The billing department your Tebra practice never had.
Small practices run Tebra with one or two people covering everything. A private AI system takes the repetitive half — eligibility, scrubbing, denial categorization, A/R follow-up — so those one or two people spend their day on the exceptions that actually need judgment.
Signed BAA before any PHI moves. Your data stays yours.
Small practices don't lack effort. They lack the hours — and the backup when one person is out.
One person is the whole billing department
When that person is out, claims stop going out. There is no bench and no documented process, only habit.
Small denials get written off
A $70 denial is not worth an hour on hold, so it dies. Across a year that is real money quietly left on the table.
Enterprise AI quotes make no sense here
Six-figure builds and per-seat AI add-ons do not fit a five-provider practice. You needed a flat number you can budget.
Our position: a second set of hands for your biller, human approval on everything, one flat number a small practice can actually budget.
The repetitive half of billing, handled — your people keep the judgment calls.
One workflow at a time, phased in. Your biller approves every action. A simple live dashboard shows clean-claim rate, denial rate, days in A/R, cash collected, and hours saved each month.
Denials & rejections
PILOTWorked when there is time; many written off
Every denial categorized the day it arrives, recoverable dollars ranked, appeal letters drafted for your biller to approve
Eligibility & benefits
PHASE 1Checked manually, often day-of
Every scheduled visit verified 48 hours ahead; problems routed to your front desk
Claim scrubbing
PHASE 1Payer rules live in billers' heads
Every claim checked against payer-specific rules, modifiers, and coding edits before it leaves
A/R worklist
PHASE 1Oldest-first, or whoever remembers
Daily follow-up list ranked by dollars, age, payer, and likelihood of payment
Claim status & posting exceptions
PHASE 2Portal checks one claim at a time; mismatches found late
Automated status checks; unposted or mismatched remittances flagged for review
Coding review & prior auth
PHASE 2–3Coder memory and document-by-document packet assembly
Charges compared to the note; auth requests drafted and tracked
Reporting
PILOTMonth-end, by hand out of Tebra reports
Live dashboard: clean-claim rate, denial rate, days in A/R, cash collected, hours saved
Nothing runs on its own until you say so. The system prepares; your team approves. Every action is logged. No patient-facing calls or messages are part of this program.
No plugin, no new login — Tebra stays exactly as it is.
We work from Tebra exports and your clearinghouse feed. Approved appeals and follow-up notes come back as files and drafts your biller submits in Tebra.
Tebra
- Scheduling, charges, claims, and patient statements stay in Tebra.
- Data leaves through Tebra exports and your clearinghouse feed.
Standard interfaces
- Claims & remits (837 / 835) via your clearinghouse
- Eligibility (270 / 271) · claim status (276 / 277)
- Tebra report exports (CSV) · FHIR R4 where available
- Secure file drop for anything else
Lonestar private AI
- Your dedicated environment — private cloud or on-prem.
- Reads, checks, categorizes, ranks, drafts. Learns your payers and rules. Every step logged.
Your billing team
- Reviews a ranked worklist and drafted appeals.
- Approves, edits, or rejects. Nothing is sent without a person.
No plugin, no new login for your staff. Approved appeals and follow-up notes come back as files and drafts your biller pastes or submits in Tebra.
Simple guardrails a one- or two-person team can actually enforce.
Human approval, always
For the first 90 days no claim, payer, or patient is touched without a person clicking approve. After that, you decide what earns autonomy.
Back office only
No patient-facing calls, texts, or messages are part of this program. We make sure you get paid for work already done.
Your data, your control
Signed BAA before anything moves. Dedicated private environment or a server in your office. Full export in standard formats whenever you want it.
Full audit trail
Every read, suggestion, edit, and approval is logged with a user and timestamp, retained for six years.
Controls we hold ourselves to: encryption at rest and in transit, VPN-only access, MFA, per-user audit logs, PHI never sent to a model endpoint without a BAA, and six-year audit-log retention. More on security.
No upfront cost. You pay nothing until you've seen results on your own data.
Prove it on your data
- 90 days of denial and A/R aging exports — files only, no Tebra access
- Recoverable dollars identified and ranked
- Appeal drafts reviewed by your own biller
- Live KPI dashboard delivered
- No integration, no invoice, no obligation
Deploy
- Denials → eligibility → scrubbing → A/R worklist, one at a time
- Your team approves everything
- Private environment stood up under BAA
- Weekly check-in with your billing lead
One flat monthly fee
- Hosting, security, monitoring, backups included
- Payer-rule, model, and Tebra changes handled by us
- Enhancement hours every month
- Monthly value report: dollars recovered, hours saved
12-month initial term, then month-to-month with 60-day notice. Implementation-rate pricing for the first two months while workflows go live. Exit any time with a full data export in standard formats within 10 days.
One number on the invoice.
What a typical AI project costs a practice versus what you pay us.
Typical AI vendor
- Large upfront build fee, paid before anything works
- Cloud bill (AWS / Azure) in your name — variable, unbudgeted
- Model subscriptions the developer chose, billed to you
- Security, backups, monitoring: your responsibility
- Model gets deprecated; nobody updates the system
- Data locked in; leaving is painful
Lonestar flat fee
- No upfront fee — the 14-day pilot is free
- Hosting, GPU, and model costs are inside our fee
- Security, backups, monitoring, on-call: ours
- Payer-rule, Tebra, and model updates every month
- Monthly report of dollars recovered and hours saved
- Your data exports in standard formats, any time
The fee is set by practice size and claim volume, and sits below the cost of adding billers. We quote after the pilot, based on what the pilot measured. See plans.
A one-time build stops working by month twelve.
The system that works in month one keeps working only if someone keeps it working.
Payer rules change every quarter
New edits, modifiers, and documentation requirements. We update the rules; your denial rate does not creep back up.
Tebra and clearinghouses ship updates
Screens, exports, and interfaces change. We absorb the breakage before your team notices.
AI models get replaced
Models are deprecated and replaced on months-long cycles. We swap and re-validate; you never see it.
Your practice changes
New providers, new payers, new specialty codes. Included enhancement hours cover it.
Same model as every software subscription you already pay for — except this one is built for your practice.
Your choice: dedicated private cloud or a server in your office.
Dedicated private cloud
- Isolated environment for your practice only — no shared tenants
- HIPAA-eligible services under a signed BAA
- Encrypted at rest and in transit; VPN-only access; MFA
- Fastest to stand up — live during the pilot
- Included in the monthly fee
On-premise server in your office
- Single AI server installed on your network
- PHI never leaves the building
- One-time hardware cost (purchased or leased); hosting portion drops out of the fee
- We maintain, patch, and monitor it remotely
- Best fit for practices with strict data-residency policies
Either way, you can move from one to the other later without rebuilding. Support runs business hours with a 99.5% uptime target and one-business-day fixes on billing-blocking issues.
Three reports. Fourteen days. Then you decide.
Sign mutual NDA and BAA. A 30-minute session with whoever does your billing to export denials (90 days), A/R aging, and your top payers. That is the only commitment.
Free pilot on those files. We present recoverable dollars, appeal-draft acceptance rate, and hours saved per month.
You decide. If the numbers work, we sign a 12-month agreement and start deploying. If not, you keep the analysis.