Azure AI Search: Evolutions for RAG

Azure AI Search Strengthens for RAG

Microsoft announces significant improvements to Azure AI Search (formerly Azure Cognitive Search), with particular focus on RAG use cases. Vector search becomes more performant, integrations more numerous, and semantic ranking more precise.

"Azure AI Search is now the foundation of choice for enterprise RAG on Azure," declares Satya Nadella at Build 2026. "Native integration with Azure OpenAI creates a unified experience."

New Features

Improved Vector Search

Vector search evolves on multiple axes:

DEVELOPERcsharp
// Advanced vector search configuration
var searchIndex = new SearchIndex("documents")
{
    VectorSearch = new VectorSearch
    {
        Algorithms = {
            new HnswAlgorithmConfiguration("hnsw-config")
            {
                Parameters = new HnswParameters
                {
                    Metric = VectorSearchAlgorithmMetric.Cosine,
                    M = 16,
                    EfConstruction = 400,
                    EfSearch = 500
                }
            }
        },
        Profiles = {
            new VectorSearchProfile("vector-profile", "hnsw-config")
            {
                Compression = new ScalarQuantizationCompression("sq-config")
            }
        }
    }
};

Semantic Ranking 2.0

Semantic reranking improves:

DEVELOPERcsharp
var options = new SearchOptions
{
    SemanticSearch = new SemanticSearchOptions
    {
        SemanticConfigurationName = "semantic-config",
        QueryCaption = new QueryCaption(QueryCaptionType.Extractive)
        {
            HighlightEnabled = true
        },
        QueryAnswer = new QueryAnswer(QueryAnswerType.Extractive)
        {
            Count = 3,
            Threshold = 0.7
        }
    }
};

These features align with our guide on reranking.

Azure OpenAI Integration

Integration becomes seamless:

DEVELOPERcsharp
// Indexing with Azure OpenAI embeddings
var skillset = new SearchIndexerSkillset("rag-skillset")
{
    Skills = {
        new AzureOpenAIEmbeddingSkill
        {
            Name = "embedding-skill",
            ModelName = "text-embedding-3-large",
            ResourceUri = "https://my-openai.openai.azure.com",
            Inputs = { new InputFieldMappingEntry("text") { Source = "/document/content" } },
            Outputs = { new OutputFieldMappingEntry("embedding") { TargetName = "vector" } }
        }
    }
};

// RAG with Azure OpenAI
var chatClient = new ChatClient(endpoint, credential);
var response = await chatClient.CompleteChatAsync(new ChatCompletionOptions
{
    AzureExtensions = {
        new AzureChatExtensionConfiguration
        {
            Type = AzureChatExtensionType.AzureCognitiveSearch,
            Parameters = {
                Endpoint = searchEndpoint,
                IndexName = "documents",
                SemanticConfiguration = "semantic-config",
                QueryType = AzureCognitiveSearchQueryType.VectorSemanticHybrid
            }
        }
    }
});

Document Intelligence Integration

Document extraction is enriched:

• Advanced OCR: Tables, charts, formulas
• Layout analysis: Structure preserved
• Custom models: Training on your documents

DEVELOPERcsharp
var skillset = new SearchIndexerSkillset("doc-processing")
{
    Skills = {
        new DocumentExtractionSkill
        {
            Name = "doc-extraction",
            ParsingMode = BlobExtractionMode.AllBlob,
            DataToExtract = BlobDataToExtract.ContentAndMetadata
        },
        new AzureOpenAIEmbeddingSkill { ... },
        new TextSplitSkill
        {
            Name = "text-split",
            TextSplitMode = TextSplitMode.Sentences,
            MaximumPageLength = 500
        }
    }
};

Check our guide on document parsing.

RAG Architecture on Azure

Recommended Architecture

Azure Blob Storage / SharePoint / SQL
              ↓
    [Document Intelligence]
              ↓
    [Azure AI Search Indexer]
              ↓
    Azure AI Search (Vector + Semantic)
              ↓
    [Azure OpenAI Service]
              ↓
    Azure App Service / Functions

Complete Configuration

DEVELOPERcsharp
// 1. Create the index
var indexClient = new SearchIndexClient(endpoint, credential);
await indexClient.CreateOrUpdateIndexAsync(new SearchIndex("rag-index")
{
    Fields = {
        new SearchableField("id") { IsKey = true },
        new SearchableField("content"),
        new SearchField("vector", SearchFieldDataType.Collection(SearchFieldDataType.Single))
        {
            VectorSearchDimensions = 3072,
            VectorSearchProfileName = "vector-profile"
        },
        new SearchableField("metadata")
    },
    VectorSearch = new VectorSearch { ... },
    SemanticSearch = new SemanticSearch { ... }
});

// 2. Configure the indexer
var indexerClient = new SearchIndexerClient(endpoint, credential);
await indexerClient.CreateOrUpdateDataSourceConnectionAsync(...);
await indexerClient.CreateOrUpdateSkillsetAsync(skillset);
await indexerClient.CreateOrUpdateIndexerAsync(indexer);

// 3. RAG query
var searchClient = new SearchClient(endpoint, "rag-index", credential);
var results = await searchClient.SearchAsync<Document>(
    "How to configure the product?",
    new SearchOptions
    {
        VectorSearch = new VectorSearchOptions
        {
            Queries = { new VectorizedQuery(queryEmbedding) { KNearestNeighborsCount = 5 } }
        },
        SemanticSearch = new SemanticSearchOptions { ... }
    }
);

Performance and Limits

Benchmarks

Limits

Pricing

Updated Pricing

Additional Costs

• Semantic ranking: $10 per 1M queries
• Vector search: Included
• Azure OpenAI: Separate billing

Check our guide on RAG cost optimization.

Comparison

Azure AI Search vs Alternatives

Our Take

Azure AI Search becomes a solid option for RAG:

Strengths:

• Native Azure integration
• Performant hybrid search
• Advanced semantic ranking
• Enterprise compliance

Points of attention:

• Microsoft lock-in
• Configuration complexity
• High cost for large volumes

For Azure-first enterprises, it's a natural choice. For others, cloud-agnostic alternatives merit consideration.

Platforms like Ailog offer an independent alternative with French hosting and simplified setup.

Check our guide to best RAG platforms.