Delivered engagementAI & Agents · $25k – $35k
Meridian Billing Partners

Healthcare billing

Delivered engagement. Coralsoft designed and built this product end to end; figures below retain their evidence level.

BackendAutomationPostgresDocker / KubernetesLangGraph
agentic-ai-prior-authorization.example / dashboard
Prior authorization queue — per-payer SLA, status, and cycle time
The situation

What was in place before we started.

An autonomous AI agent — acting inside payer portals, generating documents, and tracking statuses, not a chatbot — that runs the full HIPAA-compliant prior authorization cycle for a medical billing company serving 22 outpatient clinics.

The client processes roughly 1,900 prior authorization (PA) requests per month — a mandatory step where a clinic must obtain a payer's approval before performing a procedure, or the claim gets denied. A team of 11 billing specialists manually pulled patient data and procedure codes (CPT/ICD-10) from the EHR, figured out which forms a given payer required (each with its own portal and rules), filled them out, tracked status, and drafted appeals when requests were denied.

The average PA cycle took 3–5 business days, and 18% of requests stalled due to simple form errors or lack of tracking — the clinic often found out about a denial only after the procedure had already taken place.

Constraint

HIPAA compliance: every agent action involving PHI (protected health information) had to be logged, encrypted at rest and in transit, with least-privilege access, and no patient data could leave the controlled perimeter.

Timeline
6 months reconstructed
Team
5 people reconstructed
Budget
$25k – $35k reconstructed
Platforms
2
Stackrecorded
PostgresDocker / KubernetesLangGraphClaude 3.5 Sonnet (HIPAA-compliant)PlaywrightFHIR / HL7Custom rules engineAudit logging & encryption layer
Challenges

Each constraint, and what we did about it.

Constraint

HIPAA compliance: every agent action involving PHI (protected health information) had to be logged, encrypted at rest and in transit, with least-privilege access, and no patient data could leave the controlled perimeter.

What we did

A planner-executor agent (LangGraph) runs a separate state machine per PA request in an “execute → human review on risky steps” mode. Claude 3.5 Sonnet — deployed HIPAA-compliant, with a BAA in place with the provider — interprets clinical notes and drafts medical-necessity justifications and appeals. A FHIR connector reaches the EHR, a browser-automation layer (Playwright) drives payer portals without an API, and a document-generation step produces the PDF forms, with email/portal polling for status updates.

Results

Outcome, with the source of every figure.

84%
of PA requests submitted fully autonomously, with no manual form-filling
self-reportedCoralsoft's own figure for its own delivery, stated in cases.ts. No third-party attestation in the sources.
6.5 hrs
average time from procedure order to request submission (down from 3–5 days)
self-reportedCoralsoft's own figure for its own delivery, stated in cases.ts. No third-party attestation in the sources.
41%
reduction in denials caused by form-filling errors
self-reportedCoralsoft's own figure for its own delivery, stated in cases.ts. No third-party attestation in the sources.
Showcase

How the system fits together.

3 of 3 views
agentic-ai-prior-authorization.example
PA request detail — patient data, CPT/ICD-10 codes, and payer requirements
agentic-ai-prior-authorization.example
Denial and appeal — agent-drafted appeal letter for specialist review
agentic-ai-prior-authorization.example
Agent architecture — FHIR connector, browser automation, and rules engine
Roadmap

How it was sequenced.

6 monthsreconstructed delivery
Weeks 1–4

Discovery & architecture

Pulls data from the EHR (HL7/FHIR interface): patient demographics, insurance plan, the procedure ordered, CPT/ICD-10 codes, and the physician's clinical notes.

  • EHR data pull
Weeks 5–13

Core build

Checks a rules database (updated weekly from payer bulletins) to decide whether PA is required and which forms and supporting documents are mandatory for that specific plan and procedure.

  • Payer requirement engine
Weeks 14–18

Integrations & data

Fills out the payer-portal form via browser automation (where no API exists), attaching a medical-necessity justification the agent drafts from the physician's notes using a legally reviewed template.

  • Request generation & submission
Weeks 19–24

Hardening & QA

Periodically checks the portal/email notifications, updates the clinic's internal system, and notifies the billing team of the outcome.

  • Status tracking
Weeks 25–26

Launch & handover

On denial, analyzes the denial reason, pulls a relevant clinical precedent, and automatically drafts an appeal letter for human review — the final decision to file an appeal always stays with a specialist.

  • Appeal drafting

More work like this.

Every case study is a Coralsoft delivery story, with the evidence behind each figure kept visible.