Evidence you can defend when the number gets audited
We help providers and payers navigate coverage and medical codes for payment integrity and reimbursement. Our lab is Hammer Labs.
Our view of the market
Providers and payers are both getting squeezed for cash. Cost is outpacing revenue on the provider side, and the rules that decide what gets paid are moving faster than anyone can track. Everybody needs a better reimbursement strategy, and almost nobody has trustworthy data underneath the one they already have.
The reason is structural, not organizational. CMS rebuilds its coverage database in place with no changelog. Payers republish their rate files monthly and nothing requires the prior version to stay available. So the number a strategy was built on frequently cannot be retrieved by the time it has to be defended. Between two of our own dated extracts, 22 coverage policies were revised.
How we do it
Deep research into the source record itself: the CMS coverage record, national and local coverage determinations, coverage articles and the published comment history; state Medicaid policy across 28 states; commercial payer published rate files; Medicare physician and lab fee schedules; NCCI procedure-to-procedure edits; and hospital price transparency filings.
We preserve every source with its vintage, so an answer holds up a year later when someone asks which version applied on the date of service.
What we bring to healthcare
- Coverage Intelligence
- The living map of where a healthcare product gets paid, kept current. Read more →
- GAVEL
- The payer rulebook as software: will it be covered, what will it take, what will they pay, what did they just do. Read more →
- Agent workforce
- Agents that check enrollment, find notices and follow up on payment through the portals your staff already use. Read the case study →
- Studio engagements
- Claims edits, prior authorization, clinical policy review: the workflow owned end to end. Read more →
- World models
- Agents tested against the rules, with published refusal classes. Read more →
- Genesee
- The benchmark for payment-integrity agents: a synthetic hospital book with every leaked dollar labeled, no PHI. Read more →
Track record
Careers spent building AI and data systems inside these institutions, not a team that learned the domain in order to sell into it.
36M documents a quarter
Clinical policy review taken from one analyst, one document, one day, to a review backlog at zero.
CMS and 28 states
Medicare plus state Medicaid programs, read at the source.
25 people, from zero
An AI organization built from the first hire at a national payment integrity platform.
Two patents
On the methods that made that scale possible.
What it looks like for one specialty
Take an orthopaedic and spine business. What does each payer actually publish for 27447, 27130, 22633 or 22551, in which market, on which date, and can you still prove it next year? We answer that, and we refuse to answer past our evidence.
What we refuse to say
- No unsupported denial rates. If the record does not carry it, no figure is served.
- No payment probabilities. What a payer published is not what it will pay on a claim.
- No claim that a payer covers a code. Coverage is reported with its citation and its date.
That is what makes it usable where the number gets audited. The refusal classes are published by our lab ↗.