Summary
Agents provide self-service support through natural-language interactions. They can resolve common problems and transfer a conversation with its context to a customer service representative when human assistance is required.
Agents can incorporate information from other systems and use tools to schedule meetings, assign cases, or send emails. Microsoft Copilot Studio provides a graphical, low-code environment for subject matter experts, makers, and developers to create and manage these agents.
This module examined how to get started with Microsoft Copilot Studio, create and deploy agents, and deploy them for consumption across multiple channels, including:
- Introducing agents, explaining where they're being used, and providing an overview of how Microsoft Copilot Studio can be used to create agents.
- Comparing the GitHub Copilot, standard, and Copilot chat harnesses.
- Comparing the standard harness experience with the new GitHub Copilot harness experience.
- Examining how to create and work with agents in different environments to tailor agent content based on factors such as teams and regions.
- Exploring how generative AI can be used in an agent.
- Reviewing the process for creating an agent and how to work with the Microsoft Copilot Studio user interface.
- Exploring how generative AI can be used to provide answers to users' questions.
- Examining the tools that are available to assist you with testing your agent.
- Reviewing the process for publishing an agent so that it can be made available across multiple channels and describing the analytics that are available when the agent is published.
Your next step is to gain a deeper understanding of how to design effective agent behavior. This learning includes:
- Gaining a deeper understanding of conversation nodes
- Examining how variables can be used to capture and store relevant data
- Grounding agents with knowledge sources
- Extending agent capabilities with tools
- Incorporating Copilot Studio workflows to enhance agent behavior