If you've spent a career inside legacy AdminConsole / ClientConsole tooling, the questions are very specific: how do rules work, how do EDI flows recover, what's seeded for my state. This page is the short index into the answers, with deep links to the surfaces.
PRE_SUBMIT_VALIDATE rule artifacts. Content-
addressed YAML, scoped against the 7+1 dimensional matrix (org,
site, facility, billing entity, payer, program, service line,
state), with effective dating + a paste-a-claim simulator that
returns the winning rule and the also-rans. The legacy
usp_ScrubClaim_<Payer> procedure becomes a YAML
file you can diff.
Replay surface with DLQ + exponential backoff + idempotent re-submit. Per-row failure reason, retry count, audit log per replay. "We lost a batch" used to be an incident; on the new platform it's a row in a table with a button.
See the EDI engine →Five-trigger modifier injection (credential, POS, time, telehealth, group) plus a modifier validation rule kind with its own simulator. Per-payer precedence is explicit, not buried in T-SQL. The state-rate-code and HA-HZ / U1-UD modifier catalogues are master-side and quarterly-refreshed.
See modifier rules →Trading-partner directory with a capability matrix per partner. 270/271, 276/277, 277CA, 278, 834, 835, 837P, 837I, 999, TA1 — each capability enforced at config time. Credentials live in Key Vault, not on the application surface. Companion guides are Monaco YAML + live 837/270 preview.
See the transaction inventory →Nine first-class program categories; deep seed for OH, NY, NJ, PA; rate-code seed for TX, CA, NY, FL, PA. Codes, modifiers, group-size rate bands, EVV configuration are all configuration-layer, not application-layer. Adding a category = one canonical_program row + state-scoped YAML.
See program coverage →Quarterly + annual ingestion harness on the master-data roadmap. HCPCS Level II quarterly, ICD-10 annually, POS / TOB / Revenue / modifier catalogs already live. NCCI PTP + MUE seeded for BH today, full CMS on the roadmap (which removes the $5K-$15K/year/customer commercial-scrubber spend).
See master data →Auto-correction handler registry + per-CARC success rate views. A materialised view exposes success rate per CARC; new handlers are TypeScript functions registered against a CARC, not stored procedures. Phase two will close the loop with an autonomous agent that proposes corrections, dry-runs them, and queues human-approved re-submits.
See AI-first development → Versioned YAML mappings + visual editor + Monaco +
diff + dry-run. Per-customer SSIS / Talend pipelines
become per-customer YAML files in /admin/ingestion/mappings.
Roll back by re-publishing a prior version; no service window.
Everything an analyst would have asked back-office engineers to do in the legacy system is here.
None of the legacy categories of work go away — claim scrubbing, modifier injection, EDI dispatch, denial categorisation, EVV compliance, fee schedule maintenance. What changes is who does the work and how it's expressed.
