Microsoft has released Azure Content Understanding for general availability, bringing production-grade reliability, expanded global coverage, and deep integration across the Foundry ecosystem. The update delivers major improvements for intelligent document processing (IDP), Retrieval-Augmented Generation (RAG), and agentic applications, giving organizations a unified way to transform unstructured content into structured, AI-ready data.

Direct integration with Microsoft Foundry Models

One of the most requested preview features is now live. Customers can connect Content Understanding directly to any supported Microsoft Foundry model deployment. This allows teams to choose the right model for their latency, quality, and cost needs while keeping generative and specialized models in a single pipeline. The service is also expanding to 12 global regions, improving performance and meeting data residency requirements.

Faster time-to-value with prebuilt analyzers

Alongside custom analyzers, Microsoft now offers a full catalog of prebuilt analyzers for finance, procurement, contracts, identity, RAG ingestion, and more. These bring Azure Document Intelligence capabilities together with foundation models for richer, more accurate extraction.

Purpose-built RAG analyzers

RAG solutions require high-quality, structured content. Content Understanding now offers:

• Multi-page table extraction

• Figure understanding with descriptions and Chart.js or Mermaid.js outputs

• Layout-aware markdown with headers, sections, and annotations

These enhancements produce dense, structured chunks that dramatically improve retrieval quality and grounding.

Prebuilt analyzers for images, audio, video, and documents

Beyond PDFs and forms, new analyzers support multimodal content and produce RAG-ready outputs—no custom training required.

Domain-specific analyzers for high-value workloads

The GA release includes specialized extraction pipelines for tax forms, financial statements, invoices, procurement documents, mortgage files, identity verification, and more.

Now integrated into Foundry IQ and Azure AI Search

Content Understanding now powers ingestion for both Azure AI Search and Foundry IQ, giving search and agentic applications richer metadata, better embeddings, and a unified ingestion path without building separate pipelines.

Agents can now use Content Understanding as a tool

Agents can read and understand files using structured outputs backed by standard schemas and prebuilt analyzers. A Content Understanding MCP server is coming soon for multi-agent ecosystems.

Improved document processing with segmentation and routing

Complex documents can now be split into subsections—such as separating application forms, certificates, and supporting files—and routed to specialized analyzers. Confidence scores, grounding options, granular extraction control, and support for lightweight “mini” and “nano” models help reduce cost and latency.

Customer momentum

We’re unleashing the power of AI to transform audit for 95,000 professionals worldwide. By integrating Azure Content Understanding into KPMG Clara, we’re converting complexity into clarity—turning millions of documents into insights that drive productivity and redefine the future of audit.”

– Thomas Mackenzie, Global Audit Chief Digital Officer, KPMG

Leading organizations are already using Content Understanding at scale:

• KPMG uses it in its Clara audit platform to turn invoices, statements, and leases into structured insights with confidence scores.

• DataSnipper uses it to power Excel-native AI Extractions for traceable, audit-ready workflows.

• SJR integrates it into its GX Manager to deliver personalized digital experiences at enterprise scale.

By embedding Microsoft Azure Content Understanding in DataSnipper, we are turning unstructured documents into structured, actionable data – directly within Excel. Together, we are enabling faster reviews, reliable evidence, and AI you can trust.”

– Vidya Peters, CEO, DataSnipper

Get started

Content Understanding is now the recommended starting point for new document and file-processing workloads. Developers can explore the new experience in Microsoft Foundry, review documentation, and try GitHub samples at aka.ms/cu-samples.