AWS updates Amazon Bedrock’s Data Automation capability

AWS has updated the Data Automation capability inside its generative AI service Amazon Bedrock to further support the automation of generating insights from unstructured data and bring down the development time required for building applications underpinned by large language models (LLMs).

Bedrock’s Data Automation, according to AWS, is targeted at developers who can use the capability to accelerate the development of generative AI-driven applications by helping build components or workflows, such as automated data analysis for insights, in a simplified manner.

AWS has integrated Data Automation with Bedrock’s Knowledge Bases capability to help developers extract information from unstructured multimodal data and use it as context for retrieval augmented generation (RAG) use cases.

The latest update to Data Automation includes support for modality enablement, modality routing by file type, extraction of embedded hyperlinks when processing documents, and an increased overall document page limit of 3,000 pages.

“These new features give you more control over how your multimodal content is processed and improve Bedrock Data Automation’s overall document extraction capabilities,” AWS wrote in a blog post.

AWS said that enterprise developers can use the modality enablement feature to configure which modalities — image, document, audio, and video — will be processed amongst all data for a particular project or application.

Developers also have the choice to route specific file types as modalities, and what that means is developers will be able to process JPEG or JPG files as documents, and MP4 or M4V files as video files instead of their original image or audio type via Data Automation.

Another feature that has been added to Data Automation is the embedding of hyperlinks found in PDFs as part of the output or insights generated.

“This feature enhances the information extraction capabilities from documents, preserving valuable link references for applications such as knowledge bases, research tools, and content indexing systems,” the cloud services provider wrote.

Additionally, AWS has also increased support for processing documents, up to 3,000 pages per document from 1,500 pages per document, in Bedrock Data Automation.

The increased limit will enable developers to process larger documents without splitting them, the cloud services provider said, adding that this also simplifies workflows for enterprises dealing with long documents or document packets.Currently, Amazon Bedrock Data Automation is generally available in the US West (Oregon) and US East (Northern Virginia) regions.