Delivering 97% prescription verification accuracy for largest eyewear company with QDox
RetailBusiness Impacts
60%
reduction in manual efforts
77%
document-level extraction accuracy
97%
field-level data extraction accuracy
4000
prescriptions processed per month
Customer Key Facts
- Country : Canada
- Industry : Retail
Problem Context
The client is one of the largest online contact lens retailers in North America and the largest seller of prescription eyeglasses online worldwide. They receive ophthalmic prescriptions from their end customer as part of the prescription glasses and contact lens purchasing. These prescriptions are used to claim insurance by end customers using the client’s website/application. They must verify the submitted ophthalmic prescriptions, extract data and store it for compliance and claims. They, therefore, required a scalable solution to automate this process of extraction and verification of prescriptions with high accuracy.
Challenges
- Handling a large set of documents in a batch
- Entity extraction in the semi-structured documents
Technologies Used
Amazon Textract
Amazon SageMaker
AWS Lambda
Amazon Simple Queue Service (SQS)
AWS CodePipeline
Elastic Container Registry (ECR)
AWS Identity and Access Management (IAM)
AWS Secrets Manager
AWS Systems Manager
Amazon S3
Amazon Aurora RDS
AWS Codecommit
Solution
Quantiphi deployed QDox, a cognitive document processing solution built on Amazon Textract and various AWS machine learning services to automate the verification of ophthalmic prescriptions submitted in varying formats, including printed and handwritten in English and French.
Results
- Ensured scalability of the extract and endpoint concurrency for a large set of documents.
- Utilized vision and textual feature embedding models to facilitate entity extraction tasks for semi-structured documents.
