Notice: Microsoft has announced the retirement of eight legacy Azure Language features: key phrase extraction, sentiment analysis and opinion mining, custom text classification, conversational language understanding (CLU), custom question answering, orchestration workflow, summarization, and entity linking. As of March 31, 2029 these capabilities no longer work through Language Studio, Microsoft Foundry, or the REST APIs. The core capabilities (PII detection, language detection, named entity recognition, text analytics for health) are not affected.
What is Azure Language?
Azure Language is a Foundry Tools service from Microsoft Azure for Natural Language Processing that automatically analyzes and understands text. Microsoft distinguishes between actively evolving core capabilities and legacy capabilities that remain supported for existing implementations.
Core capabilities include PII detection (including a preview feature for anonymization via synthetic replacement), language detection, prebuilt and custom named entity recognition, and text analytics for health for extracting medical information from unstructured text. Legacy capabilities that remain stable and supported include sentiment analysis and opinion mining, key phrase extraction, entity linking, custom text classification, conversational language understanding, orchestration workflow, question answering, and summarization for text, conversations, and call center transcripts (extractive and abstractive).
For integration into AI agents, Azure Language provides an MCP server (Model Context Protocol), available as a remote server through the Foundry Tool Catalog or as a locally hosted server. It also offers prebuilt agents such as the Intent Routing agent for managing conversation flows and the Exact Question Answering agent for word-for-word answers to common business questions (both currently in preview).
Core Features
- PII detection: identifies personally identifiable information in text and conversations, including a preview anonymization feature
- Named entity recognition: available prebuilt or trainable on custom entity categories via custom NER
- Sentiment analysis and opinion mining: detects sentiment and links it to specific text elements
- Key phrase extraction: extracts the main concepts and topics from documents
- Summarization: extractive and abstractive summarization of text, conversations, and call center transcripts
- Conversational language understanding: custom models for intent and parameter recognition for chatbots and voice assistants
- Azure Language MCP server and prebuilt agents for integration into agent-based applications
Typical Use Cases
Customer feedback analysis: Companies automatically analyze customer reviews, support tickets, and social media posts. Sentiment analysis and opinion mining help identify dissatisfied customers for prioritized handling.
Document processing and compliance: Contracts, emails, and reports are automatically classified and searched for relevant information. PII detection supports data protection requirements, while NER extracts contract parties, dates, and amounts for downstream processing.
Conversational AI: Chatbots and virtual assistants use conversational language understanding to interpret user intent and extract parameters from natural language queries. The same capabilities can be integrated into agent-based architectures via the MCP server.
Healthcare: Text analytics for health extracts structured medical information from clinical documents without requiring you to train your own model.
Benefits
- Preconfigured features for immediate use without in-house ML expertise
- Custom models trainable for domain-specific requirements
- REST API and SDKs for common programming languages (C#, Java, JavaScript, Python)
- MCP server and prebuilt agents for integration into modern agent architectures
- Container support for several features for on-premises operation for compliance or latency reasons
Integration with innFactory
As a Microsoft Solutions Partner, innFactory supports you with Azure Language: we implement NLP pipelines for document analysis, build intelligent chatbots and agents with language understanding, and integrate text analytics into your existing business processes.
Contact us for a non-binding consultation on Azure Language and Microsoft Azure.
Typical Use Cases
Frequently Asked Questions
What can Azure Language do?
Azure Language is a Foundry Tools service for automated text analysis: PII detection, named entity recognition, sentiment analysis, key phrase extraction, summarization, and language understanding for chatbots. Many features can be used without training your own model.
How does Azure Language differ from Azure OpenAI?
Azure Language offers specialized, mostly deterministic NLP functions such as entity recognition, PII detection, or text classification with predictable behavior. Azure OpenAI is better suited for generative tasks like free-form text generation, open-ended summarization, or conversing with an LLM.
Can I train custom models?
Yes, custom named entity recognition, custom text classification, and conversational language understanding allow training domain-specific models with your own training data. Other features like sentiment analysis or key phrase extraction are preconfigured and cannot be customized.
Which features are currently recommended versus legacy?
Core features with ongoing investment currently include PII detection, language detection, named entity recognition (prebuilt and custom), and text analytics for health. Features such as sentiment analysis, key phrase extraction, summarization, entity linking, custom text classification, conversational language understanding, orchestration workflow, and question answering are considered legacy features that remain supported for existing implementations.
Can I use Azure Language with AI agents?
Yes, the Azure Language MCP server lets you integrate language features as tools into agents, either as a remote server through the Foundry Tool Catalog or as a locally hosted server. Azure Language also offers prebuilt agents such as the Intent Routing agent and the Exact Question Answering agent (both currently in preview).
Can Azure Language be used in a GDPR-compliant way?
When using European Azure regions, the service can be operated in a GDPR-compliant manner; Microsoft offers corresponding Data Processing Agreements. For on-premises or edge requirements, Docker containers are additionally available for several features.
Note: All product information on this page has been compiled with care, but is provided without guarantee and may be outdated or incomplete. Cloud services evolve rapidly — features, pricing, SLAs, and availability change frequently. Authoritative and up-to-date information can only be found on the official product page of Azure (official documentation). This page does not represent an offer by Azure.
