AB-620: Copilot Studio AI Agent Builder Exam Prep

Name: AB-620: Copilot Studio AI Agent Builder Exam Prep
Rating: 4.4 (277 reviews)

Prepare for the AB-620 Exam with Instructor led Hands-On Labs and Videos

Created byKuljot Singh Bakshi | 70,000+ Enrollments Worldwide

Last updated 9/2026

English

English [Auto],Finnish [Auto],

What you'll learn

Build enterprise-grade AI agents using Microsoft Copilot Studio and Microsoft Foundry
Integrate APIs, MCP servers, connectors, Power Automate and enterprise knowledge sources into AI agents
Design multi-agent and RAG-based AI solutions for real-world business scenarios
Monitor, test, deploy and manage AI agents using Application Insights and Power Platform ALM strategies

Course content

12 sections • 64 lectures • 9h 27m total length

Course Introduction2:29

What to expect from this section of the course1:37
Introduction to Microsoft Copilot Studio13:47
Copilot Studio Licensing Guide3:01
Lab: Copilot Studio Free Trial Sign Up (Hands-On Lab)1:33
Lab: Copilot Studio Walkthrough (Hands-On Lab)4:41
Copilot Studio Updates - A Quick Note1:40
Generative AI Orchestrator/Planner for Microsoft Copilot Studio4:30
Explore how the generative AI orchestration planner in Microsoft Copilot Studio enables agentic planning and dynamic tool execution at runtime, using default low-code orchestration and plugin descriptions.
Lab: Creating Web Grounding Standalone Agent (Hands-On Lab)18:34
Lab: Creating a SharePoint Site Embedded Agent (Hands-On Lab)9:10
Understanding Microsoft "Work IQ"6:02
Explore how Microsoft work iq acts as the intelligence layer that personalizes Microsoft 365 Copilot, using data, context, and skills and tools to tailor enterprise knowledge and secure role-based access.
Embedded v/s Standalone Copilot Studio Agents4:20
Compare embedded and standalone Copilot Studio agents. Use embedded agents in the same Copilot chat and invoke them with the add symbol, while standalone agents open a separate tab.

What to expect from this section of the course1:25
Lab: System-Built Topics and Custom Topics (Hands-On Lab)6:01
Explore system-defined and custom topics in Copilot Studio, including triggers and boolean-based routing, and compare topic-based control with the generative AI orchestrator for agent execution.
Lab: Create a Custom Topic with Variable Management (Hands-On Lab)12:52
Build a custom topic in Copilot Studio using variable management and conditional branching to guide users through Microsoft certification and custom AI Udemy courses, with escalation when unsupported.
Lab: Designing a Message-Based Adaptive Card (Hands-On Lab)17:08

What to expect from this section of the course1:39
Lab: Adding MSN Weather Connector as an Action to Agent (Hands-On Lab)9:56
Prompt Engineering: Introduction and Best-Practises5:51
Master prompt engineering to craft precise, context-rich prompts that maximize Copilot and other generative AI outputs. Use the four-part framework: goal, context, expectations, and source to drive focused, effective results.
Lab: Working with Prompt Actions in Copilot Studio (Hands-On Lab)14:16
Lab: Running a Power Automate Flow with Copilot Studio (Hands-On Lab)15:59
Build autonomous agent flows in Copilot Studio by integrating a resume screener prompt action with a Power Automate flow, then trigger and email structured results.
Lab: Sentiment Analysis Flow using AI Builder (Hands-On Lab)12:02
Introduction to MCP Servers5:39
Lab: MCP Server Agent in Copilot Studio (Hands-On Lab)11:21
Lab: Custom Fallback Topics (Hands-On Lab)5:33
Lab: Copilot Studio "Analytics" Tab for Monitoring and Usage (Hands-On Lab)5:45
Lab: Copilot Studio Agent "Settings" Tab (Hands-On Lab)6:02
Explore the settings tab in Copilot Studio to configure generative ai, security, channels, and connections. Learn how entities, topics, language understanding, and moderation shape agent behavior and publishing across channels.

What to expect from this section of the course1:48
Lab: Multi-Agent Workflow with Another Copilot Studio Agent (Hands-On Lab)10:43
Introduction to A2A10:32
Lab: A2A in Copilot Studio (Hands-On Lab)4:39
Introduction to the Microsoft Foundry Ecosystem8:45
Lab: Creating a Microsoft Foundry Resource (Hands-On Lab)9:45
Lab: Deploying our First Foundry LLM (Hands-On Lab)12:18
Lab: Creating our First Agent using the Agent Service (Hands-On Lab)19:59
Lab: Multi-Agent System with Copilot Studio and Foundry (Hands-On Lab)14:18
Explore a multi-agent setup with Copilot Studio and Microsoft Foundry, where a parent agent queries the MS Learn MCP server to fetch documentation pages and code samples.

What to expect from this section of the course2:34
Explore Copilot Studio integration with the Power Platform and Dataverse, outline learning paths for CRM and ERP, and complete capstone building custom Dataverse tables, deploying a managed solution in production.
Introduction to the Microsoft Power Platform12:27
Introduction to Microsoft Dataverse14:18
Explore how Microsoft Dataverse serves as the single source of truth for Power Platform apps, providing secure cloud storage with standard and custom tables, row- and column-level security, and relationships.
Lab: Walkthrough of Power Platform (Hands-On Lab)6:11
What we will be creating as a Capstone Project3:35
Lab: Creating our Dataverse Table (Hands-On Lab)9:46
Lab: Populating Dataverse Table with Agent (Hands-On Lab)17:18
Create a Copilot Studio agent to run an AI Builder sentiment analysis flow that populates a Dataverse review table with review text, sentiment scores, and metadata.
Environments, Solutions and CI/CD in Microsoft Power Platform21:19
Lab: Creating a Managed Copilot Studio Solution (Hands-On Lab)12:52

What to expect from this section of the course2:47
Explore retrieval augmented generation with enterprise data to ground chatbots. Learn retrieval, augmentation, and generation steps, vector embeddings for similarity search, and multimodal RAC with Azure AI Search.
What is RAG (Retrieval Augmented Generation) ?16:28
Introduction to Multi-Modal RAG with Azure AI Search (Hands-On Lab)15:10
Lab: Deploying the RAG Infra on Azure (Hands-On Lab)9:28
Lab: Creating the AI Search Index (Hands-On Lab)14:32
Lab: AI Search RAG with Copilot Studio (Hands-On Lab)14:42
Connect a multi-modal rag index from Azure AI Search to a Copilot Studio agent to answer Maggie's travel queries using multimodal documents.

What to expect from this section of the course1:28
Understanding Agent Evaluation with Copilot Studio7:24
Lab: Setting up Agent Evaluation (Hands-On Lab)8:01
Set up and evaluate an AI agent in Microsoft Copilot Studio using Maggi's Travel PDFs as the knowledge base and a CSV-based test set.
Lab: Viewing Agent Eval Results (Hands-On Lab)1:28
Lab: Creating a Trigger-Based Agent Eval Flow (Hands-On Lab)3:59

Requirements

No programming experience required. You will learn everything you need to know.

Description

Welcome to the Microsoft Certified AB-620: AI Agent Builder Associate Exam Prep Course — your complete guide to designing, building, integrating, and managing enterprise-grade AI agents using Microsoft Copilot Studio and the broader Microsoft AI ecosystem.

The world of AI is rapidly evolving from simple chatbots to intelligent, autonomous, agentic AI systems capable of reasoning, orchestration, automation, and enterprise integration. Microsoft’s AB-620 certification is designed specifically for professionals who want to build these next-generation AI agent solutions using technologies such as Copilot Studio, Microsoft Foundry, MCP, APIs, connectors, RAG pipelines, multi-agent architectures, and Power Platform services.

In this course, you will learn how to design AI agent architectures, integrate enterprise systems, connect APIs and custom connectors, use MCP servers, configure agent topics and flows, implement adaptive cards, and create advanced agent workflows. We will also explore multi-agent collaboration using A2A concepts, enterprise knowledge retrieval using RAG, and integrations with Microsoft Foundry and Azure services.

This course is heavily hands-on and aligned with the official Microsoft Learn AB-620 study guide. Along the way, you will work with real-world demonstrations and practical implementation scenarios to strengthen both your conceptual understanding and technical skills.

We will also cover testing, monitoring, Application Insights, environment variables, solutions, Power Platform pipelines, and ALM strategies required for deploying and managing enterprise AI agents in production environments.

Whether you are preparing for the AB-620 certification exam, exploring Copilot Studio, or looking to build modern enterprise AI agent solutions using Microsoft technologies, this course will help you develop the practical skills and confidence needed to succeed.

Who this course is for:

Developers and IT professionals preparing for the Microsoft AB-620 certification exam
Copilot Studio users who want to build enterprise-grade AI agents and multi-agent solutions
Power Platform developers interested in integrating APIs, MCP servers, RAG and Microsoft Foundry into AI agents
AI engineers, solution architects and consultants exploring enterprise agentic AI solutions using Microsoft technologies

AB-620: Copilot Studio AI Agent Builder Exam Prep

What you'll learn

Explore related topics

Course content

Introduction1 lecture • 2min

Course Outline and Learning Recommendations1 lecture • 4min

AI Foundations4 lectures • 32min

Microsoft Copilot Studio Foundations11 lectures • 1hr 9min

Author Topics in Copilot Studio Agents4 lectures • 37min

Working with Agents, Tools, Actions, Plugins and Connectors11 lectures • 1hr 34min

Multi-Agent Systems with Copilot Studio9 lectures • 1hr 33min

Work with Power Platform Environments, Dataverse and Connectors9 lectures • 1hr 40min

Copilot Studio with Azure AI Search - "RAG" at Scale6 lectures • 1hr 13min

Agent Evaluation5 lectures • 22min

Requirements

Description

Who this course is for: