The Free Sample Audit
One practice. One month of remittances, or one payer's denied and underpaid lines. We audit that slice, a certified coder reviews every finding, and you receive a findings report within five business days of complete data arriving. No fee. No obligation to continue.
Deliverables
The deliverable is a single findings report, built to be put in front of a payer or a partner without embarrassment. This is exactly what is in it:
| Deliverable | What it contains |
|---|---|
| Contract comparison | Every paid line in the sample compared against your contracted rate for that payer. Lines paid below contract — beyond the tolerance threshold — are flagged with the expected amount, the amount actually paid, the variance, and the contract evidence behind the finding. |
| Denial classification | Every denial in the sample classified by reason code and appeal deadline, so you can see which denials are still recoverable, which are urgent, and which have already aged out. |
| Orthopedic-specific checks | Modifier 25/59/X usage, 90-day global-period overlaps, laterality, and implant and hardware reimbursement patterns — the places generic reviews don't look, and where ortho money hides. |
| Benchmark variance | For any payer without a contract rate on file, lines are compared against a Medicare baseline and labelled benchmark variance — flagged as a question to investigate, never claimed as an underpayment. |
| Findings report, ranked | Each finding with expected vs. paid, the supporting evidence, the recoverable amount, and its deadline — ranked so the largest, most time-sensitive recoveries come first. Every finding passes certified-coder review before the report is delivered. |
| Review call | A short call to walk the findings, answer questions, and measure the slice against our stated bar: at least 10× the would-be fee in credible recoverable revenue. |
What the sample audit does not include: appeal drafts.
The free deliverable is the findings report, so the free work stays bounded and expectations stay clean on both sides. Drafted appeals are part of the paid 90-day pilot that follows a successful sample. And to be direct: the sample audit does not guarantee recoveries — payers decide appeals. What it guarantees is an honest measurement of what is sitting in the slice you send.
Your Part
Exports you already have. No new software, no integration project, no staff time beyond pulling the export and signing one agreement.
| Item | Detail | Status |
|---|---|---|
| One slice of claim data | Option A: one month of claim and remittance exports (837/835 files), or Option B: one payer's denied and underpaid lines. Whichever is easier to pull is the right one. | Required |
| BAA — or de-identified data | Data travels one of two ways: fully de-identified, or under a signed Business Associate Agreement over a secure transfer. No BAA and no de-identification, no audit — that rule protects you as much as it protects us. | Required, before transfer |
| Contract rate schedule | The practice's contracted rates for the payer(s) in scope, so paid amounts can be compared against what was actually agreed. Without it, the audit still runs — against a Medicare baseline, with findings labelled benchmark variance. | Optional — strengthens findings |
Data is used only to perform the audit: never sold, never shared outside the engagement, never used to train models. All work runs on HIPAA-eligible, BAA-covered infrastructure, encrypted in transit and at rest, with role-based access and audit logging.
Tolerance Discipline
A line only flags when it was paid below the expected amount by more than a set tolerance — the greater of $1 or 2%. Deductibles the payer applied correctly, and variances inside tolerance, are not findings. Inflating a report with noise is how audit vendors lose credibility, so the engine is built not to.
Every finding is then reviewed by a certified coder before it reaches the report. The engine proposes; a qualified human signs off. If the sample slice is genuinely clean, the report will say so — that is a useful answer too, and it costs you nothing.
Before touching any real book, the engine was scored against a synthetic test dataset: 22 claim lines with errors planted on purpose. It flagged $1,580 in recoverable revenue with zero false positives — and the two lines that matter most are the ones it left alone: a deductible the payer had applied correctly, and a line paid $3 under, inside tolerance.
These are prototype test results on fictional claims — not client results, and not a projection of what any practice will recover. We publish them because the discipline they demonstrate is the product.
Timeline
After the Sample
Billing companies can run this exact sample on one of their clients, under their own brand — see the partner program.