Hire an AI medical billing and coding specialist

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• A table of each code, with what the notes support, the quote behind it and what to check
• A list of missing documentation and a plain question to send to the provider
• A pre-submission checklist marked pass, fail or not supplied
• A denial triage with the cause type, the deadline to look up and the next action (if a denial)
• An appeal letter draft with blanks for patient details (if a denial)
• A note on who must review it before anything goes to a payer

How it does the job well

• Code what is documented. Never assume a diagnosis, raise a code or add a service. That is a compliance risk. Ask the provider instead.
• You cannot verify current code sets or payer policy unless the user supplies them or you have live web access. Mark such points Unverified and tell the user to check the current ICD-10-CM, CPT and HCPCS books and the payer's policy. The payer's rule wins where it is stricter.
• Never say a claim is ready to submit. The sign-off note always appears.
• Never bill for a service that is not documented as done.
• Overpayments. If a review shows possible overpayment, tell the user to stop the pattern and take it to their compliance lead. Do not advise ignoring it.
• Code what is documented — never what is assumed. Coding must reflect what the provider documented in the medical record. Never infer diagnoses, upcode procedures, or assign codes for conditions not documented. This is fraud.
• Specificity is required in ICD-10. ICD-10 demands the highest level of specificity available. "Diabetes" is not sufficient — "Type 2 diabetes mellitus with diabetic chronic kidney disease, stage 3" is. Unspecified codes should be a last resort, not a default.

Mistakes it avoids

• Mixing cash and accrual figures, or different periods, in one comparison.
• Rounding or formula errors that do not tie back to totals.
• Giving personal investment advice.

What a good result looks like

An example, not your task: A provider asks if the notes support the E/M level or modifier they billed.

• Code review. Each code or modifier with what the notes support, a quote, a status and what to check.
• Documentation questions. Plain questions for the provider where the notes do not support a code.
• Claim checklist. Demographics, provider, coding, linking and filing checks, each marked pass, fail or not supplied.
• Denial triage. The cause type, the deadline to look up and the next action for each denied claim.
• Appeal letter draft. A letter with blanks for patient details, the coding and clinical reasons, and an enclosures list.
• Sign-off note. A reminder that a certified reviewer must approve the work before it goes to a payer.

How the work is checked

• Totals tie out, and differences are explained line by line.
• Forecast assumptions are listed and easy to change.
• Anything not checked by a qualified person is marked as needing review.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

A second opinion

Before it hands back, your AI reviews the draft the way Byron Sharp would. This is a way of reviewing the work. It does not mean they wrote or endorse it.

Capabilities

• Set the job and the facts. It confirms the task, payer, specialty and year of service. Missing facts are asked once, then marked MISSING.
• Read the notes. It finds what the provider documented, such as diagnoses, procedures, time and medical need. It does not assume anything else.
• Check each code. It compares each code and modifier to the notes. It marks each Supported, Not supported, Needs a provider question, or Unverified.
• Scrub the claim. It runs a checklist on demographics, provider details, code accuracy, diagnosis links, place of service and filing window.
• Triage a denial. It sorts the cause as administrative, clinical, authorization or coding, and names the appeal deadline to look up in the notice.
• Draft and hand over. It writes the provider questions, the appeal letter or the KPI readout, with a sign-off note for the reviewer.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id medical-billing-coding-specialist only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/medical-billing-coding-specialist/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• Visit notes or an operative report, with patient name, date of birth, member ID and record numbers removed
• The codes and modifiers billed, or the codes you want checked
• The payer, the provider's specialty and the year of service
• The denial notice or reason codes, if this is a denial (optional)
• The payer's policy text for the service, if you have it (optional)

Helpful context (optional)

• The payer's published policy for the service. It beats general rules where it is stricter.
• Your own denial log, so triage can spot repeat causes by payer.

Tools and access

• Claude, ChatGPT or Gemini in an account your practice has approved: Billing notes hold private health details, so use a tool your privacy lead has cleared.

Functions and deliverables

• Medical billing and coding specialist: Check medical coding and claims against the visit notes, triage denials and draft appeal letters. A certified coder signs off before you submit.
  • A table of each code, with what the notes support, the quote behind it and what to check
  • A list of missing documentation and a plain question to send to the provider
  • A pre-submission checklist marked pass, fail or not supplied
  • A denial triage with the cause type, the deadline to look up and the next action (if a denial)
  • An appeal letter draft with blanks for patient details (if a denial)
  • A note on who must review it before anything goes to a payer
  Source: Michael Sitarzewski and Agency Agents contributors.
• Offer letter drafter: Share the role, location, approved pay and start date. Draft an offer letter, pay table and manager notes for review.
  • A drafted offer letter
  • A pay package table with first-year total
  • Terms and conditions text
  • Talking points for the offer call
  Source: Anthropic.
• Loyalty claim checker: Test a loyalty scheme claim before you pay for it. Paste the plan and its numbers. You get the sums checked, the costs added and a plain verdict.
  • A table of where the gain comes from
  • The costs and the room-to-buy check
  • A side-by-side with reaching more buyers
  • A plain verdict and the data to ask for
• Nonprofit annual appeal planner: Plan your nonprofit annual appeal. Share donor groups and gift history, and get appeal letters, gift amounts, a send calendar and results measures.
  • A plan that splits your donors into groups, each with its own message.
  • Suggested gift amounts for each group, based on past giving.
  • An appeal letter or email with a P.S. and matching reply wording.
  • A dated send calendar that spaces out your asks.
  • A scorecard to track response rate, average gift and cost to raise a dollar.
  Source: Nonprofit Skills.
• me.md builder: Answer a few questions, or let your AI use what it remembers. Get one private file that tells any AI who you are.
  • One me.md file with up to ten dated sections
  • A list of facts taken from memory, for you to check
  • A short note of what it removed for privacy
  • The two next steps, to save it and set it up
• Byron Sharp: Review a brand growth plan using Byron Sharp’s public work. Check buying evidence, reach more potential buyers and make the product easier to find.
  • Your problem worked on the way Byron Sharp would: their questions, their method, their push-back
  • The answer or the work, in Byron Sharp's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step

Tasks you can use it for

• A provider asks if the notes support the E/M level or modifier they billed.
• A claim came back denied and you need to sort the cause and draft an appeal.
• You want a pre-submission checklist run on a batch of claims.
• You have KPI numbers, such as denial rate and days in accounts receivable, and want a plain readout of what they suggest.

Limits to check before starting

• You want the final code or claim approved. A certified coder or billing professional does that, using the current code books and the payer's policy.
• You want to reply to a patient about a bill or a question. Use the Patient service reply writer.
• You are building billing software or a payments integration. Use the Payments and billing engineer agent.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Protect patient data first. Remove names, dates of birth, member IDs and record numbers. Use this AI only in a way your practice's privacy rules allow.
• Add the pack as a skill, or paste SKILL.md and then references/source-agent.md into a new chat.
• Paste the notes, the codes billed, the payer, the specialty and the year. Say the job, such as "Check these codes against the notes" or "Triage this denial and draft an appeal."
• Read the missing documentation list. Ask the provider to answer the questions. Do not code from guesses.
• Look up each code in the current code books, and read the payer's policy. The AI cannot confirm they are current.
• Have a certified coder or billing lead approve everything before you submit, appeal or refund anything.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/medical-billing-coding-specialist for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.