For independent medical billing companies
Let each biller carry more accounts without burning out
Lunan builds AI work queues for your billing company, starting with denials: sorted by cause, dollars and deadline, with appeal drafts your billers approve. It runs on tools that sign a business associate agreement, inside the systems you already use.
More than half of health plans already use AI tools in administrative work. Only 25% of provider organizations do. Source: CAQH Index, via Dataspring, Feb 2026
What gets in the way
Your fee is under attack
You charge a share of collections, and at least one AI-first billing company has advertised a flat 1.8% fee to new practice clients. You need to carry more accounts per biller to hold your price.
Denials eat the day
Your billers spend hours working out why a claim was denied, hunting for records and rewriting the same appeal letters. Meanwhile payers are using AI on their side of the claim.
Every new client means another hire
Growth stalls because each practice you sign needs more biller hours, and good billers are hard to find and slow to train.
What we set up
Denial sort and appeal drafts
Each morning the queue reads new denials and remittances from your clearinghouse or billing system, matches them to claims, and ranks them by reason, payer, dollars and appeal deadline. For your most common denial types it drafts an appeal from your own past winning appeals and lists the records needed. A biller approves, edits or rejects every draft and submits it personally.
Before: billers open each denial cold and rewrite letters from memory. After: they start the day with a ranked list and a draft and checklist on each item.
Client monthly summary
For each practice you serve, the system drafts a plain-language monthly letter: what was collected, what is stuck and why, and what the practice needs to fix on its side. Your account manager edits and sends it.
Before: spreadsheet exports and a rushed email at month-end. After: a one-page client letter waiting for approval.
Claim-status follow-up, electronic first
The queue reads electronic claim-status responses from your clearinghouse first. Portal checks run only where the payer's terms allow it, and phone calls go to an AI voice vendor you choose. Every answer lands in the same list your billers already work.
Before: billers decide claim by claim whether to log in or pick up the phone. After: only the claims that truly need a person reach one.
How it works
- 1
Free conversation
We walk through how denials and status checks move through your shop today, who touches them and where time goes. You get a short written read of the best first queue. No pitch, and the notes are yours.
- 2
Build the first queue
We set up denial sorting and appeal drafts for one client practice or payer group first, on a plan covered by a business associate agreement, with your billers testing every step against a two-week baseline.
- 3
Run and widen
Once the first queue holds up against the baseline, we extend it to more payers and clients, add the client summary and status follow-up, monitor it, and review results with you each month.
Ways to work together
Each is scoped and quoted after the conversation.
Learn
Owners who want staff using AI safely on real billing work now.
- Live workshop for billers and account managers
- Appeal and payer-correspondence prompt library built from de-identified examples
- A one-page AI tool policy for patient data, naming which plans have a business associate agreement
- Review of AI features already in your billing system and clearinghouse
Build
Firms where denial work limits how many accounts each biller can carry.
- Denial intake, sorting and ranking queue
- Appeal drafts for your most common denial types, from your own winning appeals
- Connection to your clearinghouse and billing-system exports
- Before-and-after measurement against a two-week baseline
- Biller training and a written runbook
Run
Firms that want the queue kept current as payers and rules change.
- Monitoring and fixes
- Appeal template and payer-rule updates
- Monthly accounts-per-biller review with you
- Help adding clients, payers, the client summary and status follow-up
Guardrails
We build only on plans that sign a business associate agreement (BAA), such as Claude Enterprise with its HIPAA setting turned on or an API account under a signed BAA, and never on consumer or self-serve plans like ChatGPT Plus or self-serve ChatGPT Business, which do not come with one. A biller approves every appeal, status note and client letter before it leaves your shop, and the system never changes a code or submits a claim. Portal checks run only where the payer's terms allow automation. Lunan holds no HIPAA certification; we tell you in writing where each piece of data goes so you can check it against your own policies. This page is not legal advice; confirm HIPAA questions with your compliance lead or attorney.
Questions
Will this replace my billers?
It changes their day more than it removes them. The goal is that each biller carries more accounts by approving and fixing drafts instead of starting every denial from scratch. Whether you hire fewer people as you grow is your call; we will not promise headcount cuts.
Which AI tools can we use with patient data?
Only plans covered by a business associate agreement. As of October 2026, Anthropic offers one for Claude Enterprise (with the HIPAA setting on) and for its API after a BAA is signed; a third-party summary reports that OpenAI's consumer and self-serve plans, including self-serve ChatGPT Business, do not include one. Terms change, so we check each tool's current terms with you in writing before anything is built. This is not legal advice.
What software does it work with?
It reads exports from your clearinghouse and the billing or practice-management systems your clients use, and puts its output where your billers already work, such as a work queue or a shared sheet. Some systems and payer portals limit automated access, so we confirm each one during the free read.
AI billing companies are raising big rounds and one advertised 1.8% to practices. Why work with you?
Most funded AI revenue-cycle companies sell to practices directly, and some are buying billing firms. Lunan works for your firm, inside your tools and on your data, so you keep your clients and the improvement. Lunan is new to billing, so we start every engagement with a measured baseline rather than promises.
Do you build AI phone agents for payer calls?
No. Several vendors already sell AI agents that call payers for claim status. We check electronic status through your clearinghouse first, help you choose a voice vendor if you need one, and connect its results to your queue.
Reading list
- Will AI Replace Medical Billing Companies? An Owner's Honest Read
An owner-level read of how AI pressures billing companies in 2026 and five moves to make this quarter.
- AI Denial Triage and Appeal Drafts: A Setup Where Billers Approve Every One
A five-step denial queue a billing company can build on its own data, with the rules that keep billers in charge.
- Which AI Tools Can a Billing Company Use With Patient Data? A 2026 BAA Checklist
A plan-by-plan BAA check for Claude and ChatGPT and a ten-line AI policy a billing company can adopt.
- Build, Buy or Partner: AI for Claim Status, Payer Calls and Denials
Which AI billing work to build, which to buy, and why payer phone calls are a partner decision, with vendor pricing as published.
- Accounts per Biller: The Number to Measure Before You Spend on AI
How to baseline accounts per biller, weigh AI costs against freed capacity, and decide whether to build now.
Find the hours hiding in your denial queue
Start a free conversation. You get a short written plan for your first AI work queue and the two-week baseline to measure it against.