Deep Dive into AI Tools in Copilot Studio - BRK315

Explore advanced AI tools and agent development capabilities in this 43-minute conference session from Microsoft Ignite 2025. Discover how to leverage industry-leading AI tools within Copilot Studio to create powerful agents that integrate seamlessly across the Microsoft 365 Copilot ecosystem and beyond. Learn about the three core experience types in Copilot Studio and understand the distinctions between retrieval, task, and autonomous agents. Master the implementation of Model Context Protocol (MCP) for enhanced agent functionality and explore the comprehensive suite of tools and connectors available for agent integration. Gain insights into human-in-the-loop workflows for information requests and understand strategic decision-making between using agents versus traditional workflows. Examine practical applications through AI prompts for invoice information extraction and witness a complete agent demonstration showcasing real-world scenarios with human oversight and AI-driven decision making. Presented by Microsoft experts Derah Onuorah and Wassim Wehbi, this intermediate-level breakout session provides hands-on knowledge for building sophisticated AI agents using low-code approaches in Copilot Studio.

Syllabus

0:00 - What is Copilot Studio – Three Experience Types
00:03:34 - Types of Agents – Retrieval, Task, and Autonomous Agents
00:08:22 - Using Model Context Protocol MCP with Copilot Studio
00:13:08 - Overview of tools and connectors integrated into the agent
00:20:11 - Introducing Human-in-the-Loop Request for Information in Agents
00:24:52 - Explaining When to Use Agents vs. Workflows
00:29:25 - AI Prompts for Extracting Invoice Information
00:32:33 - Introduction to Full Agent Demo in Copilot Studio
00:38:29 - Running the Agent Demo with Human-in-the-Loop and AI Decision Making