Microsoft Copilot Studio guidance documentation

Microsoft Copilot Studio guidance documentation provides best practices, implementation, and architecture guidance information from the team that works with our enterprise customers.

Building agents

Overview

• Topics best practices
• Trigger phrases best practices
• Slot filling best practices

How-To Guide

• Pass context variables from your webpage to an agent
• Pass web an analytics tracking ID from a webpage
• Optimize agents to minimize latency

Generative AI

Concept

• Generative AI strategies for math and data queries
• Generative answers based on public websites

How-To Guide

• Instruction guidance for generative AI

Conversational user experience

Overview

• Introduction to conversational experiences

How-To Guide

• Design the ideal experience
• Practical tips for designing a conversational experience

Governance and security

Concept

• Best practices: Take a phased approach

Planning a conversational AI project

Concept

• Projects best practices
• Getting started
• Building a team
• Project planning
• Production environment strategy

Testing conversational agents

Concept

• Plan and create a performance test
• Test conversational agents using Direct Line
• Performance test reference sample and guideline
• Best practices for improving agent performance

Power CAT Copilot Studio Kit

Concept

• Overview

How-To Guide

• Install the Kit
• Configure users and teams
• Configure agents
• Configure tests
• Run tests

Measuring and improving agents

Concept

• Measuring engagement
• Measuring outcomes
• Deflection optimization
• Custom analytics strategy
• Optimize prompts

Related guidance content

Concept

• Power Platform adoption
• Power Platform real-world case studies
• Power Apps
• Power Automate
• Power Pages

Architecture

• Power Platform Well-Architected: Design principles of intelligent application workloads
• Power Platform and Copilot Studio Architecture Center: Reference architectures and solution ideas