Huntington Bank: Redacting sensitive data from 400M+ documents with AWS

This article describes how Huntington Bank redacted sensitive data from 400+ million documents using AWS services, reducing processing time from years to months.

• Used Amazon Textract to detect sensitive information like Social Security numbers and account numbers from documents
• Employed AWS Step Functions with distributed map state to process millions of documents daily at scale
• Leveraged AWS DataSync and AWS Direct Connect to securely transfer encrypted documents to and from on-premises storage
• Implemented redaction workflow using Python libraries with confidence score validation for accuracy verification
• Achieved 10 million documents processed per day with 95%+ redaction accuracy and 5% of original cost estimate

The solution demonstrates how AWS services enable large-scale compliance initiatives and document processing while maintaining security and meeting PCI DSS requirements.