Measured, with the method attached.
Benchmarks with their denominators, plus simple math on what denials cost you.
About 96 in 100 land inside their range: the measured 95–97% coverage.
What 95–97% coverage looks like. 1. Each claim gets a predicted payment range before it goes out. 2. Payers pay. Each dot is the amount actually paid. 3. About 96 in 100 land inside their range: the measured 95–97% coverage. Illustrative, synthetic data.
Prediction figures were measured out-of-fold on one urgent-care provider’s ~5-year de-identified claims history (17,463 insurance lines); the model calibrates to each provider’s own payers over time, so they are not a guaranteed rate. Grouper throughput: 20,000 claims on a single 2-vCPU container, zero errors.
What denials cost you today.
Plain arithmetic on your inputs. Not a VAERION forecast.
Use your measured rate, by dollars if you can.
Revenue × denial rate. Not every denial is lost or preventable; we’ll size what’s recoverable from your remittances.
Where the numbers come from.
No denominator, no number.
Out-of-fold evaluation
Scored on unseen claims: 17,463 lines and 41,143 adjudication records, no PHI.
One provider, labeled as one
One urgent-care provider’s five-year history. The model calibrates to your payers.
The grouper is the reference
DRGs match Medicare by construction. We measure speed: p95 4.2 ms, zero errors in 20,000 claims.
No targets as results
Customer results will appear with cohort, interval and denominator once measured.
The screens behind the result.
About the numbers.
Are these customer results?
Will we get the same accuracy?
How fast do results show?
What’s recoverable for us?
See it on your claims.
Bring a de-identified sample. See which past denials it would have flagged.