Domain 3: Implement AI and analytics capabilities in database solutions

Domain 3 is where Azure Cosmos DB meets AI. It covers full text search, vector data and vector indexes, sharded DiskANN, vector and hybrid search, and Retrieval-Augmented Generation (RAG); durable agent memory stores, conversation state, and semantic memory; and operational analytics with Microsoft Fabric, from choosing between Azure Cosmos DB and Cosmos DB in Fabric to mirroring, T-SQL and Spark queries, and the Spark connector.

20–25%Of the exam

3Objectives

28Practice questions

5Official exercises

Skills measured

Objectives and skills below are quoted from the official study guide. The count beside each skill is how many exam simulator questions test it.

3.1 Design and implement AI retrieval

Design and implement full-text search1
Design for vector data, including vector data type, vector indexes, and size3
Design and implement vector partitioning and sharded DiskANN implementations1
Implement vector search1
Design and implement hybrid search4
Optimize full-text and vector performance1
Design and implement Retrieval-Augmented Generation (RAG)2

3.2 Design and implement agent memory stores

Design durable memory stores2
Implement conversation state3
Implement semantic memory2

3.3 Design and implement operational analytics on Azure Cosmos DB

Choose between Azure Cosmos DB and Cosmos DB in Microsoft Fabric1
Configure and monitor data mirroring into Microsoft Fabric4
Query Azure Cosmos DB mirrored JSON data in Microsoft Fabric with T-SQL and Spark2
Read and write data into Azure Cosmos DB from Microsoft Fabric with Cosmos DB Spark connector1

Official training and documentation

Microsoft Learn modules and product documentation for each objective, chosen for the current skills measured. Module durations are Microsoft's estimates.

3.1 Design and implement AI retrieval

Microsoft Learn modules