for timely filing (CARC 29).
2.  States the date of the original, timely submission.
3.  References the attached proof of submission (e.g., "Attached is the clearinghouse acceptance report [Report #] dated [YYYY-MM-DD], confirming receipt by your systems well within the [Number]-day filing limit.").
4.  Politely requests that the claim be reprocessed for payment based on this evidence.
Goal: Draft a concise appeal letter to [Payer Name] to overturn a CARC 29 denial.
Prompt 04 Draft an Appeal for Bundling/Inclusive Service (CARC 97)
[Explain why the modifier is appropriate. For example: "The denied procedure was performed on a separate anatomical site from the primary procedure, as documented in the attached operative report. Therefore, per NCCI guidelines, modifier 59 is appropriate to designate it as a distinct procedural service."]
Goal: Write an appeal letter explaining why the service is not bundled and should be reimbursed.
Prompt 05 Reframe Clinical Jargon for a Payer
[Paste the jargon-heavy clinical note section here.]
Tested on Claude, ChatGPT and Gemini

While these prompts are effective, some organizations turn to dedicated software for this process. Tools like Appeal Management AI (Symplr/SafeNet) aim to automate parts of the denial and appeals workflow.

Coding Accuracy & Documentation Query Prompts

Coding errors are a massive drain on the healthcare system. One report attributes an annual cost of $36 billion to these mistakes, stemming from lost revenue and compliance penalties. While base LLMs are not yet accurate enough for autonomous coding, they are excellent at reviewing documentation against coding guidelines.

33.9%

According to a 2024 study cited in *Frontiers in Nephrology*, GPT-4’s exact-match accuracy for ICD-10-CM codes was only 33.9% when prompted with just diagnosis descriptions, highlighting the need for human oversight. Source: frontiersin.org

Use these prompts to have an AI act as a second set of eyes, flagging potential documentation gaps or suggesting more specific codes based on the text.

Prompt 06 Check Documentation for CPT Code Requirements
[Paste key requirements for the code. For example: "Requires documentation of at least 30 minutes of continuous service," or "Must include documentation of [specific component 1] and [specific component 2]."]
Tested on Claude, ChatGPT and Gemini
Prompt 07 Identify More Specific ICD-10 Codes
[e.g., "Patient has been non-compliant with medication and presents with blurred vision and foot ulcer. A1c is 9.8. This is Type 2 diabetes."]
Tested on Claude, ChatGPT and Gemini
Prompt 08 Draft a Provider Query for Documentation Clarity
Act as a CDI specialist. I need to write a polite and compliant query to a provider to clarify their documentation for coding purposes. The note is ambiguous.
**Context:**
*   **Patient:** [PATIENT ID]
*   **Date of Service:** [YYYY-MM-DD]
*   **Ambiguous Documentation:** `[e.g., "Patient has anemia."]`
*   **Issue:** The term "anemia" is not specific enough. To code accurately, I need to know if it is acute, chronic, blood-loss related, or another type.
**Instructions:**
Write a non-leading query for Dr. [PROVIDER].
1.  Start by referencing the patient and date of service.
2.  State the current documentation.
3.  Explain what clarification is needed for accurate coding.
4.  Provide multiple-choice options based on common clinical indicators, if appropriate, but also include a free-text option.
5.  End with a thank you. Do not suggest any specific diagnosis that would lead to higher reimbursement.
Tested on Claude, ChatGPT and Gemini
Prompt 09 Explain CPT Modifier Usage
Act as a coding educator. A junior biller is asking when to use CPT modifier 25.
**Question:** "When do I use modifier 25?"
**Instructions:**
Explain the proper use of modifier 25 in simple terms.
1.  Define what modifier 25 signifies ("Significant, Separately Identifiable E/M Service").
2.  Provide a clear "Use When" scenario (e.g., a patient comes in for a scheduled procedure, but also presents with a new, acute problem that requires a separate E/M workup).
3.  Provide a clear "Do Not Use When" scenario (e.g., the E/M service was just the routine pre-operative assessment for the scheduled procedure).
4.  Emphasize the importance of documentation supporting the separate nature of the E/M service.
Tested on Claude, ChatGPT and Gemini
Prompt 10 Find HCC Codes from Documentation
[Paste a comprehensive clinical summary, including problem lists, history, and assessment/plan.]
Tested on Claude, ChatGPT and Gemini

Prior Authorization Request Prompts

Prior authorizations are a major administrative burden. AI can help by drafting a structured request that includes all the necessary elements a payer needs to make a decision, reducing back-and-forth communication.

Prompt 11 Draft a Prior Authorization Request
[Paste a summary of the patient's history, symptoms, what treatments have already failed, and why the provider believes this procedure is the necessary next step.]
Tested on Claude, ChatGPT and Gemini
Prompt 12 Summarize Payer’s Prior Auth Policy
[Paste the full, dense text of the payer's medical policy here.]
Tested on Claude, ChatGPT and Gemini
Prompt 13 Check Status of Multiple Authorizations
[Paste your unstructured notes, e.g., "9/3 called on auth 12345 for patient Smith, still pending. auth 67890 for Jones approved today. auth 54321 for doe denied, needs peer to peer. called again on 12345, rep said needs 2 more days."]
Tested on Claude, ChatGPT and Gemini

Patient Billing Communication Prompts

Communicating with patients about their bills requires empathy, clarity, and accuracy. AI can help draft messages that are easy to understand and provide clear calls to action, freeing you up to handle more complex patient inquiries.

Prompt 14 Explain a Bill to a Patient
Act as a patient financial services representative. I need to write a simple, clear email to a patient explaining their recent bill.
**Bill Details:**
*   **Total Charge:** $[Amount]
*   **Insurance Paid:** $[Amount]
*   **Insurance Adjustment:** $[Amount]
*   **Patient Responsibility:** $[Amount]
*   **Reason for Patient Balance:** [e.g., "Applied to annual deductible," "Copay amount," "Coinsurance," "Non-covered service"]
**Instructions:**
Write a short, friendly email that:
1.  States the date of service and provider seen.
2.  Breaks down the bill using the details above in a simple, non-technical way.
3.  Clearly explains *why* they have a balance (e.g., "Your insurance plan requires you to pay for medical costs up to your $[Amount] deductible each year. This charge was applied to that deductible.").
4.  Provides a clear link to the payment portal and a phone number for questions.
Tested on Claude, ChatGPT and Gemini
Prompt 15 Offer a Payment Plan
Act as an empathetic billing coordinator. A patient has called and said they cannot afford to pay their full balance at once.
**Context:**
*   **Patient Balance:** $[Amount]
*   **Practice Policy:** We can offer a payment plan of up to [Number] months. A minimum payment of $[Amount] is required.
**Instructions:**
Draft a script for what to say to the patient. The script should:
1.  Acknowledge their concern and show empathy.
2.  Clearly explain the payment plan option (e.g., "We can definitely set up a payment plan for you. We can split the balance of $[Amount] into [Number] monthly payments of $[Amount].").
3.  Confirm that this arrangement will prevent any further collection activity as long as payments are made on time.
4.  Ask if they would like to set that up now.
Tested on Claude, ChatGPT and Gemini
Prompt 16 Pre-service Good Faith Estimate Notice
[e.g., "Lab work (approx. $[Amount])", "Anesthesia (approx. $[Amount])"]
Tested on Claude, ChatGPT and Gemini
Prompt 17 Follow-Up on Unpaid Bill (First Notice)
Act as a billing office manager. I need to write a gentle first reminder email for an invoice that is now 15 days past due.
**Invoice Details:**
*   **Invoice Number:** [Number]
*   **Date of Service:** [YYYY-MM-DD]
*   **Amount Due:** $[Amount]
*   **Due Date:** [YYYY-MM-DD]
**Instructions:**
Draft a polite email that:
1.  Serves as a friendly reminder about the outstanding balance.
2.  Includes all relevant invoice details.
3.  Assumes the patient may have simply forgotten and avoids an accusatory tone.
4.  Provides a direct link for payment and a phone number in case they have questions or believe there is an error.
5.  Mentions that they can call to discuss payment plan options if needed.
Tested on Claude, ChatGPT and Gemini

Some companies, like Avelis Health, are building AI-powered services to help patients (and their employers) analyze and negotiate bills, indicating a growing trend in using technology for patient-facing financial interactions.

Compliance & Audit-Readiness Prompts

Compliance is non-negotiable. AI can be a useful tool for summarizing complex regulations or creating audit checklists, but it should never be the final authority. Always verify AI-generated compliance information against official sources like CMS, OIG, and payer manuals.

The consequences of non-compliance are severe. In November 2024, UCHealth agreed to a $23 million settlement to resolve allegations that it violated the False Claims Act by automatically upcoding certain emergency room visits to the highest level CPT code without sufficient documentation. This case underscores the risk of relying on automated systems without robust human oversight.

$23 Million

UCHealth agreed to pay $23 million in November 2024 to settle allegations of False Claims Act violations related to the automated upcoding of E&M claims. Source: justice.gov

Prompt 18 Summarize a Payer’s Billing Policy
Act as a compliance analyst. I have the full text of a local coverage determination (LCD) from a Medicare Administrative Contractor (MAC). I need a simple summary of the key "rules" for billing a specific service.
**LCD Document Text:**
`[Paste the entire LCD text here.]`
**Instructions:**
1.  Identify the specific service or CPT codes the policy applies to.
2.  Create a section called "Covered Diagnoses" and list the ICD-10 codes that support medical necessity.
3.  Create a section called "Documentation Requirements" and summarize the specific elements the clinical note must contain.
4.  Create a section called "Limitations" and list any frequency limits or non-covered scenarios.
5.  The final output should be a one-page summary an auditor or biller can quickly reference.
Tested on Claude, ChatGPT and Gemini
Prompt 19 Create a Self-Audit Checklist
Act as a HIM director. I want to perform a proactive internal audit of our coding for CPT code [Code].
**Instructions:**
1.  Research the official CMS, AMA, and NCCI guidelines for CPT code [Code].
2.  Generate a checklist of 10-15 questions for the audit.
3.  The questions should be in a "Yes/No" format.
4.  Examples of questions:
*   "Does the documentation clearly support the medical necessity for the procedure?"
*   "Are all required components of the service (e.g., time, elements) documented?"
*   "Is the correct modifier used if billed with another service?"
*   "Does the diagnosis code linked to the procedure meet payer policy requirements?
Goal: Create a comprehensive audit checklist based on official guidelines.
Prompt 20 Translate OIG Work Plan Item into Action Steps
[Paste the description of the OIG work plan item here.]
Tested on Claude, ChatGPT and Gemini
Prompt 21 Identify Risks in a Billing Process Description
[e.g., "The provider finalizes their note in the EHR. A biller then reviews the note, selects the CPT and ICD-10 codes, and enters them into the practice management system. The claim is then batched and sent to the clearinghouse at the end of the day."]
Tested on Claude, ChatGPT and Gemini

Productivity & Workflow Prompts

Beyond core billing tasks, AI can help with training, process documentation, and data analysis.

Prompt 22 Create a Training Document for a New Biller
Act as a senior revenue cycle trainer. I need to create a simple onboarding guide for a new medical biller about how to handle a denial for "missing or invalid diagnosis code."
**Instructions:**
Write a short, step-by-step guide.
1.  Title the guide: "How to Resolve a Denial for Invalid Diagnosis."
2.  **Step 1:** Explain how to identify the denial on an EOB.
3.  **Step 2:** Explain where to look in the patient's record to find the correct diagnosis from the provider's note for that date of service.
4.  **Step 3:** Explain how to correct the claim in the practice management system.
5.  **Step 4:** Explain how to resubmit the corrected claim to the payer.
6.  Add a "Pro Tip" about checking for more specific codes to prevent future denials.
Tested on Claude, ChatGPT and Gemini
Prompt 23 Draft a Standard Operating Procedure (SOP)
Act as a process improvement manager. I need to document our practice's procedure for patient registration.
**Core Steps (Unordered):**
*   Scan insurance card and driver's license.
*   Have patient fill out demographic sheet and medical history form.
*   Run eligibility check with the payer.
*   Collect copay.
*   Enter all data into the PM system.
**Instructions:**
Organize these steps into a formal Standard Operating Procedure (SOP).
1.  Give it a title: "SOP: New Patient Registration".
2.  Create a "Purpose" statement.
3.  List the steps sequentially and add detail to each one (e.g., for "Run eligibility check," add "Verify active coverage, copay, and deductible status.").
4.  Create a "Key Responsibilities" section clarifying that this is a front-desk staff duty.
Tested on Claude, ChatGPT and Gemini
Prompt 24 Summarize an EOB for Data Entry
[Paste the full text of a single-claim EOB here.]
Tested on Claude, ChatGPT and Gemini
Prompt 25 Generate Excel Formula for A/R Aging Report
Act as a data analyst specializing in Excel. I have a spreadsheet of outstanding claims.
*   Column A has the Patient Name.
*   Column B has the Patient Balance.
*   Column C has the Date of Service.
I need an Excel formula to categorize each claim into an aging bucket (0-30, 31-60, 61-90, 91-120, 120+ days) based on the Date of Service in Column C relative to today's date.
**Instructions:**
Provide a single Excel formula that I can place in Column D. The formula should use the `TODAY()` function and `IF` statements to return one of the following text values: "0-30", "31-60", "61-90", "91-120", or "120+".
Tested on Claude, ChatGPT and Gemini

By integrating these prompts into your daily work, you can leverage AI as a powerful assistant to improve efficiency, accuracy, and compliance. Remember to always use your professional judgment and verify any AI-generated content before it goes out the door. For more on how AI is reshaping the profession, explore our AI in medical billing and coding hub.

Can ChatGPT write medical denial appeals?

Yes, ChatGPT can draft a strong first version of a medical denial appeal letter. You must provide it with the de-identified claim details, the payer’s specific denial reason (CARC/RARC codes), and the clinical evidence from the medical record. A human expert must always review, edit, and finalize the letter before submission.

Are AI prompts safe to use with patient information?

No, it is not safe to use standard AI tools like the public version of ChatGPT with identifiable patient information. These tools are not HIPAA-compliant by default. You must de-identify all data by removing the 18 HIPAA identifiers (names, dates, MRNs, etc.) before pasting it into a public AI.

How do I de-identify a medical note for AI?

To de-identify a medical note, you must remove or replace all 18 PHI identifiers. Systematically find and replace names with placeholders like [PATIENT], remove specific dates of birth or service (keeping the year is okay), delete addresses, and remove all medical record numbers, account numbers, and other unique identifiers.

Can AI help with medical coding?

AI can assist with medical coding but cannot yet replace a certified coder. AI is best used for tasks like scanning documentation for missing elements required by a CPT code, suggesting more specific ICD-10 codes based on clinical indicators, or flagging inconsistencies. All AI-suggested codes must be verified by a human.

What are the best AI prompts for analyzing claim denials?

The best prompts for analyzing claim denials ask the AI to act as a revenue cycle analyst. Provide it with de-identified text from multiple EOBs and instruct it to extract all CARC and RARC codes, count the frequency of each, and present the results in a sorted table to identify your top denial trends.

Will AI replace medical billers and coders?

No, AI is not expected to replace medical billers and coders but rather to change the nature of their work. AI will automate repetitive tasks like data entry and first-draft creation, allowing humans to focus on more complex, high-value work like handling complex appeals, auditing, and ensuring compliance, which still requires critical thinking.

Sources (25)
  1. U.S. Department of Justice. “UCHealth Agrees to Pay $23M to Resolve Allegations of Fraudulent Billing for Emergency Department Visits.” November 12, 2024. https://www.justice.gov/opa/pr/uchealth-agrees-pay-23m-resolve-allegations-fraudulent-billing-emergency-department-visits
  2. Whistleblower Law Collaborative LLC. “Qui Tam Whistleblower Awarded Nearly $4 Million After Alleging Improper Billing by UCHealth.” November 20, 2024. https://www.whistleblowerllc.com/qui-tam-whistleblower-awarded-nearly-4-million-after-alleging-improper-billing-by-uchealth/
  3. HHS Office of Inspector General. “UCHealth Agrees To Pay $23M To Resolve Allegations Of Fraudulent Billing For Emergency Department Visits.” November 12, 2024. https://oig.hhs.gov/fraud/enforcement/uchealth-agrees-to-pay-23m-to-resolve-allegations-of-fraudulent-billing-for-emergency-department-visits/
  4. Brown, LLC. “6 Biggest False Claims Act Healthcare Fraud Cases of 2024.” February 19, 2025. https://www.firstamendmentfirm.com/blog/6-biggest-false-claims-act-healthcare-fraud-cases-of-2024/
  5. Gibson Dunn. “False Claims Act 2024 Year-End Update.” January 23, 2025. https://www.gibsondunn.com/false-claims-act-2024-year-end-update/
  6. PracticeForces. “Certified Coding: Ensuring Accuracy in Medical Billing.” April 11, 2024. https://practiceforces.com/blog/certified-coding-ensuring-accuracy-in-medical-billing/
  7. Medmio. “Autonomous Coding You Can Verify — An Auditable Accuracy Benchmark.” https://www.medmio.com/blog/autonomous-coding-benchmark
  8. NEJM AI. “Validation of 13102 ICD-10-CM Codes Using a Large Language Model-Based System.” December 31, 2025. https://ai.nejm.org/doi/full/10.1056/AIoa2400035
  9. Coding & Billing Solutions. “The Real Cost of Coding Errors: Why Auditing & Compliance Are More Critical Than Ever in 2026.” March 27, 2026. https://www.codingandbillingsolutions.com/blog/the-real-cost-of-coding-errors
  10. Coding & Billing Solutions. “The Hidden Cost of Medical Coding Errors: How to Stop Revenue Leakage in 2026.” October 22, 2025. https://www.codingandbillingsolutions.com/blog/the-hidden-cost-of-medical-coding-errors
  11. Frontiers in Nephrology. “AI integration in nephrology: evaluating ChatGPT for accurate ICD-10 documentation and coding.” September 1, 2024. https://www.frontiersin.org/journals/nephrology/articles/10.3389/fneph.2024.1423838/full
  12. JMIR Formative Research. “Developing an ICD-10 Coding Assistant: Pilot Study Using RoBERTa and GPT-4 for Term Extraction and Description-Based Code Selection.” February 11, 2025. https://formative.jmir.org/2025/1/e54541
  13. arXiv. “Training a Large Language Model for Medical Coding Using Privacy-Preserving Synthetic Clinical Data.” March 6, 2026. https://arxiv.org/html/2403.04273v2
  14. Allied Health Schools. “How a Medical Coder’s Accuracy Impacts a Practice’s Revenue.” December 12, 2025. https://www.alliedhealthschools.com/medical-billing-coding/how-a-medical-coders-accuracy-impacts-a-practices-revenue/
  15. Journal of AHIMA. “How to Identify and Address High-Risk Coding Errors.” July 17, 2023. https://journal.ahima.org/page/how-to-identify-and-address-high-risk-coding-errors
  16. Avelis Health. “Terms And Conditions.” April 17, 2025. https://www.avelishealth.com/terms
  17. symplr. “6 Tips to Navigate Insurance Peer-to-Peer Reviews.” September 30, 2024. https://www.symplr.com/blog/2024/6-tips-to-navigate-insurance-peer-to-peer-reviews
  18. symplr. “How to Overcome Payer Rejection During Enrollment.” May 5, 2020. https://www.symplr.com/blog/2020/how-to-overcome-payer-rejection-during-enrollment
  19. symplr. “The Complete Guide: How to credential, privilege, and evaluate advanced practice providers.” https://www.symplr.com/assets/pdfs/guides/symplr-guide-credentialing-privileging-evaluating-apps.pdf
  20. symplr. “Be ready for the 2025 healthcare compliance updates.” January 1, 2025. https://www.symplr.com/blog/2025/be-ready-for-the-2025-healthcare-compliance-updates
  21. Specter Insights. “Hottest Startups – September 2025.” October 6, 2025. https://www.specter.com/hottest-startups/september-2025
  22. IntuitionLabs.ai. “symplr Contract – Payer contract compliance Software.” https://www.intuitionlabs.ai/software/symplr-contract
  23. PitchBook. “Avelis 2026 Company Profile: Valuation, Funding & Investors.” https://pitchbook.com/profiles/company/600767-17
  24. Facebook. “Alaffia Health taps AI to detect errors in hospital bills.” https://www.facebook.com/monacolife.net/posts/alaffia-health-taps-ai-to-detect-errors-in-hospital-bills-httpstechcrunchcom2022/476978437798305/
  25. Hitech C-Suite. “Healthcare IT Startups funded by Y Combinator (YC) 2026.” https://www.hitechcsuite.com/healthcare-it-startups-funded-by-y-combinator-yc-2026/

See Zekai first in Google

This article is provided for general information only and does not constitute professional advice. Facts, product details, and figures were accurate to the best of our knowledge at the time of publication and may have changed since. Zekai is an independent publisher and is not affiliated with the companies mentioned. Spotted an error? See our Corrections & Removal Policy.
#AI tools#Medical Billing#tier-a

The weekly AI briefing for your profession

One weekly email: the AI changes that actually affect your profession — tools, deals, and what to do about them.

Free · 1 email/week · profession-segmented · unsubscribe anytime

See Zekai first in Google