RCI for Payers & Insurance

Claims validation, coding verification, and adjudication intelligence

Claim denials create friction for everyone — providers rework, patients delay care, and your team spends time on appeals that could have been prevented. Nearly 2 in 5 people who challenge a bill get it reduced or eliminated, which means many denials shouldn’t have happened in the first place. RCI gives your adjudication systems the billing rules and coding context to make accurate decisions faster — reducing preventable denials at the source.

What RCI does for you

Claims adjudication

Task	How RCI helps
Validate billing form	Verify the provider used the correct form (CMS-1500 vs UB-04) for the facility type
Check place of service	Confirm the POS code matches the facility type and care setting
Verify modifier usage	Check that modifiers are appropriate for the service and setting
Validate coding	AI agent verifies CPT code matches the documented procedure
Check medical necessity	AI agent evaluates whether the diagnosis supports the procedure against NCD/LCD criteria
Calculate expected payment	RVU × GPCI × CF gives you the expected Medicare payment for benchmarking

Provider network management

Task	How RCI helps
Fee schedule validation	Compare contracted rates against MPFS-calculated expected payments
Geographic adjustments	GPCI data by facility location for accurate geographic pricing
Facility classification	CCN-based facility type verification (is this really a CAH? a SNF?)

Example: Automated claims review

A claim comes in for CPT 99214 at facility 170001, billed on a CMS-1500 with POS 22.

Validate against knowledge

curl -X POST https://api-dev.rcintell.com/v1/knowledge/resolve \
  -H "X-API-Key: kp_test_..." \
  -H "Content-Type: application/json" \
  -d '{
    "ccn": "170001",
    "cpt": "99214",
    "payer": "Medicare",
    "care_setting": "outpatient"
  }'

RCI returns the complete billing context. Your system can now automatically check:

Claim field	Expected (from RCI)	Claim value	Result
Billing form	CMS-1500	CMS-1500	Pass
POS code	22	22	Pass
Payment system	MPFS	—	Benchmark: $107.42
Rate type	Facility	Facility	Pass

Validate medical necessity

When reviewing a claim for a high-cost procedure:

curl -X POST https://api-dev.rcintell.com/v1/agents/medical-necessity \
  -H "X-API-Key: kp_test_..." \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Does M17.11 support medical necessity for CPT 27447?",
    "cpt_code": "27447",
    "dx_codes": ["M17.11"],
    "clinical_notes": "Patient records indicate severe osteoarthritis with failed conservative treatment."
  }'

RCI’s AI agent evaluates the claim against NCD/LCD criteria and returns an assessment with confidence scores — audited by SENTINEL.

Adjudication pipeline integration

Claim received
    │
    ├─ RCI: Validate facility type (L2) — correct billing form?
    ├─ RCI: Validate care setting (L3) — correct POS code?
    ├─ RCI: Check payer rules (L4) — timely filing met?
    ├─ RCI: Verify coding (L5) — appropriate code for service group?
    ├─ RCI: Calculate expected payment (L6) — within fee schedule range?
    │
    ├─ If flagged → RCI: Medical necessity check (agent)
    │
    └─ Decision: Pay / Deny / Pend for review

Bulk validation

For processing claims at scale, call the REST API directly with your pipeline. Each call resolves in milliseconds — no agent overhead for rule-based checks.

import httpx
import asyncio

async def validate_claim(client, claim):
    resp = await client.post("/v1/knowledge/resolve", json={
        "ccn": claim["facility_ccn"],
        "cpt": claim["cpt_code"],
        "payer": "Medicare",
        "care_setting": claim["care_setting"],
    })
    knowledge = resp.json()

    expected_form = knowledge["layers"][1]["data"].get("billing_form")
    expected_pos = knowledge["layers"][2]["data"].get("pos_code")

    errors = []
    if claim["billing_form"] != expected_form:
        errors.append(f"Wrong billing form: expected {expected_form}")
    if claim["pos_code"] != expected_pos:
        errors.append(f"Wrong POS: expected {expected_pos}")

    return {"claim_id": claim["id"], "errors": errors, "knowledge": knowledge}

What payers care about in each layer

Layer	Adjudication value
L1 Location	GPCI values for geographic payment calculation and fee schedule benchmarking
L2 Facility	Verify the facility type matches the billing form and payment system
L3 Setting	Confirm POS code and rate type (facility vs non-facility) are correct
L4 Payer	Your own rules reflected back — timely filing, appeal deadlines, prior auth
L5 Service Group	Coding rules for the service category — documentation requirements, bundling rules
L6 Service	MPFS-calculated expected payment as a benchmark against contracted rates