Deterministic coding, AI-assisted, audit-grade.
Every claim flows through 13 deterministic stages backed by the same national code catalogs your auditor uses — then a governed AI layer suggests reversible patches you can accept in one click.
Thirteen stages. One auditable trace.
Rules execute deterministically and in order. The AI layer never overrides a rule — it only proposes patches that can be safely auto-applied when they meet your tenant's confidence threshold.
- • Reversible patches with field-level diffs
- • Per-tenant confidence threshold (Conservative → Aggressive)
- • Every decision written to the audit log with rule + rationale
- 1Catalog resolveEvery code on the claim is normalized against the active national catalog (ICD-10-CM, CPT, HCPCS, Revenue, POS).
- 2Crosswalk validationCPT ↔ ICD plausibility checked against the platform-managed crosswalk.
- 3Modifier compatibilityModifier 25, 59, 50, RT/LT and payer-specific modifier rules applied.
- 4NCCI editsCMS National Correct Coding Initiative procedure-to-procedure edits, including Column 1/2 logic.
- 5MUE limitsMedically Unlikely Edits — per-line maximum units enforced before submission.
- 6Bundling & global periodsGlobal surgical packages, post-op windows, bundled labs and supplies flagged.
- 7Place-of-service validityPOS / setting compatibility against the procedure (e.g., facility vs. office).
- 8Payer-specific rulesYour contracted payer overrides — prior-auth required CPTs, frequency, age/sex edits.
- 9Facility editsFor institutional claims: Type of Bill, MS-DRG, Revenue code validity, ancillary requirements.
- 10Duplicate detectionSame patient + DOS + procedure detection across submitted claims, with override workflow.
- 11Documentation floorHeuristic check that note length and structure supports the billed E/M level before AI refinement.
- 12Fee schedule / pricingLines compared against contracted allowables; under-pricing and over-pricing both flagged.
- 13AI refinementGoverned AI reviews remaining risk, suggests reversible patches, and writes the rationale to the audit log.
The national code sets you need, kept current, on every plan.
MedIQ centrally licenses and maintains the catalogs that drive correct coding — you never re-pay for them, never re-import them, and never run on stale data.
Full diagnosis catalog with annual updates, includability flags, and short/long descriptions.
Procedure codes with descriptors, RVUs, and modifier guidance — refreshed on the AMA cycle.
Supplies, DME, drugs, and non-physician services with category and coverage flags.
UB-04 revenue codes, TOB validity, ancillary requirements for institutional claims.
CMS-published procedure-to-procedure edits and Medically Unlikely Edits, updated quarterly.
Payer-specific modifier rules — when 25, 59, 50, X{EPSU} apply (or don't).
POS codes with compatibility rules against the billed CPT.
Bring your contracted allowables; the pricing stage compares line totals against them.
Tenant admins attest to required code-set licenses (e.g., AMA) — tracked in the admin console.
MDM floor first. AI refinement on top.
MedIQ derives the medical decision-making floor for an E/M visit from documented problems, data, and risk — deterministically. Governed AI only refines from that floor upward, never below it, and every change is logged with the rationale.
- Documentation-supportedIf the note doesn't carry the level, we don't bill it.
- Conservative-by-defaultAuto-apply patches only run when above your tenant's confidence threshold.
- ReversibleEvery patch can be undone with a single click; the original is preserved.
- 99213 · Office visit, established
- 36415 · Venipuncture (no Mod 59)
- J3301 × 4 · Triamcinolone (MUE: 2)
- 99214 · Office visit (MDM floor: moderate)
- 36415 -59 · Venipuncture
- J3301 × 2 · Triamcinolone (within MUE)
See the pipeline trace on your claims.
We'll seed a sandbox with your top procedure codes and walk through the 14 stages on a real day's worth of charges.