Back office ops · Production

Associa transforms document classification with GenAI IDP Accelerator and Amazon Bedrock achieving 95% accuracy

The problem

Associa manages approximately 48 million documents across 26 TB of data but lacked automated classification, forcing employees to spend countless hours manually categorizing incoming documents—a time-consuming, error-prone process creating operational bottlenecks and reduced productivity.

Workflow diagram · grounded in source
1
Incoming document received
trigger
“Every day, employees spend countless hours manually categorizing and organizing incoming documents”
2
First-page OCR extraction
ai_action
“Multimodal prompting with OCR data (using the Amazon Textract analyze_document_layout)”
3
Generative AI document classification
ai_action
“Associa built a generative AI-powered document classification system using Amazon Nova Pro on Amazon Bedrock that achieves 95% accuracy at an average cost of 0.55 cents per document”
4
Unknown document routing
routing
“Accurate Unknown document classification is critical for downstream human review and operational efficiency at Associa”
5
Workflow integration
integration
“The document classification system, developed using the Generative AI Intelligent Document Processing (GenAI IDP) Accelerator, is designed to integrate seamlessly into existing workflows”
Reported outcome

Associa's generative AI-powered classification system achieves 95% overall accuracy at 0.55 cents per document using Amazon Nova Pro on Amazon Bedrock, with the CIO foreseeing significant reduction of manual effort and reporting substantial cost savings.

Reported metrics
Overall classification accuracy95%
Classification cost per document0.55 cents
Documents correctly classified443 out of 465
accuracy improvement (full PDF to first-page)91% to 95%
Show all 8 reported metrics
overall classification accuracy95%
classification cost per document0.55 cents
documents correctly classified443 out of 465
accuracy improvement (full PDF to first-page)91% to 95%
cost reduction per document (full PDF to first-page)1.10 cents to 0.55 cents
Unknown document accuracy improvement40% to 85%
manual effort reductionsignificant reduction of manual effort in document processing
cost savingssubstantial cost savings
Reported stack
Amazon BedrockAmazon TextractGenAI IDP AcceleratorAmazon Nova ProAWS CloudFormationAnthropic Claude Sonnet 4
Source
https://aws.amazon.com/blogs/machine-learning/how-associa-transforms-document-classification-with-the-genai-idp-accelerator-and-amazon-bedrock?tag=soumet-20
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Associa's generative AI-powered classification system achieves 95% overall accuracy at 0.55 cents per document using Amazon Nova Pro on Amazon Bedrock, with the CIO foreseeing significant reduction of manual effort an…

What tools did this team use?

Amazon Bedrock, Amazon Textract, GenAI IDP Accelerator, Amazon Nova Pro, AWS CloudFormation, Anthropic Claude Sonnet 4.

What results were reported?

Overall classification accuracy: 95%; Classification cost per document: 0.55 cents; Documents correctly classified: 443 out of 465; accuracy improvement (full PDF to first-page): 91% to 95% (source-reported, not independently verified).

How is this back office ops AI workflow structured?

Incoming document received → First-page OCR extraction → Generative AI document classification → Unknown document routing → Workflow integration.

Related back office ops cases