Simplifying Search Development with the New Azure AI Search Wizard

AI Search is set to continue evolving. It is becoming a RAG agent.

Search technology continues to evolve, and Azure AI Search is providing updates to simplify workflows and improve developer productivity. Based on customer feedback and real-world usage scenarios, we have announced a new experience for the
"Import Data" wizard in the Azure portal. This update allows for more efficient index creation and makes it even easier to integrate search functionality into applications through a low-code interface.

What is the new "Import Data" wizard?

Azure AI Search has rolled out the "Import Data (new)" wizard, which integrates the functionality of the traditional keyword search "Import Data" wizard with the vectorization wizard features for RAG and multimodal RAG workflows.

This new wizard makes the processes of index creation, skillset configuration, and data source connection more intuitive and modern, allowing you to easily leverage the latest features of Azure AI Search (such as applied AI, vectorization, and multimodal search). You can quickly prepare an index that is ready to connect to your application code or orchestrator.

Key Changes

• Integrated experience:
  The functionality of the traditional "Import Data" wizard has been integrated into the new "Import Data (new)" wizard. All wizards are now consolidated into a single interface.

• Expanded functionality:
  The new wizard supports keyword search, RAG, and multimodal RAG, and can seamlessly connect to embedding model deployments for vector generation and chat completion models for image captioning.

• Phased rollout:
  Both the "Import Data" and "Import Data (new)" wizards will coexist during a transition period. A banner will be displayed to encourage migration to the new experience. The traditional wizard is scheduled to be retired eventually.

Why should you switch to the new wizard?

The new Import Data (new) wizard provides Azure AI Search users with faster implementation and experimentation environments. The main benefits are as follows:

✅ Streamlined index creation

Simplifies all steps from data source connection to schema generation, reducing time and effort.

✅ Integrated vectorization support

Unlocks advanced search scenarios such as RAG and multimodal RAG. Easily integrates with Azure OpenAI embedding models and Azure AI Vision multimodal skills.

✅ Improved search performance

Semantic ranker and advanced enrichment features improve the relevance of keyword search results.

✅ Expanded data source options

Connect to new data sources such as Azure Queues, OneDrive for Business, and SharePoint Online using Azure Logic Apps connectors.

Comparison table of old and new wizards

Note:
If you configure the RAG multimodal flow to reference images during queries, the knowledge store will be automatically configured.

Step-by-step flow of the new wizard

1️⃣ Connect to data source

Supports a wide range of data sources such as Azure Blob Storage, SQL Database, and Cosmos DB. Configure connection information and access settings.

2️⃣ Customize to your solution

You can choose from the following options.
Semantic ranking features can be added to any of these.

• Keyword:
  For traditional search using text-based relevance and structured queries.
  Phrase extraction, OCR, entity extraction, text splitting, etc., are available.

• RAG:
  Build a RAG-ready index that integrates text-based embedding generation (chunking + vectorization). Image processing is not required.

• Multimodal RAG:
  For multimodal applications including image analysis, captioning, chunking, and vectorization.
  Saves images in the knowledge store for reference during search.

3️⃣ Configure skillsets

Depending on the selected option, you can configure the following AI enrichment features.

• Extract useful text information from unstructured data

• Document Intelligence Layout

• Azure AI Vision Multimodal

• Azure OpenAI embedding models

• OCR

• GenAI prompts for image captioning (Chat Completion), etc.

4️⃣ Define index schema

Automatically samples the data source to infer schema and data types.
Customize attributes such as searchable, filterable, retrievable, and sortable.

5️⃣ Run the wizard

Create objects such as indexes, skillsets, indexers, and data sources to import and process your data.
Once processing is complete, you can check the progress and any errors in the 'Indexers' section on the left pane.

Summary of the new wizard's benefits

The new Import Data (new) wizard minimizes complexity by integrating workflows and combining the latest AI capabilities.
It supports vector search, multimodal features, and semantic ranking, enabling you to build smarter, faster, and more accurate search solutions.

Next steps

Please try out the new 'Import Data (new)' wizard.
From experienced developers to Azure AI Search beginners, anyone can easily build powerful search solutions.
Feedback can be submitted via the link at the bottom of the wizard.

Reference materials

• Documentation: Import Data Wizards Overview

• Community Support: Microsoft Q&A