AWS Finance and Global Business Services builds an automated contract-processing platform using Amazon Textract and Amazon Comprehend

Use case typeIntelligent document processingUpdated Jun 13, 2026

AWS Finance and Global Business Services built an automated end-to-end contract-processing platform for incoming contracts and agreements. The workflow uses Amazon Textract for OCR, table and form extraction, Amazon Comprehend for text analysis and custom classification of non-standard sections, and a custom web UI for validation and reporting.

Organization
AWS Finance and Global Business Services
Industry
Finance
Location
United States
Published
November 2020

Reported outcomes

5x

processing time reductionTime & speed

150 hours/monthemployee hours per month before automation1 minutestime to process a contract30 hours/monthanalyst effort after automation

Strategic outcomes

New product / capabilityBuilt automated contract-processing platformSpeed & agilityAccelerated contract parsing and storageScale & capacityEnabled live processing of hundreds of contractsCost efficiencyReduced manual processing workload

Catalog median for time & speed deployments: −60% across 727 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 3 of 3

  • 1Document Processing
  • 2Intelligent Automation
  • 3Workflow Automation
  • Manual contract review and data entry took more than 150 employee hours per month.
  • Multiple analysts had to populate Excel in batches, creating slow and labor-intensive processing.
  • Contracts are stored in Amazon S3 and trigger asynchronous Amazon Textract DocumentAnalysis jobs through Amazon SQS and AWS Lambda.
  • Textract outputs are processed into extracted terms and metadata, with Amazon Comprehend custom classification identifying non-standard sections for human review.
  • Extracted data is stored in Amazon DynamoDB and surfaced through a custom Angular web UI, with reporting via Tableau in Amazon AppStream 2.0.
Technologies
Amazon TextractAmazon ComprehendAmazon SageMakerAmazon S3Amazon SQSAWS LambdaAmazon DynamoDBAmazon CloudFrontAmazon CognitoAmazon API GatewayAmazon AppStream 2.0AWS CloudFormationTableau
  • The platform parses and stores contractual data in under a minute per contract.
  • The automated process reduced work to about 30 hours per month for a single analyst.
  • The team reported a 5x reduction in processing time and the system is live with hundreds of contracts per month.
Architecture

Contracts land in Amazon S3 and trigger asynchronous Amazon Textract DocumentAnalysis jobs via Amazon SQS and AWS Lambda. Textract output is stored in S3 for downstream processing, where AWS Lambda functions use Amazon Comprehend custom classification to identify non-standard contract sections after the team built training data with an annotation app hosted on Amazon SageMaker. Extracted contract terms and audit metadata are stored in Amazon DynamoDB. Accounting users interact with a custom Angular web UI hosted on Amazon S3 and Amazon CloudFront, authenticated by Amazon Cognito and fronted by Amazon API Gateway. Reporting is delivered through Amazon AppStream 2.0, with the solution codified in AWS CloudFormation and deployed through a CI/CD pipeline.

Implementation partners1
AWS ProServe
Sources & evidence1
Groundedness: 5/5Type: Blog PostPublished: Nov 12, 2020Publisher: AWSEvidence: VendorConfidence: Medium
Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases
Customer evidence
More from AWS Finance and Global Business Services
Provider evidence
AWS provider pageAWS ranked cases
Market context
Finance AI use casesFinance industry insightsUnited States country insights
Browse the catalog
All AI use casesAI use-case typesIntelligent document processing use cases

Was this useful?