Azure AI Search-Retrieval-Augmented Generation

Overview

Power agentic RAG, search and retrieval

Focus on exponential growth with an enterprise-ready vector database that comes with security, compliance, and responsible AI practices built in.

Watch the customer video
Build better applications with sophisticated retrieval strategies backed by decades of research and customer validation.

Read the blog
Accelerate AI app development and simplify your RAG pipeline with seamless platform and data integrations for data sources, AI models, and frameworks.

Watch the video

Data ingestion

Integrated data pipeline

Vector database

Search and knowledge retrieval

Security

Platform integrations

Simplified data ingestion

Automatically upload data from a wide range of supported Azure and third-party sources.

a computer-generated image of several objects

USE CASES

Explore common applications to get started

Generative AI app with real-time audio

Build a VoiceRAG application for speech-to-speech interactions with the Azure OpenAI gpt-4o-real time preview model.

Vector search

Move beyond text to create image search engines and multimodal experiences for images, PDFs, and audio.

AI assistant

Gain better analytics and improve support with an agent that provides relevant information to answer a question fast.

Retrieval-augmented generation (RAG)

Ground your generative AI application with proprietary data.

A woman sitting on a bench holding a laptop and a cup of coffee.

Pricing

Azure AI Search pricing

Azure AI Search has pricing tiers that scale to meet the storage and throughput needs of large scale RAG applications.

See Azure AI Search pricing

Azure products work better together

Azure AI Foundry Agent Service

Create AI agents that automate complex business processes—completing specific tasks while keeping humans in control.

Learn more

Azure OpenAI in Foundry Models

Explore cutting-edge models for your copilot and generative AI applications.

Learn more

Azure AI Foundry

Get everything you need to develop generative AI applications and custom copilots on one platform.

Learn more

Microsoft Copilot Studio

Customize Microsoft 365 Copilot or build your own agent experiences.

Learn more

Azure AI Content Understanding

Accelerate the transformation of multimodal data into insights.

Learn more

Customer stories

See how customers innovate with Azure AI Search

"Azure AI Search helped us provide both vector and hybrid search, which significantly improved the usability of the service and provided the right context for searches."

John Shang, Director, Data Engineering and EHR Systems, BILH

A white counter with a white and black Fujitsu device.

“The particularly useful aspect was the ease of integration between Foundry Agent Service (previously Azure AI Agent Service) and Azure AI Search—it saved us time, and the implementation was very easy from the Azure portal.”

Hirotaka Ito, Lead Engineer of AI, Digital System Platform Unit, Fujitsu

A tree with green leaves in front of a building with a glass door.

UBS uses Azure AI Search to streamline legal searches and quickly access specific information in a vast library of documents. This has improved productivity and accuracy, helping UBS better serve clients.

A woman sitting in a chair looking at a computer screen.

"Customers will have greater success with our product if we augment it with context. Our Azure-based bot uses retrieval-augmented generation (RAG) to guide customers through coding in our specific Python dialect to work in their particular application domains."

Charles Crain, Chief Technology Officer, Artificial

RESOURCES

Get the latest Azure AI Search news and resources

A close-up of a man wearing a black jacket.

Microsoft Learn plan

Create a custom RAG solution

Discover how to customize with RAG in Azure AI Foundry, Azure AI Foundry SDK, or Azure OpenAI to search documents for answers.

Get started

Report

Customizing generative AI for unique value

Get expert insights on how customizing AI models is helping the world's leading companies drive more value in this new report from Microsoft and MIT Technology Review Insights.

Read the blog

Research

Raising the bar for RAG excellence

See how query rewriting and a new reranking model improve RAG response quality.

Learn more

Developer series

RAG Time: Ultimate Guide to Mastering RAG

Learn how to build AI retrieval systems with videos, hands-on demos, and code samples.

Learn more

A woman laughing while another woman leans on her arm.

Documentation

Azure AI Search documentation

Read an introduction to Azure AI Search.

Learn more

A woman sitting at a desk using a laptop.

GitHub

Azure AI GitHub samples

Explore community-driven Azure AI examples.

Browse samples

FAQ

Frequently asked questions

Azure AI Search is an enterprise knowledge retrieval system that powers sophisticated retrieval-augmented generation (RAG) applications and enterprise search engines. With advanced AI and search technology, and seamless platform integrations, Azure AI Search delivers end-to-end RAG systems built for app excellence, enterprise-readiness, and speed to market.
RAG excellence: Set the bar for app excellence with the latest rigorously tested AI technology, delivered ahead of the market.
Velocity: Built for end-to-end RAG development, move from idea to production fast with integrations and features across the toolchain.
Enterprise scale: Deliver high-performance RAG applications without compromising scale or cost.
Hybrid search in Azure AI Search executes a query for both text search (or keyword search) and vector search in one request. The results are then merged together using Reciprocal Rank Fusion (RRF). Hybrid search is useful for RAG scenarios—vector search is effective at finding information from queries posed in natural language and full text search is able to find specific data like someone’s name or a product code.
RAG is an AI technique that combines retrieval-based techniques with generative models to produce accurate and contextually relevant responses. It retrieves information from external sources like databases, documents, or the web to enhance the generation of results. This ensures that responses generated by large language models are accurate, relevant, and enriched with the most up-to-date information. Learn more about RAG here and RAG in Azure AI Search here.
Read the Azure Service Level Agreements (SLA).
Azure AI Search is available in more than 30 countries and regions worldwide, with new locations added regularly.
Built-in indexers are available for Azure Cosmos DB, Azure SQL Database, Azure Blob Storage, and Microsoft SQL Server hosted in Azure Virtual Machines. Use Azure Data Factory, with more than 80 connectors, or Azure Logic Apps to connect to your data source. Alternatively, push data into an Azure AI Search index, which has no restrictions on data source type.
Azure AI Search works with several popular formats, including Microsoft Word, PowerPoint, and Excel; Adobe PDF; and PNG, RTF, JSON, HTML, and XML.
Azure AI Search is an enterprise retrieval and search engine used in custom apps that supports vector, full-text, and hybrid search over an indexed database. It’s used in RAG-based applications on Azure and integrates with Azure OpenAI and Azure AI Foundry models. Microsoft Search enables the ability to search across Microsoft 365 apps, such as Outlook and SharePoint, to help users find relevant information within their organization.
Azure AI Search and Azure OpenAI are two different services.
Azure AI Search is a knowledge base and retrieval system built for RAG and enterprise search applications. Azure OpenAI provides AI models and agent development tools for enterprise applications. Azure AI Search and Azure OpenAI are used together for AI agents that rely on private, enterprise data.
Azure AI Search supports vector search, keyword search, and hybrid search, combining vector and non-vector fields in the same search corpus. It stores the data you query over, allowing it to function as a vector store for applications that require long-term memory, a knowledge base, or grounding data for retrieval-augmented generation (RAG) architecture. This makes Azure AI Search a versatile tool that can be used as a vector database when needed.
Azure AI Search uses advanced algorithms like exhaustive k-nearest neighbors (KNN) and Hierarchical Navigable Small World (HNSW) for approximate nearest neighbor (ANN) search, enabling vector similarity queries to find semantically similar information. It also employs BM25 relevance scoring for full-text search, computing search scores based on the strength of the match.
Azure AI Search is built on an enterprise-ready foundation that supports diverse data types and retrieval methods. It integrates vector, full-text, and hybrid search capabilities, enabling developers to store, index, and deliver search applications over various data formats. The system supports advanced indexing, JSON document indexing, and AI enrichment through cognitive skills, enhancing search relevance with semantic ranking.
Azure AI Search provides native indexing support for multiple databases, including SQL databases like Azure SQL Database and Azure SQL Managed Instance, as well as NoSQL databases and various data stores such as Azure Blob Storage, ADLS Gen2, and Azure Tables through built-in indexers. For other data sources, content can be integrated into AI Search indexes using Azure Logic App connectors. The Azure AI Search Push API allows for maximum flexibility by enabling data to be pushed in a specific format regardless of the source.
As of November 2023, Azure Cognitive Search is now Azure AI Search. All existing deployments, configurations, and integrations continue to function as they previously did. There are no changes to pricing and no breaking changes to APIs or SDKs.