Skip to content
DINVEXA
Healthcare SaaS & Artificial Intelligence Verified Scope
Back to All Projects

Clinic SaaS — Insurance AI Decision Engine

A healthcare decision-support system that processes insurance eligibility responses, organizes source evidence, normalizes terminology, and applies deterministic business rules to support office-visit coverage decisions.

Clinic SaaS platform preview
Technical Accuracy & Review Notes
  • Internal enterprise deployment: Public URL withheld to protect client confidentiality.
  • Important: System does NOT make medical diagnoses; it supports administrative insurance coverage decisions only.
  • Deterministic business rules are strictly authoritative; LLMs do not adjudicate final coverage.
  • Zero protected health information (PHI) or proprietary client credentials are exposed.
01Executive Overview

Background & Business Context

Clinic SaaS is an intelligent healthcare decision-support platform engineered to solve one of administrative healthcare's most complex challenges: evaluating insurance eligibility responses for office visits. The platform processes semi-structured insurance eligibility data (such as X12 271 responses), extracts and normalizes clinical terminology, retrieves relevant policy evidence, and applies deterministic, fully auditable business rules to determine coverage eligibility.

02The Challenge

Operational Bottleneck & Goals

Healthcare clinics lose hundreds of administrative hours manually interpreting dense, non-standardized insurance eligibility documents. Human interpretation errors lead to claim rejections, surprise patient bills, and delayed care. Conversely, relying purely on generative AI is clinically and legally unacceptable because generative models can hallucinate coverage policies.

03Engineering Execution

Solution & Architecture

Dinvexa engineered a hybrid decision-support architecture where AI and deterministic rules play strictly separated roles. Large Language Models (Google Gemini via LangGraph) are utilized strictly for semi-structured text parsing, candidate evidence extraction, and terminology normalization against medical ontologies. Final coverage decisions are executed by an authoritative deterministic rules engine, guaranteeing 100% reproducibility, zero hallucinations, and comprehensive decision traceability.

04Deliverables

Verified Key Features

Insurance eligibility response processing (automated extraction from semi-structured EDI / benefit data)
Source evidence extraction and normalization across multi-carrier formats
Terminology resolution mapping complex clinical terms to standardized codes
Evidence candidate classification and semantic matching
Knowledge-base retrieval powered by pgvector embeddings
Deterministic office-visit decision rules (Authoritative engine with zero LLM override on final verdicts)
Complete decision traceability with human-in-the-loop audit validation
Healthcare workflow integration designed for clinical administrative desks
05Interface Gallery

Visual Modules & Layouts

Module views and production interface captures for Clinic SaaS.

Eligibility Decision Engine & Clinical Console
Module 01Verified Photo

Eligibility Decision Engine & Clinical Console

High-precision decision support interface showing automated EDI data extraction, confidence scoring, and deterministic rule workflows.

Module 02Screenshot Placeholder

Eligibility Parsing & Evidence Extraction Console

Clinical workbench displaying parsed insurance EDI responses with extracted evidence candidates.

Verified Asset Specification

Eligibility Parsing & Evidence Extraction Console

Clinical workbench displaying parsed insurance EDI responses with extracted evidence candidates.

Module 03Screenshot Placeholder

Deterministic Rule Evaluation Engine

Authoritative decision matrix showing rule match status, copay verification, and validation trace.

Verified Asset Specification

Deterministic Rule Evaluation Engine

Authoritative decision matrix showing rule match status, copay verification, and validation trace.

Module 04Screenshot Placeholder

Traceability & Audit Verification Log

Complete chronological audit trail showing raw input, extracted tokens, and final deterministic verdict.

Verified Asset Specification

Traceability & Audit Verification Log

Complete chronological audit trail showing raw input, extracted tokens, and final deterministic verdict.

Project Overview

CategoryHealthcare SaaS & Artificial Intelligence
Scope of ContributionFull-stack architecture, LangGraph state machine modeling, deterministic decision engine design, clinical frontend interface, and HIPAA-conscious data isolation.
Live VerificationInternal deployment / client confidential
StatusPublished in Portfolio

Confirmed Technology Stack

Only verified technologies confirmed against codebases or production sites are presented.

Python & FastAPIVerified

High-throughput asynchronous backend orchestration

LangGraphVerified

Stateful agentic pipeline and extraction graph workflows

PydanticVerified

Strict data validation and schema enforcement

PostgreSQL & pgvectorVerified

Relational persistence and semantic policy retrieval

RedisVerified

In-memory caching and message queuing

Google GeminiVerified

Evidence candidate extraction and text normalization

React & ViteVerified

High-performance clinical decision-support interface

Key Highlights

  • Deterministic business rules remain strictly authoritative
  • LangGraph + Gemini pipeline for evidence extraction & ontology normalization
  • pgvector similarity search with 100% decision audit traceability
Need a Similar Solution?

Let's engineer your next digital platform.

We deliver production-ready software, responsive web portals, and reliable cloud integrations.

Start a conversation

Have an ambitious project in mind?

Whether you are planning a restaurant operations platform, bespoke CMS, high-impact agency website, or intelligent SaaS decision engine—DINVEXA is ready to engineer it.