Simplify Search Development with the New Azure AI Search Wizard

Introducing the New "Import Data" Wizard

Recently, Azure AI Search rolled out the "Import Data (new)" wizard, an enhanced experience that merges the capabilities of the older "Import Data" wizard for keyword search with the integrated vectorization wizard previously used for RAG and multimodal RAG workflows.

This update offers a more intuitive and modernized process for creating indexes, configuring skillsets, and connecting to data sources, ensuring that you can take advantage of the latest features available in Azure AI Search for applied AI, vectorization, and multimodal search with ease so you can have indexes ready to plug to your application code/orchestrator as needed.

Key Changes at a Glance:

• Unified Experience: An improved experience of the legacy "Import Data" wizard functionality is now part of the new "Import Data (new)" wizard, consolidating all wizards into a single interface.
• Enhanced Capabilities: The new wizard supports keyword search, RAG, and multimodal RAG while connecting seamlessly to embedding model deployments for vector generation and to chat completion models for image verbalization.
• Phased Rollout: Both wizards “Import Data” and “Import Data (new)” will coexist during the initial phase, with a banner guiding users to the new experience. The classic "Import Data" wizard will eventually be retired.

Why Switch to the New Wizard?

The Import Data (new) wizard brings a host of benefits to Azure AI Search users, enabling faster implementation and experimentation. Key features include:

1. Streamlined Index Creation: From connecting to a data source to generating a schema, the new wizard simplifies every step of building a query-ready index, saving time and effort.
2. Integrated Vectorization: Unlock advanced search scenarios with integrated support for vector and multimodal workflows, including RAG and multimodal RAG capabilities. Easily connect to Azure OpenAI embedding models or Azure AI Vision multimodal skills.
3. Improved Search Performance: Enable semantic ranking and advanced enrichment capabilities to deliver more relevant results for keyword search.
4. Expanded Data Source Options: Connect to new data sources like Azure Queues, OneDrive for Business, and SharePoint Online using Azure Logic Apps connectors.

Please note that the classic Import Data wizard will be retired after a transitional period of coexistence. We strongly recommend switching to the Import Data (new) wizard as soon as possible. This will allow you to familiarize yourself with the enhanced functionality and start getting its benefits right away.

Image 1 - Import data (new) wizard

What Can You Do with the New Wizard?

Whether you’re performing keyword searches or exploring vectorized search workflows, the Import Data (new) wizard has you covered. Here’s a breakdown of its capabilities:

** Note that knowledge store is automatically configured in the RAG Multimodal flow, if you configure to export your extracted images for reference at query time.

Step-by-Step Workflow for the New Wizard

Using the Import Data (new) wizard is simple and intuitive. Here’s a quick overview:

1. Connect to a Data Source:

• • Choose from a wide array of supported Azure data sources, including Blob Storage, SQL Database, Cosmos DB, and more.
  • Specify connections and configure access.

Image 2 - Choosing your data source (built-in and Logic Apps connectors)

2. Customize for your Solution

Choose from the following options. They all include the option to add semantic ranking functionality. Note that these options are available for built-in indexers only. Logic Apps connectors have different options that create a RAG-ready index (including chunking and vectorization).

• Keyword: For traditional search applications that rely on text-based relevance and structured queries. Includes optional capabilities for phrase extraction, OCR, entity extraction, and split text.
• RAG: Ideal when you need a RAG-ready index for your application, with integrated vectorization (chunking + embeddings), without requiring specialized image handling.
• Multimodal RAG: Designed for RAG scenarios that involve advanced embedded image analysis, verbalization, chunking and vectorization, and storage for later use, enabling index creation tailored to multimodal applications.

Image 3 - Scenarios offered: Keyword Search, RAG and Multimodal RAG

3. Configure Skillsets:

Depending on the option you choose, there are AI enrichment options available to:

• Extract meaningful text insights from unstructured data.
• Configure built-in skills like Document Intelligence layout, Azure AI Vision multimodal, Azure OpenAI embedding, OCR, GenAI prompt (Chat Completion) for image verbalization and more.

Image 4 - Keyword search built-in skills (Extract phrases, split text, extract entities, OCR

Image 5 - RAG built-in skills (Vectorize your text, enrich and vectorize images, etc.)

Image 6 - Multimodal RAG built-in skills (Document layout model, image verbalization, multimodal embeddings) and image output to knowledge store

4. Define Index Schema:

• • Automatically sample your data source to infer schema and data types.
  • Customize field attributes like searchable, filterable, retrievable, and sortable.

Image 7 - Index mappings

5. Run the Wizard:

• • Create objects like indexes, skillsets, and indexers.
  • Import and process your data to make it available for query. Note: This process may take some time to complete. After the wizard finishes, you can monitor the progress of your indexer by selecting it from the Indexers section in the left-hand pane of your search service. This view will show the current status and any errors encountered during execution.

Image 8 - Child object creation: index, indexer, skillset, datasource

Benefits of the New Wizard

By consolidating workflows and integrating modern AI capabilities, the Import Data (new) wizard surfaces the most used AI Search features while minimizing complexity. With support for vectorized search, multimodal capabilities, and semantic ranking, the new wizard enables you to build smarter, faster, and more accurate search solutions.

Next Steps

We encourage you to explore the Import Data (new) wizard and experience the enhanced capabilities firsthand. Whether you’re a seasoned developer or new to Azure AI Search, this streamlined workflow makes it easier than ever to build powerful search solutions. Also, if you have any feedback, please let us know via the link at the bottom of the wizard.

Image 8 - Feedback option

Resources

• Documentation: Import Data Wizards Overview
• Community Support: Microsoft Q&A