Agentic Coding with Azure Cosmos DB: From Idea to Working App

Part of the Modern AI Apps with Azure Cosmos DB Series. What happens when you move beyond autocomplete and start building applications with an AI coding agent as part of your workflow? This session shows how developers can go from a prompt to a working Azure Cosmos DB application using agentic development tools, while learning where agents accelerate work—and where human architecture decisions still matter. Presenters: Sajeetharan Sinnathurai - Azure Cosmos DB Principal Program Manager -   / sajeetharan   Jay Gordon - Azure Cosmos DB Senior Program Manager   / jaygordon0042   • Start with the Agent Kit — Give your coding agent the context it needs to build Cosmos DB apps the right way - https://aka.ms/CosmosDB-agentkit • Subscribe to Azure Cosmos DB on YouTube -    / azurecosmosdb   • Follow Azure Cosmos DB on X -   / azurecosmosdb   • Follow Azure Cosmos DB on LinkedIn -   / azure-cosmos-db   Chapters: 0:00 - Introduction 2:24 - Series Overview: Modern AI Apps with Azure Cosmos DB 3:51 - Tooling Overview: VS Code, Copilot, SDKs, and Emulator 5:08 - Deep Dive: Azure Cosmos DB Agent Kit 6:46 - Why Use the Agent Kit? 8:11 - Installation and Setup 11:28 - Agent Kit Directory and Rule Structure 14:47 - Q&A: Data Explorer vs. VS Code Extension 18:04 - Starting the Linux Emulator 20:53 - Defining the Project: Movie Review Website 22:56 - Planning with AI (Plan Mode) 27:58 - Transition to Agent Mode 31:10 - Scaffolding and Building the Application 34:57 - Q&A: Optimizing Performance and Reducing Costs 36:56 - Running the App and End-to-End Testing 40:19 - Q&A: Understanding Skill Content and Context 41:46 - Generating Documentation with AI 45:03 - Connecting and Exploring Data with VS Code Extension 49:33 - Data Model Modifications and Iteration 55:54 - Next Episode Preview: Designing Better Data Models 59:08 - Q&A: Rule Sets and Final Thoughts #azurecosmosdb #aicoding [eventID:27239]