What challenges did Ernst & Young face before AI implementation?

Tax forms are complex, varied, and often span hundreds of pages, making manual ingest and transcription slow and error-prone during tax season.

What was the impact of the AI solution?

Processed about 16,000 federal and state tax forms spanning 200,000 pages. Reduced client onboarding from weeks or months to days. Improved extraction speed, accuracy, and operational efficiency for high-volume tax workflows.

Source

MicrosoftScaled productionEvidence: Medium65/100

EY automates K-1 tax data extraction with Azure AI Document Intelligence

Use case typeDocument processing automationUpdated Jul 11, 2026

Global financial service provider EY accelerated work for clients with Azure AI Document Intelligence. Tax work is timely and requires absolute accuracy, and the forms are complicated, are in various formats, and often span hundreds of pages. EY was an early adopter of Microsoft Azure AI Document Intelligence and integrated Microsoft OpenAI within a secure, cloud-native elastic Azure architecture to automate structured and unstructured K-1 tax data processing at scale.

Organization: Ernst & Young
Industry: Professional Services
Location: United Kingdom
Published: July 2026

Reported outcomes

Tax forms processed: Approximately 16,000 forms

Other quantified impact

Pages processed: 200,000 pages

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Manual interventions reduced: 90% decrease

Microsoft Customer StoriesJul 11, 2026Customer storyExplicit claimMedium evidence strength

reduced manual interventions by up to 90 percent

Normalized claim

Tax forms processed: 16,000 forms increase

Microsoft Customer StoriesJul 11, 2026Customer storyExplicit claimMedium evidence strength

processed about 16,000 federal and state tax forms

Normalized claim

Pages processed: 200,000 pages increase

Microsoft Customer StoriesJul 11, 2026Customer storyExplicit claimMedium evidence strength

spanning 200,000 pages of both structured and unstructured data

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Ernst & Young
Provider: Microsoft
Maturity: Scaled Production
Linked source: Microsoft Customer Stories

EY was an early adopter of Microsoft Azure AI Document Intelligence and integrated Microsoft OpenAI within a secure, cloud-native elastic Azure architecture to automate structured and unstructured K-1 tax data processing at scale

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 2 of 2

1Document processing automation
2Intelligent document processing

Built and trained document intelligence models for specialized tax form types.
Leveraged generative AI to create synthetic data for model training and avoid storing client data.
Scaled the solution across EY's Global Tax Platform with cloud-native Azure architecture and continuous feedback for model improvement.

Technologies

Azure AI Document Intelligence Microsoft OpenAI Microsoft Azure

Processed about 16,000 federal and state tax forms spanning 200,000 pages.
Reduced client onboarding from weeks or months to days.
Improved extraction speed, accuracy, and operational efficiency for high-volume tax workflows.

Architecture

EY built specialized document intelligence models on Azure AI Document Intelligence and integrated Microsoft OpenAI into a secure, cloud-native elastic Azure architecture. The solution used synthetic data generation for training, continuous feedback loops for model improvement, and Microsoft Unified support to scale compute resources and manage capacity for high-volume tax processing.

Sources & evidence1

Evidence: Medium65/100Evidence strength