Findings closed by evidence.
Fraud, waste, and abuse findings are not closed when they are flagged, alleged, or asserted. A disposition closes only when the evidentiary chain supports it — and the chain stays verifiable after it closes.
Closes by evidence — or it does not close.
FRAUD
Intentional deception, surfaced as a finding that must prove itself.
WASTE
Resources spent without result — tracked to the provision that allowed it.
ABUSE
Practices inconsistent with program rules, closed on evidence, not opinion.
The evidence standard
Every finding carries its evidence.
Every closure can be re-verified, link by link, end to end.
The record stands on its own — no trust required.
In a market where manufactured evidence has burned buyers before, this system does not ask to be trusted; it shows the chain. Every disposition on the record can be opened, checked, and checked again — by the reviewer who closed it, and by anyone who answers for it later.
Live today — South Carolina
Figures are the operating South Carolina instance, as of October 12, 2026. The health score of 29.5 is 16.9 coverage + 12.6 confidence. Scope: 523 provisions presented (493 carrying verification sets); verification dashboard of record: 496 provisions, 154 verified, health 29.5. Coverage is of public federal Medicaid regulation — the Code of Federal Regulations and the public sources around it. The CMS Edition is designed for CMS and state Medicaid programs and extends state by state from live instances.
Our story
THE PROVING GROUNDS
The Atlas is the proving ground: an evidence discipline aimed at Medicaid regulation — 499 tracker rows worked end to end, every disposition carried by its evidentiary chain. What holds up there is what ships.
THE STANDARD
One rule for every domain the work touches — federal, Medicaid-state, medical, military, banking: meet the governing standard, and build past it. A standard is claimed only when an outside body grants it.
THE SNOWBALL
The Agentic ❄️ Snowball rolls what the proving ground earned into every next build — each stage proven before the roll widens, replicated across Production, Research, and Quality.
The Machine, Assembling — how this operation grew
The Agentic Armor white paper — the evidence layer, in depth (17.3-P · PDF)
The Standard
Meet — and exceed — the governing standard in every domain the work touches. Where domains overlap, the strictest rule governs the design.
Federal — designed to the CMS ARC-AMPE baseline for Medicaid systems.
Medicaid-state — designed to state Medicaid enterprise requirements, alongside the federal baseline.
Medical — designed to HIPAA safeguards, with HITRUST as the assurance reference.
Military — designed to the DoD Risk Management Framework (RMF) and DISA STIG hardening.
Banking — designed to FFIEC guidance and GLBA safeguards.
Design intent only: these are the standards the system is designed to meet and built to exceed. No certification is claimed for any of them — a standard is claimed only when the granting body grants it.
The proof
Bureau 1, complete — the exemplar bureau is finished: 223 L4 questions and 53 leadership questions authored, then independently re-counted clean.
93% — independent corroboration of the legacy audit chain, disclosed in full, exactly as it came back. The disclosure is the honesty proof.
The retraction — this system caught its own headline figure inflated and retracted it to the defensible number, in public. The record was corrected, not defended.
Scope, stated plainly: the exemplar is one bureau, proven end to end. Nothing beyond it is claimed until it is counted the same way.
The operating model
The Agentic ❄️ Snowball — every stage proven before the roll widens, replicated across the ×3 lanes, under perpetual optimization: each pass is measured against the last, and the standard only moves one way.
PRODUCTION
Builds and ships the Atlas and its editions — the evidence discipline inside the product, not around it.
RESEARCH
Works the regulation ahead of the build — provisions, changes, and the questions each reviewer must be able to answer.
QUALITY
Re-counts, re-verifies, and tries to break the chain before anyone else does — independent of the build it checks.
Capabilities run across the model — compliance evidence systems, the Atlas editions, and a Project Management Wing that runs delivery the same way the Atlas closes findings: on evidence.
Built for the people who answer for program integrity
For CMS and state Medicaid leadership — the directors, reviewers, and program integrity officers who sign their names to findings — the Atlas states each program’s compliance position plainly, with the proof behind every disposition within reach. Designed to meet, and built to exceed, state and federal security standards. Leadership reads a calm surface; the evidence underneath is one step away.