AWS launches Amazon Bedrock Data Automation, a unified service that extracts insights and structure from complex documents—moving beyond simple text scanning to understand context, relationships, and visual elements at scale.

What happened: Amazon introduced Bedrock Data Automation (BDA) within its Bedrock platform. BDA processes documents through a single API by automatically splitting them along logical boundaries, classifying sections into document types, extracting text and tables, and analyzing visual elements like charts and graphs. The service supports up to 3,000 pages and 500 MB per API request, and offers both standard output (summaries, extracted text, captions) and custom output using blueprints—artifacts that define extraction logic for specific document types, with projects able to contain up to 40 document blueprints.

Why it matters: Traditional OCR (optical character recognition) tools extract text but cannot understand context or meaning, creating manual bottlenecks in processing insurance claims, invoices, contracts, and medical records. BDA's ability to validate extracted data and provide confidence scores addresses this gap. AWS describes the service as enabling organizations to 'transform their document processing workflows with minimal development effort' by combining extraction with orchestrated AI agents that can handle complex, multi-document tasks.

What to watch: The architecture integrates BDA with AWS Step Functions (workflow orchestration), DynamoDB (metadata tracking), and Amazon Bedrock Knowledge Base (contextual search across multiple documents). Organizations can process thousands of documents concurrently using asynchronous jobs and task tokens, making the approach suitable for high-volume enterprise document processing pipelines.