At Microsoft Ignite this year, we’re excited to announce that Azure Content Understanding in Foundry Tools is now generally available (GA). Over the past months, we’ve seen preview usage across industries, from large consultancies to healthcare leaders, with invaluable customer feedback shaping this release.

With this GA release, we’re enabling flexibility and control with model choice, production-grade reliability, expanded region availability, and broader scenario coverage. In addition, this update brings tight integration with Microsoft Foundry Models, Foundry IQ powered by Azure AI Search, and agent ecosystems—making it easier than ever to build intelligent solutions.

Let’s walk through what’s new.

Content Understanding now connects directly to Foundry Models

One of the biggest requests we heard during preview was: “Can I bring my own Gen AI model deployment?” 

Now you can.

You can now connect Content Understanding to any supported Microsoft Foundry model deployment:

• Make the quality and latency trade-offs based on the demands of your use case.
• Save cost and consolidate usage by using pay-as-you-go or Provisioned Throughput Units (PTUs) and leveraging global or data zone deployments.

Figure 1: Content Understanding now extends to support Foundry model deployments, for enabling generative AI capabilities.

Content Understanding pairs specialized models and generative AI where each excels: specialized models handle OCR, Layout, Transcription and produce confidence scores, while GenAI powers field extraction, segmentation, and figure analysis. The result is high quality output with predictable cost control and clear architecture.

We’re also expanding availability to 12 regions worldwide, improving data residency options and lowering latency for global workloads. To learn more about supported regions, click here.

Faster time-to-value with prebuilt analyzers

Alongside the existing ability to create custom analyzers, we now provide a library of prebuilt analyzers optimized for top use cases like RAG ingestion, finance, procurement, and more. These capabilities have been some of the most popular in Azure Document Intelligence and we’re now extending them with foundation models in Content Understanding.

Purpose-built analyzers for RAG

Retrieval-Augmented Generation (RAG) solutions succeed when responses are contextual, accurate, and relevant. Achieving that requires transforming unstructured content into structured knowledge. With GA, Content Understanding introduces RAG-specific analyzers designed to extract tables, paragraphs, and sections—and now, to analyze figures as well. This richer structure improves grounding, so your RAG solutions deliver answers that truly reflect the source content.

Layout-aware enrichment across documents

Content Understanding now adds the ability to:

• Extract tables across pages into a single markdown or JSON structure
• Parse figures and converts charts/diagrams into structured representations (Chart.js, Mermaid.js)
• Generate layout-aware markdown with headers, sections, and annotations represented to dramatically improve chunking and retrieval quality

This level of structural fidelity translates to context-dense chunks which results in more accurate responses.

Figure 2: Figure analysis provides both description and structured mermaid code for diagrams providing much better grounding for downstream generations.

Figure 3: Figure analysis provides both description and structured ChartJS or Mermaid code for diagrams providing much better grounding for downstream generations.

A full catalog of prebuilt RAG analyzers

In addition to documents, we are also making prebuilt analyzers for video, audio and images giving you RAG-ready outputs with no custom training.

For details on prebuilt analyzers, check out Azure Content Understanding in Foundry Tools prebuilt analyzers – Foundry Tools | Microsoft Learn.

High-quality extraction paired with built-in enrichment creates richer, structured representations of your content. This means downstream models don’t just see raw text. They receive contextually organized data that includes summaries, segmentation, and custom schemas.

With this foundation, models can reason more effectively, improving accuracy and relevance in tasks like question answering, summarization, and retrieval. In short, better data representations lead to better intelligence.

Domain-specific analyzers

Our new domain-specific prebuilt analyzers accelerate intelligent document processing (IDP) pipelines in Finance (statements, tax forms, W-2s, 1099s), Contracts & procurement (Contracts, Invoice, PO, etc.), Mortgage & lending, Identity verification, and more.

These analyzers provide specialized field extraction for specific document types and formats, powered by rich knowledge bases of real-world examples.

For the full list of supported analyzers and details on how to try them out see Azure Content Understanding in Foundry Tools prebuilt analyzers – Foundry Tools | Microsoft Learn.

Now integrated into Foundry IQ and Azure AI Search

These latest capabilities like layout-aware enrichment with diagram and chart descriptions are now available across both Azure AI Search and Foundry IQ. This means that whether you’re building agentic workflows or enterprise search experiences, you can leverage Content Understanding for ingestion to deliver richer, more relevant context for every query. With seamless integration, developers no longer need to build custom pipelines or manage complex ingestion steps.

This gives search applications richer, more relevant embeddings and metadata for every file—no extra pipelines required.

See Azure Content Understanding skill – Azure AI Search | Microsoft Learn for more details.

Agents can now use Content Understanding as a tool

Agents need rich context to act intelligently. Whether it’s a customer application, a receipt, or a contract, every detail matters. Content Understanding now ensures that context is complete and consistent by delivering:

• Standardized representation for any supported file, so agents can process diverse content seamlessly.
• Consistent schema across modalities, enabling reliable interpretation regardless of format.
• Prebuilt domain-specific analyzers, accelerating extraction of critical details without custom development.

Coming soon: a Content Understanding MCP server, enabling integration into multi-agent systems and tool ecosystems.

Upgrades for document processing

If you’ve been using Content Understanding during preview or migrating from Document Intelligence, this GA release delivers a host of updates designed to enhance usability and streamline your workflows.

Splitting and classification for complex documents

Now, complex documents can be automatically segmented into their constituent pieces. For example, separating a visa application into the application form, marriage license, and proof of residency.

Each segment is then routed to a purpose-built analyzer, ensuring that every piece extracts the right structured representation for downstream systems to process. In the visa example, the marriage certificate can leverage a pre-built analyzer, while the application can route to an analyzer custom built for this application.

Figure 4: New splitting and classification capabilities enable the creation of complex workflows by breaking documents into smaller sections and analyzing each with purpose-built analyzers.

This new capability unlocks powerful scenarios in finance, government, healthcare, and beyond—where documents are rarely one-size-fits-all, and precision matters at every step.

Granular control to reduce cost and complexity

We’ve added support for grounding and confidence scores to more scenarios, added more granular control and more complex schemas to improve flexibility and reduce cost. These updates also help lower operational costs and latency by optimizing extraction workflows and enabling faster, more efficient processing. Key enhancements include:

• Confidence scores and grounding across all document extraction methods
• Per-field controls for when confidence and grounding are applied
• Automatic selection of extraction method, so customers don’t need to explicitly select between extract or generate for each field.
• Support for complex object types within schemas
• Additional control of content extraction method used (read or layout) and the generative AI model types, now including mini and nano models.

As a result: lower cost, lower latency, and the right structured outputs to enable straight-through-processing.

A clear path forward

With GA, we now recommend Content Understanding as the starting point for new file-processing workloads and related agentic applications. It carries forward the power of Document Intelligence while adding field extraction powered by foundation models, multimodal capabilities, RAG optimization, and integration of analyzers within Microsoft Foundry.

Customer momentum: Leading brands deriving value with Content Understanding

Customers across industries are already building high-scale production workloads with Content Understanding.

KPMG: Bringing clarity to complex audits with Azure Content Understanding

KPMG Clara AI is redefining audit intelligence by tackling the surge of unstructured data in modern audits. By integrating Azure Content Understanding, Clara transforms documents like invoices, bank statements, and lease agreements into structured, actionable insights. Using advanced analyzers for field and table extraction, Clara delivers standardized outputs (HTML/Markdown) and confidence scores—empowering auditors to validate results quickly, reduce manual effort, and scale precision across thousands of engagements worldwide.

“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

DataSnipper: Transform unstructured documents into actionable data

DataSnipper is an intelligent automation platform for faster audit and finance workflows. Its latest capability, AI Extractions, uses Azure Content Understanding to interpret and transform unstructured documents, such as payroll reports and bank statements, into data that teams can directly work with in Excel. Every extracted value links back to its source for fully traceable reviews.

By combining DataSnipper’s Excel-native experience with Azure Content Understanding, AI Extractions helps teams structure and verify data at scale while improving accuracy and auditability. This translates into teams spending less time doing manual work and more time exercising judgment.

“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

SJR: The high stakes of scaling personalization

SJR, a WPP company, is setting a new industry benchmark for digital personalization, driving transformation across web experiences with its first-to-market AI solution, Generative Experience (GX) Manager. Read more about SJR’s integration of Content Understanding in GX Manager here.

“Our collaboration with Microsoft made GX Manager more than a website upgrade. It helped us deliver a paradigm shift in digital marketing that has opened new possibilities for intelligent content that drives conversion.”

– Selena Cameron, Global CEO, SJR

Get started today

Content extraction for contracts and forms remains one of the most impactful use cases, consistently delivering measurable ROI. Begin by targeting a single high-friction process—such as an approval workflow, a media archive, or an application-document pipeline—that can benefit from greater accuracy, cost reduction, and automation. Use Content Understanding to analyze these files and unlock structured insights. Get started in the new Content Understanding user experience in Microsoft Foundry.

And be sure to check out these on-demand breakout sessions from Microsoft Ignite 2025:

• Innovation Session: Build & Manage AI Apps with Your Agent Factory
• Introducing Microsoft Foundry Tools

To learn more:

• Visit What is Azure Content Understanding in Foundry Tools? – Foundry Tools | Microsoft Learn
• Explore GitHub samples here: aka.ms/cu-samples