JARED MATFESS is an AI Architect at AvePoint who partners with a diverse portfolio of clients, including Fortune 500 companies and State & Local Government agencies. He is a 7-time Microsoft MVP and has more than 20 years' experience working in the Microsoft ecosystem and is an expert at assisting organizations in their digital transformations by leveraging advanced technologies, including AI.

Foreword xxi

Chapter 1 Navigating the Copilot Ecosystem 1

What Is GenAI? 2

How Does GenAI Work? 3

GenAI Key Terms and Definitions 3

The Risk of Bias in GenAI 4

OpenAI Brings GenAI to the World 4

ChatGPT Gains Excitement 5

ChatGPT Data Leaks 5

Microsoft's Strategic Investment in OpenAI 6

Retrieval-Augmented Generation 7

Azure OpenAI RAG Pattern 8

Enterprise Adoption of the RAG Pattern 9

Microsoft M365 Copilot 10

The Rise of the Copilots 12

Copilot, aka Bing Chat Enterprise 12

Microsoft's Copilot Portfolio 13

Copilot for Sales 14

Copilot for Service 16

Copilot for Security 17

Copilot in Microsoft Viva 19

Viva Goals 19

Viva Engage 20

Viva Amplify 21

Viva Roadmap 22

Additional Copilots 22

Aligning Copilots with Company Personas 24

Introducing Copilot Studio 25

Conclusion 26

Chapter 2 Introduction to Copilot Studio 29

Copilot Studio's Core Audience 29

Citizen Developers Overview 30

Citizen Developer Challenges 31

The Role of IT 32

Copilot Studio: The Platform 33

Copilot Studio Prerequisites 33

Accessing Copilot Studio 34

Power Platform Environments 35

Microsoft Dataverse 36

Dataverse Core Components 37

Environment Management 39

Creating Your First Agent 39

Adding Knowledge Sources 42

Testing Your Copilot Agent 47

Copilot Studio Topics 50

Conversational Boosting with GenAI 52

Modifying Your User Experience with Topics 55

Lights, Camera, Actions! 57

Conclusion 69

Chapter 3 Publishing Your Copilot Agent 71

Channels 71

Publishing Your Agent to Teams + Microsoft 365 74

Publishing Your Agent 79

Testing Your Agent in Microsoft Teams and M365 Copilot 86

Conclusion 90

Chapter 4 Microsoft 365 Copilot Declarative Agents 91

The Spectrum of Copilot Agents 91

M365 Copilot Agents 92

Declarative Copilot Agents 93

The Agent App Package 93

App Manifest 93

App Icons 94

Declarative Agents Manifest 94

Plugin Manifest 95

Configuration Options 96

The AI Orchestrator 96

Creating a Declarative M365 Copilot Agent with the Copilot Studio Agent Builder 97

Configuring Your Agent 98

Setting Instructions 99

Configuring Knowledge 101

Actions and Capabilities 103

Starter Prompts 104

Creating Your Agent 105

Adjusting Sharing Permissions 106

Application Manifest: [...] 108

The [...] File 110

Code Versus Configuration 112

Test Driving Your Agent 112

Updating Your Declarative Agent 114

Copilot Studio Agent Builder Limitations 115

Data Storage 115

Application Lifecycle Management 116

User Experience 116

When to Use Declarative Agents 117

Conclusion 118

Chapter 5 Planning ALM for Your Copilot Agents 119

The ALM Framework 120

Requirements Gathering 120

Design 121

Development 121

Testing 122

Deployment 122

Maintenance and Retirement 123

ALM Summary 123

Power Platform Environments 123

Environment Strategy 124

Environment Costs 124

Managed Environments 125

Managed Environments Considerations 127

Solutions 127

Creating a Solution 129

Exporting a Solution 132

Importing a Solution 135

CI/CD Pipelines 140

Conclusion 141

Chapter 6 Deep Dive into Agent Templates 143

Prerequisites 144

Selecting Our Agent Template 145

Citizen Services Template 146

Knowledge Sources 146

Topics 147

Apply for a Service Topic 148

Data Collection Topic and Adaptive Cards 149

Road Closures Topic 154

Conversational Boosting Topic 155

Summarizing Topics 157

Building Our Agent from the Citizen Services Template 157

Updating Knowledge Sources 158

Updating Topics 159

Testing Your Agent 166

Conclusion 169

Chapter 7 Real-World Use Cases and Inspiration 171

Agents in the Contact Center 171

The Scenario 172

Technical Setup 173

Testing the Contact Center Agent 175

Contact Center Charlie Agent Summary 176

Agents in the Public Sector 176

Technical Setup 176

Testing Our Shared Mailbox Agent 182

Summary of the Shared Mailbox Agent 185

Agents in Human Resources 186

Technical Setup 186

Testing Your Heidi from HR Agent 194

Summary of the Updated Heidi from HR Agent 195

Conclusion 195

Chapter 8 Building an Autonomous Agent 197

Autonomous vs. Semi-autonomous Agents 197

Considerations for Autonomous Agents 198

How Microsoft's Autonomous Agents Work 198

Autonomous Agent Use Case 199

Microsoft Form Configuration 200

Salesforce Sales Cloud Setup 201

Configuring Your Agent 202

Enabling Orchestration 203

Creating a Trigger 204

Configuring your Trigger 207

Setting Up Knowledge Sources 213

Creating Actions 218

Creating Inputs 220

Formatting Variables 221

Getting the Account ID 222

Creating a Condition 224

Creating a Contact 226

Creating an Opportunity 227

Associating Your Opportunity to a Contact 228

Configuring "Run as" User 231

Configuring the Create Lead Action 232

Configuring Your Agent Instructions 235

Testing Your Agent 236

Publishing and Monitoring 242

Conclusion 242

Chapter 9 Optimizing and Measuring Your Agent 245

Prerequisites 245

Agent Analytics 246

Optimizing Your Agent for Cost 248

Enhancing User-Focused Agent Performance 251

Quick Replies 251

Starter Prompts 253

Capturing User Feedback 255

Conclusion 265

Chapter 10 Copilot Studio and Azure AI Foundry: Better Together 267

Azure AI Search 268

Optimizing for Cost 269

Creating Your Azure SQL Database 269

Creating a Table and Loading Data 274

Provisioning Azure AI Search 276

Configuring Your Agent for Azure AI Search 286

Testing Your Agent 288

Conclusion 289

Appendix Agent Flows 291

Index 297