Here’s a copy-paste portal lab for Azure AI Search that gets you from zero to a working searchable index using Blob Storage + the Azure portal.

This follows Microsoft’s current portal flow: create a search service, upload sample data to Blob Storage, run the Import data wizard, let it create the data source, index, and indexer, and then test queries in the portal. The wizard also supports keyword, RAG, and multimodal RAG scenarios, and the vector quickstart uses integrated vectorization for chunking and embeddings. (Microsoft Learn)

What you’ll build

Blob Storage
   ↓
Import data wizard
   ↓
Data source + Index + Indexer
   ↓
Search Explorer

Azure AI Search queries run over content loaded into a search index, and the portal wizard can create the ingestion pieces for supported sources like Blob Storage. (Microsoft Learn)

Portal lab: basic keyword search

1. Create the Azure AI Search service

In the Azure portal:

• Search for Azure AI Search
• Click Create
• Choose:
  • Subscription
  • Resource group
  • Service name
  • Region
  • Pricing tier

Microsoft’s portal guide says you create the service directly in the portal and choose region and tier there. (Microsoft Learn)

2. Create a storage container

In Azure portal:

• Open Storage accounts
• Create or open one
• Go to Data storage → Containers
• Create a container

Microsoft’s full-text portal quickstart uses a sample container named hotels-sample-data. (Microsoft Learn)

Suggested container name:

search-lab-data

3. Upload sample documents

Upload a few .json, .txt, or .pdf files to the container.

Good starter options:

• product descriptions
• FAQ files
• policy text
• hotel/sample catalog style JSON

The portal quickstart supports using your own files, not only Microsoft’s sample data. (Microsoft Learn)

4. Open the Import data wizard

In your Azure AI Search resource:

• Go to Overview
• Click Import data

Microsoft’s portal quickstart uses this exact entry point. (Microsoft Learn)

5. Choose the data source

Select:

• Azure Blob Storage

Then provide:

• Storage subscription
• Storage account
• Container name

Azure AI Search supports a pull/indexer model for supported sources like Blob Storage. (Microsoft Learn)

6. Choose search type

For the first run, select:

Keyword search

Microsoft’s full-text portal quickstart uses Keyword search as the basic first lab. (Microsoft Learn)

7. Configure the index

The wizard will infer fields from your documents and create an index.

Typical fields look like this:

id           key, retrievable
content      searchable, retrievable
title        searchable, filterable, retrievable
category     filterable, facetable, retrievable
lastUpdated  sortable, filterable, retrievable

In Azure AI Search, each document needs a unique key field in the index. (Microsoft Learn)

A simple example index shape:

{
  "name": "docs-index",
  "fields": [
    { "name": "id", "type": "Edm.String", "key": true, "searchable": false },
    { "name": "title", "type": "Edm.String", "searchable": true },
    { "name": "content", "type": "Edm.String", "searchable": true },
    { "name": "category", "type": "Edm.String", "filterable": true, "facetable": true }
  ]
}

8. Create and run the indexer

Let the wizard create an indexer and run it.

The indexer is what pulls data from Blob Storage into the search index. Azure documents indexers as the automated ingestion mechanism for supported data sources. (Microsoft Learn)

9. Test in Search Explorer

After indexing finishes:

• Open Indexes
• Select your index
• Open Search Explorer

Try queries like:

loan
interest rate
policy
support

Azure AI Search supports full-text querying over the indexed documents once content is loaded. (Microsoft Learn)

What got created

You should now have:

• Search service
• Data source
• Index
• Indexer
• Indexed documents you can query in the portal

That is the standard Azure AI Search portal pattern for supported data sources. (Microsoft Learn)

Recommended field choices

When the wizard shows fields, a good rule is:

• id → key
• title → searchable
• content → searchable
• category → filterable/facetable
• dates → sortable/filterable

That gives you a useful first search app without overcomplicating the schema. The key-field requirement is mandatory in Azure AI Search indexes. (Microsoft Learn)

Upgrade path: vector + RAG lab

Once the keyword lab works, do a second run using the vector wizard.

Use:

• Import data
• choose the vector/RAG path
• let the wizard chunk content and vectorize it

Microsoft’s vector portal quickstart says the wizard can chunk your content and call an embedding model to vectorize chunks at indexing and query time. (Microsoft Learn)

For production, hybrid search is usually better than vector-only because Azure AI Search can run full-text and vector queries in parallel and merge results. (Microsoft Learn)

Fast troubleshooting

If indexing fails, the usual causes are:

• Blob container access/config is wrong
• unsupported or malformed files
• index schema/key field problems
• service can’t access the data source

Microsoft’s ingestion docs and Q&A guidance both point to data-source access and ingestion errors as common causes. (Microsoft Learn)

Best next step

After this lab, the most useful next exercise is:

1. repeat it with the vector quickstart
2. compare keyword search vs hybrid/vector search
3. then connect it to Azure OpenAI for a RAG chatbot. (Microsoft Learn)