LTC Claims Processing Automation
InsuranceBusiness Impacts
Cost Savings of ~89% Compared to the Current Process
73% Accuracy for Extraction of Predefined Entities
Reduce ~2500 Hrs of Manual Processing Time per Year
Additional Field Data Extracted Forms the Foundation for AI/ML Use Cases
Customer Key Facts
- Location : North America
- Industry : Insurance
Problem Context
The client is a Fortune 500 insurance holding company in the business of mortgage insurance and long-term care insurance. They sought to automate the extraction of information from Accounts Payable and Claims invoice documents, a currently manual process where the team reviews an invoice, extracts relevant information, and uploads it to a structured database.
Challenges
- Large number of documents to index, classify and review manually
- Poor quality of scanned documents coming from multiple sources like Email, fax, mail
- Current workflow does not check for invoice duplication and Policy max-outs from core systems
Technologies
Cloud Vision API
Cloud Functions
Cloud Storage
App Engine
TensorFlow
Cloud Firestore
Node.js
Python
Solution
Quantiphi developed an automated OCR and entity extraction solution leveraging Document AI & Cloud Vision APIs to extract predefined entities from AP Invoices. We also built a lightweight user interface(UI) to upload individual/bulk documents & visualize the extracted data.
Result
- Secure and compliant environment to handle sensitive medical documents
- Ability to identify invoices that can be processed without human review
