{"id":1479,"date":"2025-11-20T08:00:12","date_gmt":"2025-11-20T16:00:12","guid":{"rendered":"https:\/\/devblogs.microsoft.com\/foundry\/?p=1479"},"modified":"2026-01-21T10:54:03","modified_gmt":"2026-01-21T18:54:03","slug":"announcing-azure-language-in-foundry-tools-for-deterministic-privacy-first-agents","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/foundry\/announcing-azure-language-in-foundry-tools-for-deterministic-privacy-first-agents\/","title":{"rendered":"Announcing Azure Language in Foundry Tools for deterministic, privacy-first agents"},"content":{"rendered":"<p>In today\u2019s rapidly evolving AI landscape, developers are seeking reliable, secure, and predictable language capabilities to power the next generation of enterprise-grade agents. As agentic architecture becomes central to modern applications, teams need tools that deliver stronger privacy guarantees, deterministic behavior, and seamless integration across their AI stack.<\/p>\n<p>As part of the broader transition from Azure AI Services into unified Microsoft Foundry Tools, we\u2019re excited to introduce the rebranded <a href=\"https:\/\/aka.ms\/azure-language\"><strong>Azure Language in Foundry Tools<\/strong><\/a> with new enhancements purpose-built for agent development. Today\u2019s release brings <strong>a powerful new remote MCP server<\/strong> with a comprehensive suite of Language AI tools, <strong>significant advancements in PII Redaction<\/strong>, and <strong>new deterministic intent routing features in Conversational Language Understanding (CLU)<\/strong>, giving developers the control, safety, and consistency they need to confidently build and scale intelligent agents.<\/p>\n<h2>Azure Language Remote MCP Server<\/h2>\n<p>As developers build more capable agents, a familiar challenge keeps popping up: how do you give an agent the language skills it needs, like redaction, summarization, entity extraction, or intent detection, without stitching together a bunch of separate APIs and custom routing logic?<\/p>\n<p>Earlier this year, we\u2019ve already taken steps to simplify this through <a href=\"https:\/\/github.com\/Azure-Samples\/ai-language-samples\/tree\/main\/src\/dotnet\/mcp\">local MCP tools for Azure Language<\/a>, giving developers a way to run Language capabilities directly on their machines for development. The new <a href=\"https:\/\/learn.microsoft.com\/azure\/ai-services\/language-service\/concepts\/foundry-tools-agents#azure-language-mcp-server-\">Azure Language remote MCP server<\/a> takes this even further.<\/p>\n<p>With this release, we\u2019re bringing Azure Language capabilities directly into the Model Context Protocol (MCP) as a fully managed, cloud-hosted service. Your agents can now tap into a rich library of deterministic Language AI tools with minimal integration work \u2014 no infrastructure to run, no services to host. Instead of hand-building complex pipelines, your agent can call the Azure Language MCP server and instantly gain capabilities for PII redaction, sentiment analysis, summarization, question answering, and more.<\/p>\n<h3>A complete Language AI toolbox for agents<\/h3>\n<p>Through the MCP server, agents can natively use:<\/p>\n<ul>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/personally-identifiable-information\/overview?tabs=text-pii\"><strong>PII Redaction<\/strong><\/a> for plain text and native documents (.pdf, .docx, etc. stored in Azure Storage) to detect and redact personally identifiable information (PII) from text or documents, including categories like names, email addresses, social security numbers, locations, and more.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/conversational-language-understanding\/overview\"><strong>Intent Detection<\/strong><\/a> to detect intents and entities in conversational messages as defined in your specific CLU project and deployment.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/question-answering\/overview\"><strong>Exact Question Answering<\/strong><\/a> to answer questions based on provided text documents or knowledge bases configured in Custom Question Answering projects with exact verbatims.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/language-detection\/overview\"><strong>Language Detection<\/strong><\/a> to detect the language of any input.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/text-analytics-for-health\/overview?tabs=ner\"><strong>Healthcare Entity Extraction<\/strong><\/a> to identify healthcare-specific entities in text, including support for FHIR (Fast Healthcare Interoperability Resources) standards.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/named-entity-recognition\/overview\"><strong>Named Entity Recognition (NER)<\/strong><\/a> to extract named entities such as persons, organizations, locations, dates, and others from text.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/sentiment-opinion-mining\/overview\"><strong>Sentiment Analysis with Opinion Mining<\/strong><\/a> to understand users\u2019 sentiment level and the cause of sentiment.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/summarization\/overview?tabs=text-summarization\"><strong>Text Summarization<\/strong><\/a> to generate summaries (both abstractive and extractive) of input text with options for summary length and number of sentences.<\/li>\n<li><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/key-phrase-extraction\/overview\"><strong>Key Phrase Extraction<\/strong><\/a> to surface the most important topics and concepts from text.<\/li>\n<\/ul>\n<p>All these capabilities are exposed in a way that\u2019s predictable, structured, and agent-ready.<\/p>\n<h3>Real-world use cases where Azure Language MCP server shines<\/h3>\n<p>Developers can plug Azure Language remote MCP server into their agents to unlock powerful, end-to-end workflows. Here are some of the most common patterns we\u2019re seeing:<\/p>\n<table style=\"border-collapse: collapse; width: 100%;\">\n<tbody>\n<tr>\n<td style=\"width: 21.9713%; vertical-align: top;\"><strong>Use Cases<\/strong><\/td>\n<td style=\"width: 28.7448%; vertical-align: top;\"><strong>Description<\/strong><\/td>\n<td style=\"width: 49.2838%; vertical-align: top;\"><strong>Examples<\/strong><\/td>\n<\/tr>\n<tr>\n<td style=\"width: 21.9713%; vertical-align: top;\"><strong>Privacy-preserving AI assistants<\/strong><\/td>\n<td style=\"width: 28.7448%; vertical-align: top;\">Agents handling sensitive documents or messages (such as medical notes, claims, or customer chats) can use PII Redaction via MCP before sending any content to downstream systems.<\/td>\n<td style=\"width: 49.2838%; vertical-align: top;\">A healthcare triage agent may:<\/p>\n<ul>\n<li>Receive a patient message,<\/li>\n<li>Call <strong>PII Redaction<\/strong> tool in Azure Language remote MCP to remove identifiers,<\/li>\n<li>Send the redacted text for triage suggestions,<\/li>\n<li>Respond with clinically helpful guidance while maintaining privacy protections.<\/li>\n<\/ul>\n<\/td>\n<\/tr>\n<tr>\n<td style=\"width: 21.9713%; vertical-align: top;\"><strong>Intelligent case routing and workflow automation<\/strong><\/td>\n<td style=\"width: 28.7448%; vertical-align: top;\">Agents supporting customer service or operations teams can use Intent Detection with entity extraction, Sentiment Analysis and Key Phrase Extraction to consistently route requests to the right workflow.<\/td>\n<td style=\"width: 49.2838%; vertical-align: top;\">A customer support agent may:<\/p>\n<ul>\n<li>Receive a customer inquiry,<\/li>\n<li>Call <strong>Intent Detection<\/strong> to classify it as a billing issue or card problem and identify specific entities like &#8220;credit card&#8221; or &#8220;checking account&#8221;,<\/li>\n<li>Call <strong>Key Phrase Extraction<\/strong> to surface additional details.<\/li>\n<li>Use <strong>Sentiment Analysis<\/strong> to prioritize urgent or negative messages,<\/li>\n<li>Route the request into the correct workflow with full determinism.<\/li>\n<\/ul>\n<\/td>\n<\/tr>\n<tr>\n<td style=\"width: 21.9713%; vertical-align: top;\"><strong>Knowledge-grounded enterprise assistants<\/strong><\/td>\n<td style=\"width: 28.7448%; vertical-align: top;\">Agents that answer policy, IT, or HR questions can rely on QA from Knowledge Bases to provide accurate, grounded responses without hallucinating.<\/td>\n<td style=\"width: 49.2838%; vertical-align: top;\">An IT helpdesk agent may:<\/p>\n<ul>\n<li>Receive a question like \u201cHow do I reset my MFA token from home?\u201d,<\/li>\n<li>Call <strong>Question Answering<\/strong> from a configured Knowledge Base for IT support,<\/li>\n<li>Retrieve the exact approved steps,<\/li>\n<li>Provide the employee with a correct, consistent and policy-compliant answer every time.<\/li>\n<\/ul>\n<\/td>\n<\/tr>\n<tr>\n<td style=\"width: 21.9713%; vertical-align: top;\"><strong>Document intelligence for compliance and operations<\/strong><\/td>\n<td style=\"width: 28.7448%; vertical-align: top;\">Legal, audit, and compliance teams can use NER, Healthcare Entity Extraction, Key Phrase Extraction, and Summarization to quickly turn large documents into structured insights.<\/td>\n<td style=\"width: 49.2838%; vertical-align: top;\">A compliance review agent may:<\/p>\n<ul>\n<li>Receive an uploaded document,<\/li>\n<li>Call <strong>Named<\/strong> <strong>Entity Extraction<\/strong> to identify people, organizations, and financial terms,<\/li>\n<li>Call <strong>Summarization<\/strong> to highlight key risks or obligations,<\/li>\n<li>Call <strong>Key Phrase Extraction<\/strong> to surface clauses requiring deeper review,<\/li>\n<li>Present a structured summary that accelerates the review process.<\/li>\n<\/ul>\n<\/td>\n<\/tr>\n<tr>\n<td style=\"width: 21.9713%; vertical-align: top;\"><strong>Multilingual enterprise experiences<\/strong><\/td>\n<td style=\"width: 28.7448%; vertical-align: top;\">Global agents can use Language Detection, NER, and Sentiment Detection to support users across multiple languages while maintaining consistent behavior.<\/td>\n<td style=\"width: 49.2838%; vertical-align: top;\">A multilingual HR assistant may:<\/p>\n<ul>\n<li>Receive an employee question in Spanish,<\/li>\n<li>Call <strong>Language Detection<\/strong> to identify the input language,<\/li>\n<li>Call <strong>Named Entity Extraction<\/strong> to extract entities like leave type or office location,<\/li>\n<li>Route the request into the correct localized workflow,<\/li>\n<li>Respond with guidance tailored to the employee\u2019s language.<\/li>\n<\/ul>\n<\/td>\n<\/tr>\n<tr>\n<td style=\"width: 21.9713%; vertical-align: top;\"><strong>Meeting and productivity assistants<\/strong><\/td>\n<td style=\"width: 28.7448%; vertical-align: top;\">Agents supporting teams can use Summarization, Key Phrase Extraction, and Sentiment Detection to transform long meetings, chats, or documents into actionable outcomes.<\/td>\n<td style=\"width: 49.2838%; vertical-align: top;\">A meeting assistant agent may:<\/p>\n<ul>\n<li>Receive a meeting transcript,<\/li>\n<li>Call <strong>Summarization<\/strong> to extract key decisions,<\/li>\n<li>Call <strong>Key Phrase Extraction<\/strong> to identify owners, tasks, and deliverables,<\/li>\n<li>Call <strong>Sentiment Analysis<\/strong> to understand tone around contentious topics,<\/li>\n<li>Produce an actionable meeting recap with follow-up tasks.<\/li>\n<\/ul>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h3>Getting started with Azure Language remote MCP server<\/h3>\n<p>Whether you&#8217;re integrating the Azure Language MCP server into an existing or brand-new agent, we\u2019ve made the process straightforward and flexible. You can connect using the public MCP endpoint, authenticate with the method that works best for your environment, and optionally add the tool directly through the Foundry portal. Below is a detail guide to help you start using the Azure Language remote MCP server in minutes.<\/p>\n<p><span style=\"font-size: 14pt;\"><strong>Connect to the Azure Language MCP endpoint<\/strong><\/span><\/p>\n<p>The Azure Language MCP server is available through a dedicated MCP endpoint on your Foundry resource:<\/p>\n<pre class=\"prettyprint language-default\" style=\"padding-left: 40px;\"><code class=\"language-default\">https:\/\/{foundry-resource-name}.cognitiveservices.azure.com\/language\/mcp?api-version=2025-11-15-preview<\/code><\/pre>\n<p><em>{foundry-resource-name}<\/em> is the name of your Foundry resource. It can be the same resource you use to build your agent, or a different one.<\/p>\n<p><span style=\"font-size: 14pt;\"><strong>Authenticate using a key or Microsoft Entra ID<\/strong><\/span><\/p>\n<p>The MCP server supports two authentication methods to meet the needs of both rapid prototyping and enterprise-grade deployments:<\/p>\n<ul>\n<li><strong>Key-based authentication<\/strong>\nUse the key from your Foundry resource for a quick, no-setup connection. You can find the key in the overview page of your Foundry project where the resource is used, or in Azure portal.<\/li>\n<li><strong>Microsoft Entra authentication<\/strong>\nIdeal for production agents, controlled RBAC, and cross-resource scenarios. If your agent is running in a different Foundry resource than the one you want to use in the MCP endpoint, ensure the agent\u2019s (or the project\u2019s) identity has the <strong>Cognitive Services User <\/strong>role on the MCP resource.<\/li>\n<\/ul>\n<p><span style=\"font-size: 14pt;\"><strong>Configure the MCP through Foundry Tools in the portal<\/strong><\/span><\/p>\n<p>If you\u2019re working inside the new Foundry portal, you can add the Azure Language MCP server with just a few clicks. There are a couple of options to access the MCP:<\/p>\n<ul>\n<li><strong>Option 1: Add from the Foundry Tools\n<\/strong>This option allows you to configure the MCP first, then decide later which agent to add to:<\/p>\n<ul>\n<li>In <strong>New Foundry<\/strong> portal, go to <strong>Discover<\/strong>, then select <strong>Tools<\/strong> from the left menu.<\/li>\n<li>Search for <strong>Azure Language in Foundry Tools<\/strong>, then select the entry.<\/li>\n<li>Select <strong>Connect<\/strong> button, then give it a name, provide your resource name and select the authentication method.<\/li>\n<li>Once you connect, it will be listed in <strong>Build<\/strong> -&gt; <strong>Tools<\/strong><\/li>\n<li>You can select <strong>Use in an agent<\/strong> button in the added tools page to add it into an agent.<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<p><figure id=\"attachment_1513\" aria-labelledby=\"figcaption_attachment_1513\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-1-mcp-in-tools-catalog.webp\"><img decoding=\"async\" class=\"wp-image-1513\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-1-mcp-in-tools-catalog-1024x590.webp\" alt=\"Discover Azure Language remote MCP in Tool Catalog\" width=\"600\" height=\"346\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-1-mcp-in-tools-catalog-1024x590.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-1-mcp-in-tools-catalog-300x173.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-1-mcp-in-tools-catalog-768x442.webp 768w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-1-mcp-in-tools-catalog.webp 1351w\" sizes=\"(max-width: 600px) 100vw, 600px\" \/><\/a><figcaption id=\"figcaption_attachment_1513\" class=\"wp-caption-text\"><em>Figure 1: Discover Azure Language remote MCP in Foundry Tools<\/em><\/figcaption><\/figure><\/p>\n<ul>\n<li><strong>Option B: Add directly to an agent<\/strong>\nIf you already have an agent where you want to add the MCP tools, following these steps:<\/p>\n<ul>\n<li>Open your agent in <strong>New Foundry<\/strong> portal, which opens the agent\u2019s <strong>Playground<\/strong>.<\/li>\n<li>Select <strong>Add<\/strong> button in <strong>Tools<\/strong><\/li>\n<li>Select <strong>+ Add a new tool<\/strong> at the bottom of the pop-up menu.<\/li>\n<li>Select <strong>Catalog<\/strong> tag in the pop-up <strong>Select a tool<\/strong><\/li>\n<li>Choose <strong>Azure Language in Foundry Tools<\/strong> from the available tools<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<p><figure id=\"attachment_1512\" aria-labelledby=\"figcaption_attachment_1512\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-2-add-mcp-to-agent.webp\"><img decoding=\"async\" class=\"wp-image-1512\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-2-add-mcp-to-agent-1024x559.webp\" alt=\"Add Azure Language MCP to an existing agent\" width=\"600\" height=\"328\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-2-add-mcp-to-agent-1024x559.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-2-add-mcp-to-agent-300x164.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-2-add-mcp-to-agent-768x420.webp 768w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-2-add-mcp-to-agent.webp 1404w\" sizes=\"(max-width: 600px) 100vw, 600px\" \/><\/a><figcaption id=\"figcaption_attachment_1512\" class=\"wp-caption-text\"><em>Figure 2: Add Azure Language MCP to an existing agent<\/em><\/figcaption><\/figure><\/p>\n<p>Once added, you can test requests and inspect output in the agent playground interactively.<\/p>\n<p><span style=\"font-size: 14pt;\"><strong>Test the MCP in Visual Studio Code GitHub Copilot<\/strong><\/span><\/p>\n<p>One of the easiest ways to try the Azure Language MCP server is through GitHub Copilot in Visual Studio Code, which natively supports the Model Context Protocol (MCP). By connecting the Azure Language MCP server as a custom tool, you can experiment with all its capabilities right from your editor, before wiring it into a full agent. Here\u2019s how you can set it up in just a few steps:<\/p>\n<ul>\n<li>Follow the instruction in <a href=\"https:\/\/code.visualstudio.com\/docs\/copilot\/customization\/mcp-servers#_other-options-to-add-an-mcp-server\">Use MCP servers in VS Code<\/a> to either add the MCP server to a workspace \u201cmcp.json\u201d file or your user configuration.<\/li>\n<li>Select HTTP as the server type, then provide the Azure Language MCP server endpoint with your Foundry resource as the MCP server URL.<\/li>\n<li>In the <em>json<\/em> file where the MCP server is added, add the following at the end of your added MCP object for key-based authentication:<\/li>\n<\/ul>\n<pre class=\"prettyprint language-default\" style=\"padding-left: 40px;\"><code class=\"language-default\">\"headers\": { \r\n     \"Ocp-Apim-Subscription-Key\": \"&lt;your-foundry-resource-key&gt;\" \r\n}<\/code><\/pre>\n<ul>\n<li>Save the file then click to start the server connection. A complete <em>json<\/em> would look like this:<\/li>\n<\/ul>\n<p><figure id=\"attachment_1511\" aria-labelledby=\"figcaption_attachment_1511\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-3-mcp-in-copilot.webp\"><img decoding=\"async\" class=\"wp-image-1511\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-3-mcp-in-copilot.webp\" alt=\"MCP settings in mcp.json\" width=\"700\" height=\"173\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-3-mcp-in-copilot.webp 624w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-3-mcp-in-copilot-300x74.webp 300w\" sizes=\"(max-width: 700px) 100vw, 700px\" \/><\/a><figcaption id=\"figcaption_attachment_1511\" class=\"wp-caption-text\"><em>Figure 3: MCP settings in mcp.json<\/em><\/figcaption><\/figure><\/p>\n<p>Now you can prompt GitHub Copilot to call the Azure Language tools directly!<\/p>\n<p><figure id=\"attachment_1510\" aria-labelledby=\"figcaption_attachment_1510\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-4-test-mcp-in-copilot.webp\"><img decoding=\"async\" class=\"wp-image-1510 size-full\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-4-test-mcp-in-copilot.webp\" alt=\"Test PII redaction tool in Azure Language MCP using Visual Studio Code GitHub Copilot\" width=\"624\" height=\"561\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-4-test-mcp-in-copilot.webp 624w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-4-test-mcp-in-copilot-300x270.webp 300w\" sizes=\"(max-width: 624px) 100vw, 624px\" \/><\/a><figcaption id=\"figcaption_attachment_1510\" class=\"wp-caption-text\"><em>Figure 4: Test PII redaction tool in Azure Language MCP using Visual Studio Code GitHub Copilot<\/em><\/figcaption><\/figure><\/p>\n<h2>Protecting Privacy with Confidence through PII Redaction<\/h2>\n<p>We\u2019re excited to unveil several major updates to our Text PII Redaction service, making it easier than ever to detect and mask sensitive and personally identifiable information (PII) like phone numbers, emails, and credit card numbers. The PII Redaction service handles customer data in text, conversation, and native-doc modalities and returns a JSON of redacted text and a comprehensive list of the detected sensitive PII, classification, confidence score, and more. The Text PII modality is now enhanced with several new updates.<\/p>\n<ul>\n<li><strong>Test Instantly<\/strong>: Try our\u00a0new demo\u00a0playground on Microsoft Foundry to quickly\u00a0validate\u00a0your redaction workflows.\u00a0Try it out\u00a0<a href=\"https:\/\/ai.azure.com\">on the Foundry platform <\/a>or\u00a0<a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/personally-identifiable-information\/quickstart?tabs=windows&amp;pivots=ai-foundry-portal\">review the\u00a0quickstart\u00a0tutorial<\/a>.<\/li>\n<li><strong>Smarter Anonymization<\/strong>: Automatically replace real PII with realistic, synthetic data for safer sharing and analysis.\u00a0With the new\u00a0syntheticReplacement\u00a0redaction\u00a0policy, &#8220;John Doe received a call from 424-878-9193&#8221; can be transformed into &#8220;Sam Johnson received a call from 401-255-6901.&#8221;\u202fTry it out with\u00a0our\u00a0<a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/personally-identifiable-information\/how-to\/redact-text-pii#redaction-policy-version-2024-11-5-preview-only\">how-to guide<\/a>.<\/li>\n<li><strong>Quality Improvements<\/strong>: Enjoy better accuracy, customizable confidence thresholds, and\u00a0higher AI quality with\u00a0fewer false positives.\u00a0A comprehensive list of our new releases is\u00a0summarized\u00a0on\u00a0our\u202f<a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/whats-new?tabs=csharp\"><em>What\u2019s New<\/em>\u202fpage<\/a>.<\/li>\n<\/ul>\n<p>The Language tools team engages regularly with enterprise customers to deliver state-of-the-art, customer-obsessed products, as exhibited in the customer testimony from Nationwide Building Society (NBS):<\/p>\n<blockquote><p><em>&#8220;NBS and Microsoft have partnered to elevate PII redaction accuracy from 81% to over 90%, delivering greater confidence in data privacy and ensuring regulatory compliance remains central to AI system design and implementation.&#8221;<\/em><\/p><\/blockquote>\n<h3>Test instantly with Foundry playground experience<\/h3>\n<p>Seamlessly test, configure, and validate your real-world customer data scenarios on Microsoft Foundry. Tailor every aspect of your workflow, choose from multiple API and model versions to optimize AI quality, set your preferred input language, and fine-tune redaction policies to safeguard sensitive information (with options to replace sensitive data with a character mask \u2018***\u2019 or entity-label mask \u2018[PHONENUMBER_3]\u2019). Dive deeper with advanced analytics: view confidence scores, character offsets, entity lengths, and detailed tags for comprehensive, granular insights.<\/p>\n<p><figure id=\"attachment_1570\" aria-labelledby=\"figcaption_attachment_1570\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-playground.webp\"><img decoding=\"async\" class=\"wp-image-1570\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-playground-1024x506.webp\" alt=\"Try out Text PII redaction in New Foundry portal playground\" width=\"800\" height=\"396\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-playground-1024x506.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-playground-300x148.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-playground-768x380.webp 768w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-playground-1536x760.webp 1536w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-playground.webp 1905w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1570\" class=\"wp-caption-text\"><em>Figure 5: Try out Text PII redaction in New Foundry portal playground<\/em><\/figcaption><\/figure><\/p>\n<p><figure id=\"attachment_1569\" aria-labelledby=\"figcaption_attachment_1569\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-json.webp\"><img decoding=\"async\" class=\"wp-image-1569\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-json.webp\" alt=\"Sample Text PII JSON output from Foundry portal playground. API call output includes the redacted text and a list of all detected entities along with entity type classification, character length and offset, confidence score, and Tags, which are now Generally Available\" width=\"800\" height=\"488\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-json.webp 965w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-json-300x183.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/pii-json-768x469.webp 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1569\" class=\"wp-caption-text\"><em>Figure 6: Sample Text PII JSON output from Foundry portal playground. API call output includes the redacted text and a list of all detected entities along with entity type classification, character length and offset, confidence score, and Tags, which are now Generally Available<\/em><\/figcaption><\/figure><\/p>\n<h3>Quality improvements with updated model<\/h3>\n<p>The 2025-11-15-preview model includes new entities:<\/p>\n<ul>\n<li>Airport<\/li>\n<li>City<\/li>\n<li>Geopolitical Entity<\/li>\n<li>Korean Drivers License Number<\/li>\n<li>Korean Passport Number<\/li>\n<li>Location<\/li>\n<li>State<\/li>\n<li>Zip Code<\/li>\n<\/ul>\n<p>And significant AI quality improvements in entities:<\/p>\n<ul>\n<li>Date of Birth<\/li>\n<li>Vehicle Identification Number<\/li>\n<li>License Plate<\/li>\n<li>SORT Code<\/li>\n<\/ul>\n<h2>Deterministic intent routing in AI Agents with CLU<\/h2>\n<p>Conversational AI is entering a new era\u2014one where intent routing is smarter, more flexible, and easier to implement than ever. In May 2025, we introduced the <strong>Intent Routing Agent<\/strong> to help developers orchestrate complex interactions by intelligently routing user requests to the right skill or action based on predicted intents. As conversations with agents become more complex, our agents also need to become smarter.<\/p>\n<p>We\u2019re excited to announce new understanding at the conversation-level (not just single utterances) to understand the true intent by a user.<\/p>\n<h3>Understanding the whole conversation<\/h3>\n<p>As humans, we use short utterances, like \u201cThat\u2019s correct!\u201d or \u201cTomorrow\u201d to make meaningful points in our conversations. While these phrases alone don\u2019t provide much context for a human task, in the full conversational context, these are critical phrases to make a point quickly. We\u2019re happy to announce that CLU not only understands individual utterances, but it can make sense of the full conversation to identify the true intent of the conversation.<\/p>\n<p>By simply turning on <strong>Multi-turn understanding <\/strong>in the Microsoft Foundry playground, you can enter full conversations to provide context for the true utterance.<\/p>\n<p><figure id=\"attachment_1572\" aria-labelledby=\"figcaption_attachment_1572\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-7-clu-playground.webp\"><img decoding=\"async\" class=\"wp-image-1572\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-7-clu-playground.webp\" alt=\"Enable multi-turn understanding in the Language playground\" width=\"800\" height=\"422\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-7-clu-playground.webp 1430w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-7-clu-playground-300x158.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-7-clu-playground-1024x540.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-7-clu-playground-768x405.webp 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1572\" class=\"wp-caption-text\"><em>Figure 7: Enable multi-turn understanding in the Language playground<\/em><\/figcaption><\/figure><\/p>\n<h3>Improved entity predictions with slot-filling<\/h3>\n<p>In addition to supporting intent prediction on conversations, we\u2019re excited to announce that CLU now supports slot-filling. This means that you can associate a specific list of entities with a given intent, so your system can process the data that\u2019s most important to the business process flow. This slot-filling also helps you to identify which entities are missing to help you determine what to ask next. To see this feature in action, suppose you\u2019re setting up a flight booking agent.<\/p>\n<p>First, you can add your desired intents and entities with short natural language descriptions into Microsoft Foundry.<\/p>\n<p><figure id=\"attachment_1573\" aria-labelledby=\"figcaption_attachment_1573\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-8-add-intent.webp\"><img decoding=\"async\" class=\"wp-image-1573\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-8-add-intent.webp\" alt=\"Add intents and descriptions\" width=\"800\" height=\"347\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-8-add-intent.webp 1430w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-8-add-intent-300x130.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-8-add-intent-1024x445.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-8-add-intent-768x334.webp 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1573\" class=\"wp-caption-text\"><em>Figure 8: Add intents and descriptions<\/em><\/figcaption><\/figure><\/p>\n<p><figure id=\"attachment_1574\" aria-labelledby=\"figcaption_attachment_1574\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-9-add-entities.webp\"><img decoding=\"async\" class=\"wp-image-1574\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-9-add-entities.webp\" alt=\"Add entities and descriptions\" width=\"800\" height=\"351\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-9-add-entities.webp 1430w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-9-add-entities-300x132.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-9-add-entities-1024x450.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-9-add-entities-768x337.webp 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1574\" class=\"wp-caption-text\"><em>Figure 9: Add entities and descriptions<\/em><\/figcaption><\/figure><\/p>\n<p>Second, to set up slot-filling, toggle to the Associations tab to link each entity with at least one intent. You can also switch to the entity view to associate by entity. Note: when using slot filling, every entity must be associated with an intent.<\/p>\n<p><figure id=\"attachment_1575\" aria-labelledby=\"figcaption_attachment_1575\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-10-associate-entities.webp\"><img decoding=\"async\" class=\"wp-image-1575\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-10-associate-entities.webp\" alt=\"Associate entities with intents\" width=\"800\" height=\"392\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-10-associate-entities.webp 1430w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-10-associate-entities-300x147.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-10-associate-entities-1024x502.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-10-associate-entities-768x376.webp 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1575\" class=\"wp-caption-text\"><em>Figure 10: Associate entities with intents<\/em><\/figcaption><\/figure><\/p>\n<p>Once your associations are complete, click <strong>Quick deploy with LLM <\/strong>and add your own LLM deployment to create your new CLU deployment. In a matter of minutes, you can try out your custom deployment in the Language playground.<\/p>\n<p><figure id=\"attachment_1576\" aria-labelledby=\"figcaption_attachment_1576\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-11-enable-multi-turn.webp\"><img decoding=\"async\" class=\"wp-image-1576\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-11-enable-multi-turn.webp\" alt=\"Enable multi-turn understanding in the Language playground\" width=\"800\" height=\"422\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-11-enable-multi-turn.webp 1430w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-11-enable-multi-turn-300x158.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-11-enable-multi-turn-1024x540.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-11-enable-multi-turn-768x405.webp 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1576\" class=\"wp-caption-text\"><em>Figure 11: Enable multi-turn understanding in the Language playground<\/em><\/figcaption><\/figure><\/p>\n<p>After clicking <strong>Run<\/strong>, you can see the parsed conversation, the predicted top intent, and the associated slots.<\/p>\n<p><figure id=\"attachment_1577\" aria-labelledby=\"figcaption_attachment_1577\" class=\"wp-caption alignnone\" ><a href=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-12-test-in-playground.webp\"><img decoding=\"async\" class=\"wp-image-1577\" src=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-12-test-in-playground.webp\" alt=\"See intent predictions with the associated entities\" width=\"800\" height=\"430\" srcset=\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-12-test-in-playground.webp 1430w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-12-test-in-playground-300x161.webp 300w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-12-test-in-playground-1024x550.webp 1024w, https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2025\/11\/figure-12-test-in-playground-768x412.webp 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/><\/a><figcaption id=\"figcaption_attachment_1577\" class=\"wp-caption-text\"><em>Figure 12: See intent predictions with the associated entities<\/em><\/figcaption><\/figure><\/p>\n<p>Learn more about multi-turn conversational understanding in <a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/ai-services\/language-service\/conversational-language-understanding\/how-to\/quickstart-multi-turn-conversations\">Quickstart: Multi-turn conversational language understanding (CLU) models with entity slot filling<\/a>.<\/p>\n<h2>Summary<\/h2>\n<p>With the introduction of the Azure Language remote MCP server, enhanced PII redaction, and deterministic CLU routing, Azure Language in Foundry Tools provides a powerful and privacy-first Language AI foundation for modern agentic solutions.<\/p>\n<p>Whether you\u2019re building compliance-critical automation, customer service agents, or intelligent workflow systems, Azure Language gives you deterministic behavior, privacy protections, and tooling integration you need to innovate with confidence.<\/p>\n<p>We\u2019d love to hear your thoughts. Please share your comments, feedback or questions. We can\u2019t wait to see what you build!<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In today\u2019s rapidly evolving AI landscape, developers are seeking reliable, secure, and predictable language capabilities to power the next generation of enterprise-grade agents. As agentic architecture becomes central to modern applications, teams need tools that deliver stronger privacy guarantees, deterministic behavior, and seamless integration across their AI stack. As part of the broader transition from [&hellip;]<\/p>\n","protected":false},"author":188951,"featured_media":1563,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[49,45,51,1,76],"tags":[3,5,10,6,80,9,2,77,78,79],"class_list":["post-1479","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-aiagent","category-azure-ai-services","category-mcp","category-microsoft-foundry","category-msignite","tag-ai-development","tag-ai-tools","tag-ai-agents","tag-azure-ai","tag-intent-routing","tag-mcp","tag-microsoft-foundry","tag-msignite","tag-pii-redaction","tag-privacy-control"],"acf":[],"blog_post_summary":"<p>In today\u2019s rapidly evolving AI landscape, developers are seeking reliable, secure, and predictable language capabilities to power the next generation of enterprise-grade agents. As agentic architecture becomes central to modern applications, teams need tools that deliver stronger privacy guarantees, deterministic behavior, and seamless integration across their AI stack. As part of the broader transition from [&hellip;]<\/p>\n","_links":{"self":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/1479","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/users\/188951"}],"replies":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/comments?post=1479"}],"version-history":[{"count":0,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/posts\/1479\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media\/1563"}],"wp:attachment":[{"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/media?parent=1479"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/categories?post=1479"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/foundry\/wp-json\/wp\/v2\/tags?post=1479"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}