Building Declarative Agents to Chat with Your Documents in Copilot

Organizations are flooded with documents, making it hard to find precise information quickly. Building custom AI assistants that securely query your own files is the modern solution to this challenge. This text-only course guides you through creating declarative agents for Copilot, enabling you to connect AI directly to your custom data sources. You will transition from understanding basic AI extensibility to structuring agent manifests that ground conversations in your private files. What you'll learn: - Understand the foundational architecture of declarative agents and how they extend Copilot capabilities - Configure agent manifests with precise metadata, custom instructions, and conversation starters - Implement grounding techniques to ensure the agent answers questions using only your specified documents - Apply basic prompt engineering principles to guide the agent's tone, scope, and response format - Explore security and access control best practices to protect sensitive organizational data - Test and refine your declarative agent's performance using realistic document-query scenarios You will start with core definitions and architectural concepts before moving step-by-step through manifest configuration, grounding strategies, and deployment basics. Each concept is reinforced with clear written explanations and structured code snippets. This course is designed for developers, IT professionals, and tech-savvy creators who are new to Copilot extensibility, with no prior AI development experience required. Start building your own custom document-chat assistants today.