Case Studyamazon bedrockdocument classificationgenai idp
Associa Deploys Generative AI Document Classifier
8.1
Associa partnered with the AWS Generative AI Innovation Center to deploy a GenAI IDP document-classification system using Amazon Bedrock, processing a 465-document evaluation dataset. The team found a first-page-only, multimodal prompting setup raised accuracy from 91% to 95% and halved inference cost (from $0.011 to $0.0055 per document), improving Unknown-class accuracy and reducing processing time.
Key Points
- 1Achieves 95% overall classification accuracy on 465 PDFs using first-page-only prompting.
- 2Reduces inference cost from 1.10 cents to 0.55 cents per document, enabling cheaper scaling.
- 3Uses Amazon Textract OCR plus Bedrock models to improve Unknown-class detection and reduce noise.
Scoring Rationale
Strong applied evaluation and measurable gains across accuracy and cost; limited novelty beyond integration of existing AWS services.
Sources
Public references used for this report.
Practice with real Retail & eCommerce data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
250 free problems · No credit card
See all Retail & eCommerce problems
