🏥 MediVerify AI Simplifying Provider Data Validation with Agentic AI 📌 Overview
MediVerify AI is an Agentic AI–based system designed to automate provider data validation, enrichment, and quality assurance for healthcare payers and provider directory management teams.
Healthcare organizations often struggle with inaccurate provider directories, leading to member frustration, regulatory risks, and high operational costs. MediVerify AI addresses this problem by using modular AI agents to validate provider data against public sources, enrich missing information, compute quality scores, and generate actionable dashboards and reports.
This solution demonstrates how intelligent automation can significantly reduce manual verification effort while improving data accuracy and compliance.
🎯 Problem Statement
Healthcare provider directories frequently contain outdated or incorrect information such as phone numbers, addresses, specialties, and license details. Manual validation processes are time-consuming, error-prone, and do not scale efficiently.
MediVerify AI aims to:
Detect inaccuracies in provider data
Automatically enrich missing information
Assign confidence and quality scores
Flag high-risk providers for manual review
Provide dashboards and reports for operational teams
🧠 Solution Approach (Agentic AI)
The system is built using loosely coupled AI agents, each responsible for a specific task:
🔹 Data Validation Agent
Compares provider data with simulated public sources
Computes similarity scores for phone, address, license, and specialty
🔹 Information Enrichment Agent
Enriches provider profiles using a mock NPI registry
Adds qualification, hospital affiliation, and experience
🔹 Quality Assurance Agent
Detects missing data and inconsistencies
Calculates final quality score
Categorizes providers as Good, Review, or High Risk
🔹 Directory Management Agent
Displays insights via an interactive Streamlit dashboard
Generates downloadable PDF reports
Highlights providers requiring manual attention
🏗️ Architecture Overview Provider Data (CSV / PDF) ↓ Data Validation Agent ↓ Information Enrichment Agent ↓ Quality Assurance Agent ↓ Directory Management Agent ↓ Dashboard + PDF Reports
📊 Key Features
Case-insensitive CSV upload validation
Automated quality score calculation
Multiple visualizations:
Bar Chart (Provider Quality Scores)
Histogram (Score Distribution)
Pie Chart (Quality Status Breakdown)
Heatmap (Similarity Metrics Correlation)
PDF report generation
Edge-case handling for missing or incomplete data
🛠️ Tech Stack Layer Technology Programming Python Data Processing Pandas, NumPy AI / Matching Fuzzy Matching (difflib) Dashboard Streamlit Visualization Matplotlib Reports ReportLab Data Sources Mock NPI Registry, Simulated Public APIs 📂 Project Structure MediVerify/ │ ├── agents/ │ ├── validation_agent.py │ ├── enrichment_agent.py │ ├── quality_assurance_agent.py │ └── directory_management_agent.py │ ├── data/ │ ├── providers_sample.csv │ └── mock_public_sources.json │ ├── reports/ │ ├── validated_providers.csv │ ├── enriched_providers.csv │ └── quality_checked_providers.csv │ └── README.md
2️⃣ Run the Dashboard streamlit run agents/directory_management_agent.py
3️⃣ Upload CSV
Upload a provider CSV file
Click Load Dashboard
View insights, charts, and download PDF report
Missing phone numbers or addresses → flagged for review
Incomplete provider records → lower quality score
Inconsistent data → automatically detected
Empty or invalid CSV → blocked with clear error messages
📈 Business Impact
Reduces manual verification effort by 60–80%
Improves provider directory accuracy
Enhances regulatory compliance
Scales easily for large provider datasets
🚀 Future Enhancements
Integration with real NPI and Google Maps APIs
Automated email notifications to providers
Support for unstructured PDF extraction
Role-based dashboards for compliance teams
👩💻 Author
Thejaswini Techathon 6.0 MediVerify AI – Agentic Healthcare Data Quality Solution