What changed
AWS announced the general availability of Amazon Textract Insights, a new generative AI feature of Amazon Textract. This feature allows organizations to describe the fields they want in natural language and receive structured results with confidence scores and page-level grounding, without requiring model training.
What this means in practice
Organizations in financial services, insurance, healthcare, and the public sector can now process mixed, multi-document packets by defining what to extract in a blueprint. A blueprint is a JSON Schema where each field has a natural language description and an inference type. Users can start with AWS public blueprints for common document types or create their own. For mixed packets, users add blueprints for each expected document type and submit a single file from Amazon S3. Amazon Textract Insights splits the packet into logical sub-documents, matches each to its blueprint, and returns results with confidence scores and grounding.
Limitations and availability
Amazon Textract Insights is not yet available in the Asia Pacific (Tokyo) and Europe (Zurich) AWS Regions, where Amazon Textract Core is now generally available. Projects and blueprints are encrypted at rest with AWS owned keys, and users can specify their own AWS KMS customer managed key. All data is transmitted encrypted across Amazon’s secure network, with inference requests routed to optimal AWS Regions within the request’s geography. The feature is available through a new API within Amazon Textract.
