Introduction

  • 5 minutes

In this self-paced module, you use Microsoft Copilot Studio to build an autonomous agent that automates complex business tasks and improves efficiency and innovation across your organization.

Copilot Studio is a graphical, low-code tool for creating an autonomous agent and extending the agent across your enterprise. Copilot Studio provides AI-based conversations, builds automation with Microsoft Power Automate, and helps expand the autonomous agent capabilities across the Microsoft 365 ecosystem in your enterprise. These agents act as expert systems that support individuals, teams, or organizations.

To connect with enterprise data, you can select from many built-in connectors in the Copilot Studio environment.

In this module, you learn about the fundamental concept of the Copilot Studio environment, explore real-world applications, and engage in exercises that help you build and deploy an autonomous agent effectively.

Screenshot of Copilot Studio.

Overview: Autonomous agent

An autonomous agent in Copilot Studio is an AI-powered solution that manages, orchestrates, and automates complex business tasks. An autonomous agent proactively supports workflows, minimizes manual effort, and adapts to changing conditions or requirements. By streamlining tasks, such as data processing, decision support, or workflow automation, these agents help organizations achieve greater efficiency and innovation, therefore complementing human expertise while reducing the need for constant intervention.

How an autonomous agent works

An autonomous agent works independently based on autonomous prebuilt or custom triggers that are available in Copilot Studio. The agent automates long-running processes, follows human guardrails, and asks for help. Also, an advanced agent can orchestrate other agents. Each agent can dynamically decide the best action to take next. As an agent builder, you can provide instructions, knowledge, and actions that act as guardrails.

An autonomous agent has six interconnected stages that follow a circular flow:

  1. Trigger activation - Initiates tasks based on predefined conditions.
  2. Process automation - Runs tasks without human intervention.
  3. Follow guardrails - Adheres to guidelines to ensure safe operation.
  4. Seek assistance - Requests human input when needed.
  5. Orchestrate other agents - Coordinates with other agents for complex tasks.
  6. Integrate AI reasoning - Uses AI to enhance decision-making capabilities.

Screenshot of the six interconnected stages showing how the autonomous agent works.

Next steps

This unit provided an overview of what an autonomous agent is and how it relates to Copilot Studio. Next, you explore how to create an autonomous agent for a real-world scenario.