Feature · Claim scrubber

Catch the denial before the payer does.

Every claim passes through a deterministic rule engine — CPT/ICD crosswalks, modifier checks, payer rules, duplicate detection — and then an LLM risk pass for context only humans usually catch.

How it works

Rules first. Models second. Audit third.

We never gamble compliance on a black box. Rules block the obvious denials; the LLM only adds risk scoring and natural-language explanations on top. Every decision is logged.

  • Deterministic rule engine
    60+ built-in rules and a per-tenant rule editor.
  • Crosswalk validation
    CPT ↔ ICD-10 plausibility, NCCI edits, modifier compatibility.
  • LLM risk pass
    Bring your model. We score risk and explain in plain English.
  • Recommended fixes
    Not just a flag — a suggested edit your biller can accept in one click.
app.mediq.ai / claims / CLM-001 274
Live
Claim
CLM-001274 · $642.00
Payer
Aetna Commercial
First-pass accept
98.4%
Mean A/R
11d
Denial rate
-42%
Cost / claim
$0.18
AI risk score
Claude Sonnet 4.5
42
/ 100 likely to pay first pass
  • Eligibility verified · plan active
  • CPT 99214 ↔ ICD E11.9 crosswalk passes
  • Modifier 25 valid given E/M + procedure
  • !Prior auth recommended within 24h
Appeal draft (if denied)

Per ACA §2719 and plan SPD, the rendered service is medically necessary documentation

What it catches

The 80% of denials that never had to happen.

Missing or invalid modifiers

Modifier 25, 59, 51 — applied (or removed) based on payer-specific rules.

ICD/CPT incompatibilities

Crosswalks flag procedure-diagnosis pairs payers won't honor.

Duplicate claims

Same patient, same DOS, same CPT — blocked before submission.

Eligibility & authorization

Surface stale eligibility or missing prior-auth before the claim drops.

Documentation risk

LLM flags claims likely to face medical-necessity review based on note quality.

Payer-specific rules

Layer your own rules per payer; we ship a starter library out of the box.

See your top denial reasons disappear.

A 14-day pilot runs the scrubber side-by-side with your current workflow. You keep the playbook.