AI Prompts for Medical Coders: ICD-10, CPT, Claim Accuracy

Published: July 2026 | 8 min read

Medical coding is a precision profession. One wrong code means claim denials, delayed reimbursement, or compliance violations. Coders navigate thousands of ICD-10-CM codes, CPT procedure codes, HCPCS Level II codes, and payer-specific guidelines — all of which change annually. AI can't replace a certified coder's judgment, but it can dramatically speed up code research, audit preparation, and denial analysis.

The prompts below are structured for real coding workflows. Each defines the coding scenario, required references, and output format that maps to how coders actually work. These are adapted from Skillent's Healthcare AI Prompt Library.

Code Lookup & Verification

1. ICD-10-CM Code Narrowing

Role: Certified medical coder (CCS, CPC)
Task: Help narrow ICD-10-CM code options for the following diagnosis description
Input: "[e.g., type 2 diabetes with diabetic nephropathy]"
Steps:
1. List all potentially applicable ICD-10-CM codes
2. For each code: full code, description, chapter
3. Identify the most specific code based on documentation
4. Note any "use additional code" instructions
5. Flag if a combination code is more appropriate than multiple codes
Output: Table with code, description, specificity notes, recommendation
Disclaimer: Verify against current ICD-10-CM Official Guidelines for Coding and Reporting

2. CPT Code Selection Assistant

Role: CPC-certified coder
Task: Identify appropriate CPT codes for the following procedure documentation
Input: "[paste operative report or procedure note summary]"
Process:
1. Identify the procedure performed
2. List candidate CPT codes (range)
3. Note any bundling issues (NCCI edits)
4. Identify needed modifiers (-25, -59, -51, -LT/-RT, etc.)
5. Check if separate codes vs. combination code applies
Output: Recommended code(s) with modifier(s), rationale, and bundling notes
Reference: Current CPT codebook and NCCI edit files

3. HCPCS Level II Code Finder

Role: Medical coder specializing in DME/supplies
Task: Find HCPCS Level II codes for the following items/supplies
Input: [list of supplies, equipment, or drugs]
For each item:
1. HCPCS code
2. Code description
3. Category (DME, drug, supply)
4. Payment category indicator
5. Notes on coverage restrictions
Format: Table sorted by category
Flag: Items that may need advance beneficiary notice (ABN)

Claim Auditing & Compliance

4. Pre-Submission Claim Audit

Role: Coding auditor (CPMA)
Task: Audit the following claim before submission
Input:
- Diagnosis codes: [list]
- Procedure codes: [list]
- Modifiers: [list]
- Payer: [Medicare/Medicaid/Commercial]
Check:
1. Code specificity (is there a more specific code?)
2. Medical necessity (do dx codes support procedure codes?)
3. NCCI edit conflicts
4. Modifier appropriateness
5. LCD/NCD coverage verification
6. Code sequencing (principal vs. secondary)
Output: Audit findings with severity (Error/Warning/Info) and recommended corrections
Format: Audit report with line-by-line findings

5. E/M Level Verification

Role: E/M coding specialist
Task: Verify the E/M level for the following encounter documentation
Input: [paste provider note summary]
Determine:
1. E/M category (office visit, hospital, ER, etc.)
2. Level of MDM (Straightforward/Low/Moderate/High)
3. Supporting elements: problems, data reviewed, risk
4. Time-based coding eligibility (if documented)
5. Recommended E/M code
Output: MDM worksheet showing each element scored, final level, and code recommendation
Note: Apply 2021+ office visit guidelines or 2023+ hospital guidelines as applicable

6. Coding Compliance Checklist Generator

Role: Coding compliance officer
Task: Create a coding compliance checklist for [provider specialty: cardiology/orthopedics/internal medicine]
Include:
1. Documentation requirements per code level
2. Common coding errors for this specialty
3. Red flags that trigger audits
4. Medical necessity documentation requirements
5. Modifier usage rules
6. Bundling rules specific to specialty
Format: Checklist with checkboxes, organized by category
Include: References to relevant CMS publications and coding guidelines

Denial Analysis & Appeals

7. Denial Root Cause Analysis

Role: Revenue cycle analyst with coding expertise
Task: Analyze the following claim denial and identify root cause
Input:
- Denial reason code: [code]
- Payer: [name]
- Submitted codes: [list]
- Denial description: [text]
Analysis:
1. Most likely cause of denial
2. Whether this is a coding error, documentation issue, or coverage issue
3. Corrective action needed
4. Whether appeal is warranted
5. Appeal strategy if applicable
Output: Denial analysis report with recommended next steps
Include: Estimated probability of successful appeal (Low/Med/High)

8. Appeal Letter Generator

Role: Coding/billing appeal specialist
Task: Draft a claim appeal letter for [denial reason]
Input:
- Payer: [name]
- Claim number: [number]
- Patient: [de-identified]
- Original codes: [list]
- Denial reason: [text]
- Medical record evidence: [summary]
Include:
1. Clear statement of what is being appealed
2. Reason for appeal with code citations
3. Supporting documentation references
4. Medical necessity justification
5. Request for reconsideration
Tone: Professional, factual, non-confrontational
Format: Business letter, 1 page maximum

Documentation Improvement

9. CDI Query Opportunity Finder

Role: Clinical documentation improvement specialist
Task: Review the following provider documentation and identify query opportunities
Input: [paste clinical documentation]
Look for:
1. Unspecified diagnoses that could be more specific
2. Missing complications/comorbidities
3. Conditions documented but not coded
4. Vague terms ("probably," "likely," "versus")
5. Missing physician signatures or dates
6. Inconsistent documentation between sections
Output: List of query opportunities with suggested query language for each
Format: Table with issue, location in doc, suggested query, priority (High/Med/Low)
Do NOT: Lead the provider — queries must be non-leading per AHIMA guidelines

10. Operative Note Coding Prep

Role: Surgical coder
Task: Extract coding-relevant information from this operative note
Input: [paste operative note]
Extract:
1. Procedure performed (exact terminology)
2. Approach (open/laparoscopic/robotic)
3. Body part(s) involved
4. Ancillary procedures
5. Devices/grafts implanted
6. Estimated blood loss
7. Complications encountered
8. Surgeon's assessment
Output: Coding summary sheet ready for code assignment
Flag: Any missing elements needed for accurate code selection

Education & Training

11. Coding Scenario Practice

Role: Coding educator
Task: Create [number] practice coding scenarios for [specialty/topic]
Each scenario includes:
1. Patient encounter summary (2-3 paragraphs)
2. Provider documentation excerpt
3. Answer key with:
   - Principal diagnosis code + rationale
   - Secondary diagnosis codes + rationale
   - Procedure codes + rationale
   - Modifiers if needed
   - MS-DRG assignment (if inpatient)
4. Common mistakes to avoid
Difficulty: Progressive (beginner to advanced)
Format: Case study format suitable for CE or exam prep

Best Practices for Medical Coding AI Prompts

1. Always cite the code source — AI should reference ICD-10-CM, CPT, HCPCS Official Guidelines. You verify against current codebooks.

2. Never accept codes blindly — AI may reference outdated codes. Always verify against the current year's code set.

3. Use de-identified data only — Remove all PHI before entering clinical documentation into AI tools.

4. Structure the coding question — Define specialty, payer, setting (inpatient/outpatient), and code year for accurate results.

5. Follow AHIMA/ADA ethical standards — AI assists coding; it doesn't replace certified coder judgment. See our medical billing prompts for billing-specific workflows.