The Client
One of the biggest insurance brokers in the United States, has over 25 years of experience. They offer services at over 40+ locations that include auto, motorcycle, boat, life, health, pets, and other insurance services.
As an insurance broker, our client also conducts underwriting services for numerous insurance providers across the US.
Challenges
Our client faced significant challenges in their underwriting process, which required manual verification of multiple documents, consuming time and leaving room for errors.
- Manual verification in lengthy sales cycles often led to inaccuracies and customer loss.
- Needed to digitize documents like FNOLs, bank statements, and credit scores.
- Required cross-verification of handwritten or printed documents without skipping any details.
- Sought to process multiple applications simultaneously, 24/7, without human intervention.
- Demanded systematic data storage without duplication and timely sharing of verification status.
Solution
Maruti Techlabs developed a Python-based OCR model to automate underwriting:
- Data Extraction: Extracted key data from handwritten or printed documents using OCR.
- Preprocessing: Used Tesseract to process input images, detect edges, and outline characters.
- Text Detection: Deployed YOLOv5 for real-time object detection to mark character areas.
- Character Recognition: Implemented PyTorch for recognizing and digitizing characters.
- Error Handling: Flagged documents with inconsistencies for manual review, while unflagged ones proceeded for underwriting.
The model was thoroughly tested with client-provided data and deployed on their online platforms.
Results
The OCR model transformed the client's underwriting process:
- Increased data accuracy and reduced duplication through document classification.
- Automated tasks, cutting processing time by 40%, enabling focus on high-priority cases.
- Improved security with risk-scoring and fraud insights.
- Enhanced customer experience with an end-to-end claims process.
The solution optimized underwriting and delivered measurable value through AI-driven automation.
