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

2026-06-24 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Cybersecurity & Data Privacy · Depth: Intermediate, medium

Summary

Huntington National Bank, a top 10 US bank, successfully addressed the challenge of redacting sensitive customer data from over 400 million documents accumulated since 2015. For a proactive compliance initiative in 2025, the bank designed a scalable redaction workflow using AWS services including Amazon Textract, Amazon SageMaker, AWS Step Functions, and AWS Lambda. This solution reduced the estimated processing time from years to months, achieving a processing rate of approximately 10 million documents per day. The project also met critical requirements for data encryption, strict access controls, PCI DSS compliance, and on-premises replication, while exceeding 95% redaction accuracy. The overall cost was approximately 5% of the original estimate, and Huntington plans to reuse this framework for future high-volume redaction needs like mergers and acquisitions.

Key takeaway

For MLOps Engineers or AI Architects tasked with large-scale sensitive data redaction for compliance, you should consider a serverless, ML-driven architecture on AWS. Huntington Bank's success demonstrates that combining services like Amazon Textract, AWS Step Functions, and AWS DataSync can reduce processing timelines from years to months and costs by 95%, while achieving over 95% accuracy. This approach provides a robust framework for handling hundreds of millions of documents securely and efficiently.

Key insights

Huntington Bank scaled sensitive data redaction for 400M+ documents using AWS services, reducing time and cost significantly.

Principles

• Data must be encrypted at rest and in transit.
• Services used must be in-scope for PCI DSS compliance.
• Redaction accuracy must meet or exceed 95% for compliance.

Method

Transfer documents to S3 via AWS DataSync, detect sensitive data with Amazon Textract orchestrated by AWS Step Functions' distributed map state, then redact and sync back on-premises.

In practice

• Use AWS DataSync for secure on-premises to S3 data migration.
• Orchestrate document processing with AWS Step Functions for scalability.
• Leverage Amazon Textract for automated sensitive data detection.

Topics

• Data Redaction
• Amazon Textract
• AWS Step Functions
• AWS DataSync
• PCI DSS Compliance
• Financial Services

Best for: AI Engineer, MLOps Engineer, AI Architect

Related on AIssential

• Building an Event-Driven Data Validation Pipeline on AWS Using S3, Lambda, and SNS — Event-driven serverless architectures on AWS enable automated, real-time data validation workflows, eliminating manual intervention and polling.
• How Meta Rebuilt Data Ingestion for Petabyte-Scale Reliability — Migrating petabyte-scale data ingestion requires staged transitions, continuous validation, and robust rollback mechanisms to ensure zero downtime and consistency.
• Unlock seamless and cost-effective marketing campaigns with Lakebase — Lakebase merges transactional databases with data lake economics for scalable, cost-effective marketing data serving.
• Validio closes $30M Series A to address enterprise data quality challenges — Automated data quality and reliability are crucial for successful AI adoption and regulatory compliance in enterprises.
• Netflix Automates RDS PostgreSQL to Aurora PostgreSQL Migration Across 400 Production Clusters — Automating database migrations via an infrastructure-level platform reduces risk and downtime for large-scale operations.
• Stop Overpaying for S3 Sync Pipelines! Welcome to Amazon S3 Files — Amazon S3 Files allows direct S3 bucket mounting, simplifying data access and reducing storage costs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.