Azure AI Search

In the Ultra Transcenders books

AI-901AI-300AI-103

Each book explains Azure AI Search in context, with comparison tables and the common traps.

Terms in this definition

• OVER

  Gives a T-SQL window function its window: PARTITION BY, ORDER BY and, if wanted, a ROWS or RANGE frame. Rankings and running totals can then be worked out while every row is kept.

• AI enrichment

  Indexing-time process in Azure AI Search where skills such as OCR, key phrase extraction, entity recognition and image analysis turn raw content into fields that can be searched.

• Knowledge mining

  Making large amounts of content searchable and open to analysis by indexing it and enriching it with AI, usually through Azure AI Search.

• ALL

  A DAX function that ignores any filters and gives back every row of a table or every value of the named columns. Used within CALCULATE, it works as a modifier that clears filters, although REMOVEFILTERS states that intent more clearly where it is available.

Related terms

• Agentic retrieval

  Query approach in Azure AI Search where an LLM turns a conversation into several planned subqueries; these execute together, are reranked semantically, and the best chunks are merged. Classic RAG, by contrast, issues just one query.

• Autocomplete

  Query API in Azure AI Search that finishes partially typed terms using suggester fields, invoked with suggesterName. To get matching documents back, use Suggestions instead.

• Azure AI Search tool

  Tool that lets an agent query an Azure AI Search index that already exists. You supply project_connection_id and index_name; by default it runs vector_semantic_hybrid queries and returns five results (top_k 5).

• Azure OpenAI On Your Data

  Feature of Azure OpenAI, now deprecated, for grounding chat answers in your own content retrieved from a search index. Its replacement for new work is Foundry Agent Service paired with Azure AI Search or Foundry IQ.

• Chunking

  Dividing content into short passages before it is embedded and indexed. Oversized chunks mix unrelated sections and weaken meaning, so Azure AI Search advises around 512 tokens each, overlapping by about 10-25%.

• Custom Web API skill

  Azure AI Search skill (Microsoft.Skills.Custom.WebApiSkill) that sends enrichment calls to an HTTP endpoint you supply in uri, for example an Azure Function fronting a Document Intelligence model. Setting a context that ends in /* makes it run per item.

• Customer-managed keys (Azure AI Search)

  Adds a further encryption layer to Azure AI Search, protecting indexes and synonym maps with an RSA key you hold in Key Vault or Managed HSM, which the service reaches using its managed identity. Expect queries up to 30-60% slower and no added capacity.

• Document-level access control

  Azure AI Search preview capability that captures permissions from SharePoint or ADLS Gen2 at indexing time and filters results by the Entra token the caller sends with each query.