Simplify data extraction and process automation across multimodal data-centric workflows, including Intelligent Document Processing (IDP)
This Guidance shows how Amazon Bedrock Data Automation streamlines the generation of valuable insights from unstructured multimodal content such as documents, images, audio, and videos through a unified multi-modal inference API. Amazon Bedrock Date Automation helps developers build generative AI applications or automate multi-modal data centric workflows like IDP, media analysis, or retrieval augmented generation (RAG) quickly and cost-effectively. By following this Guidance, you can simplify complex tasks such as document splitting, classification, data extraction, output format normalization, and data validation, significantly enhancing your processing scalability.
Please note: [Disclaimer]
Architecture Diagram
-
Intelligent document processing
-
Medical claims processing
-
Intelligent document processing
-
This architecture diagram shows how to perform document classification and extraction using a loan origination processing example for a financial services company.
Click to enlarge
Step 1
The data science team uploads sample documents to an Amazon Simple Storage Service (Amazon S3) bucket.Step 2
The data science team uses provided blueprints and creates new custom blueprints for each document class: W2, Pay Slip, Drivers License, 1099, and Bank Statement. Each sample is processed, and generative AI prompts extract fields (such as first and last name, gross pay, capital gains, and closing balance).
Step 3
The blueprints are tested and refined. Key normalizations, transformations, and validations are added.
Step 4
The blueprints are managed and stored in the Amazon Bedrock Data Automation feature.
Step 5
Using an “Object Created” event, Amazon EventBridge triggers an AWS Lambda function when documents are uploaded to Amazon S3. This Lambda function then uses the Amazon Bedrock Data Automation feature to process the uploaded documents.Step 6
The processing workflow in the Amazon Bedrock Data Automation feature includes document splitting based on logical boundaries, with each split containing up to 20 pages. Each page is classified into a specific document type and matched to appropriate blueprints.Step 6 (continued)
The corresponding blueprint is then invoked for each page, executing key normalizations, transformations, and validations. This entire process operates asynchronously, allowing for efficient handling of multiple documents and large data volumes.
Step 7
Amazon Bedrock Data Automation stores the results in a Amazon S3 bucket for later processing and triggers EventBridge.
Step 8
EventBridge triggers the Lambda function to process the JSON results of Amazon Bedrock Data Automation. The processing results are sent to downstream processing systems.
Step 1
The data science team uploads sample documents to an Amazon Simple Storage Service (Amazon S3) bucket.Step 2
The data science team uses provided blueprints and creates new custom blueprints for each document class: W2, Pay Slip, Drivers License, 1099, and Bank Statement. Each sample is processed, and generative AI prompts extract fields (such as first and last name, gross pay, capital gains, and closing balance).
Step 3
The blueprints are tested and refined. Key normalizations, transformations, and validations are added.
Step 4
The blueprints are managed and stored in the Amazon Bedrock Data Automation feature.
Step 5
Using an “Object Created” event, Amazon EventBridge triggers an AWS Lambda function when documents are uploaded to Amazon S3. This Lambda function then uses the Amazon Bedrock Data Automation feature to process the uploaded documents.Step 6
The processing workflow in the Amazon Bedrock Data Automation feature includes document splitting based on logical boundaries, with each split containing up to 20 pages. Each page is classified into a specific document type and matched to appropriate blueprints.
Step 6 (continued)
The corresponding blueprint is then invoked for each page, executing key normalizations, transformations, and validations. This entire process operates asynchronously, allowing for efficient handling of multiple documents and large data volumes.
Step 7
Amazon Bedrock Data Automation stores the results in a Amazon S3 bucket for later processing and triggers EventBridge.
Step 8
EventBridge triggers the Lambda function to process the JSON results of Amazon Bedrock Data Automation. The processing results are sent to downstream processing systems.
-
Medical claims processing
-
This architecture diagram shows how to automate medical claims processing with multimodal input data and processing to improve efficiency and accuracy.
Click to enlarge
Step 1
Providers submit claims documents, images, and videos to Amazon S3.
Step 2
A workflow is triggered in Amazon Bedrock Data Automation.
Step 3
Developers create blueprints in Amazon Bedrock Data Automation to extract relevant data.
Step 4
Amazon Bedrock Data Automation processes documents, images, and videos by extracting text, tables, objects, transcripts; normalizing structuring the data; flagging low-confidence items for review. Amazon Bedrock Data Automation stores the data in Amazon S3 and triggers EventBridge.
Step 5
EventBridge triggers Lambda, which retrieves the Amazon Bedrock Data Automation output from the S3 bucket.
Step 6
Amazon Bedrock Agents uses the Lambda function to fetch the patient's insurance plan details from Amazon Aurora.Step 7
Amazon Bedrock Agents then updates the claims database in Aurora.
Step 8
Adjudicators verify important fields and focus on low-confidence items.
Step 9
Explanation of Coverage (EoC) documents, images, and videos are stored in Amazon S3. Amazon Bedrock Data Automation processes multimodal data with a single API and stores it in Amazon S3. It is then processed, embedded, and stored in a vector collection for Amazon Bedrock Knowledge Bases.Step 10
Amazon Bedrock Agents calculates eligibility using extracted data and indexed information.
Step 11
Amazon Bedrock Agents updates claims database and notifies the adjudicator. The adjudicator reviews and approves or adjusts claim efficiently.
Step 1
Providers submit claims documents, images, and videos to Amazon S3.
Step 2
A workflow is triggered in Amazon Bedrock Data Automation.
Step 3
Developers create blueprints in Amazon Bedrock Data Automation to extract relevant data.
Step 4
Amazon Bedrock Data Automation processes documents, images, and videos by extracting text, tables, objects, transcripts; normalizing structuring the data; flagging low-confidence items for review. Amazon Bedrock Data Automation stores the data in Amazon S3 and triggers EventBridge.
Step 5
EventBridge triggers Lambda, which retrieves the Amazon Bedrock Data Automation output from the S3 bucket.
Step 6
Amazon Bedrock Agents uses the Lambda function to fetch the patient's insurance plan details from Amazon Aurora.Step 7
Amazon Bedrock Agents then updates the claims database in Aurora.
Step 8
Adjudicators verify important fields and focus on low-confidence items.
Step 9
Explanation of Coverage (EoC) documents, images, and videos are stored in Amazon S3. Amazon Bedrock Data Automation processes multimodal data with a single API and stores it in Amazon S3. It is then processed, embedded, and stored in a vector collection for Amazon Bedrock Knowledge Bases.Step 10
Amazon Bedrock Agents calculates eligibility using extracted data and indexed information.
Step 11
Amazon Bedrock Agents updates claims database and notifies the adjudicator. The adjudicator reviews and approves or adjusts claim efficiently.
Get Started
Deploy this Guidance
Well-Architected Pillars
The AWS Well-Architected Framework helps you understand the pros and cons of the decisions you make when building systems in the cloud. The six pillars of the Framework allow you to learn architectural best practices for designing and operating reliable, secure, efficient, cost-effective, and sustainable systems. Using the AWS Well-Architected Tool, available at no charge in the AWS Management Console, you can review your workloads against these best practices by answering a set of questions for each pillar.
The architecture diagram above is an example of a Solution created with Well-Architected best practices in mind. To be fully Well-Architected, you should follow as many Well-Architected best practices as possible.
-
Operational ExcellenceRead the Operational Excellence whitepaper
Amazon S3, EventBridge, and Lambda create a seamless, automated workflow for document processing and data extraction through secure storage for various document types. Amazon Bedrock Data Automation streamlines the extraction and normalization of data, reducing manual effort and increasing accuracy. Amazon Bedrock Knowledge Bases index the processed information, making it easily searchable and accessible, while Amazon Bedrock Agents leverages this structured data to make intelligent decisions and route claims efficiently. Aurora serves as a robust database for storing and retrieving critical information. Together, these services enable a highly efficient, scalable, and reliable system that minimizes human intervention and maximizes productivity.
-
SecurityRead the Security whitepaper
Amazon S3 offers encrypted storage, Lambda executes code in isolated environments, and Amazon Bedrock leverages secure AWS infrastructure with built-in encryption and access controls. Aurora provides advanced database security features. These services create a comprehensive security approach that protects data throughout its lifecycle while maintaining strict access controls and audit trails. The ability to centrally manage security policies and leverage continuous AWS security updates and improvements allows you to maintain a strong security posture while focusing on your core business operations.
-
ReliabilityRead the Reliability whitepaper
Amazon S3 provides durable and highly available storage for documents. EventBridge helps ensures consistent event-driven processing by reliably triggering Lambda functions, which scale seamlessly to handle varying workloads without downtime. Aurora, a highly available database, offers automated backups and failover capabilities. These services offer a robust, fault-tolerant system that can withstand component failures, scale automatically, and maintain consistent performance under high loads, minimizing downtime and data loss risks.
-
Performance EfficiencyRead the Performance Efficiency whitepaper
AWS services enhance performance efficiency through scalable, high-performance solutions for document processing. Amazon S3 provides low-latency access to stored documents, while EventBridge enables real-time event processing. Lambda offers rapid, on-demand compute power. The serverless nature of Lambda and EventBridge eliminates bottlenecks associated with server provisioning. Additionally, Amazon Bedrock leverages AI models for efficient processing of complex data analysis tasks.
-
Cost OptimizationRead the Cost Optimization whitepaper
AWS services contribute to cost optimization through pay-as you-go models (meaning you only pay for resources consumed) and elimination of upfront infrastructure investments. Amazon S3 offers tiered storage options balancing performance and cost. The serverless nature of EventBridge and Lambda means paying only for actual compute time used. Amazon Bedrock provides AI capabilities without expensive in-house infrastructure or expertise, and Aurora offers performance comparable to commercial databases at a fraction of the cost.
-
SustainabilityRead the Sustainability whitepaper
AWS services contribute to sustainability by optimizing resource utilization and energy efficiency. Amazon S3 uses efficient storage technologies, while EventBridge and Lambda provide serverless architectures that minimize idle capacity. These cloud-based services significantly reduce on-premises infrastructure, lowering energy consumption and carbon emissions. Their scalability ensures optimal resource use, avoiding over-provisioning and waste.
Related Content
[Title]
Disclaimer
The sample code; software libraries; command line tools; proofs of concept; templates; or other related technology (including any of the foregoing that are provided by our personnel) is provided to you as AWS Content under the AWS Customer Agreement, or the relevant written agreement between you and AWS (whichever applies). You should not use this AWS Content in your production accounts, or on production or other critical data. You are responsible for testing, securing, and optimizing the AWS Content, such as sample code, as appropriate for production grade use based on your specific quality control practices and standards. Deploying AWS Content may incur AWS charges for creating or using AWS chargeable resources, such as running Amazon EC2 instances or using Amazon S3 storage.
References to third-party services or organizations in this Guidance do not imply an endorsement, sponsorship, or affiliation between Amazon or AWS and the third party. Guidance from AWS is a technical starting point, and you can customize your integration with third-party services when you deploy the architecture.