Autonomous Agent in Copilot Studio: Build AI that works FOR YOU

In this session, Amelia Roberts from Pragmatic Works demonstrates how to build an autonomous agent in Copilot Studio that automates repetitive tasks without requiring manual intervention. The example shown takes responses from a Microsoft Form and automatically creates tasks in Microsoft Planner, notifies users, and ensures that key information flows seamlessly across tools. The goal is simple: let AI do the repetitive work so professionals can focus on higher-value tasks.

Why Autonomous Agents Matter

Many organizations deal with repetitive processes such as responding to form submissions, assigning tasks, or sending notifications. By using Copilot Studio’s autonomous agents, these workflows can be handled automatically. This eliminates the need to constantly monitor responses and ensures the right information reaches the right people instantly.

Scenario Overview

Amelia’s example centers on a Microsoft Form designed to collect training requests:

• Participants select which Microsoft tool they’d like to take a class on.
• They choose a preferred date for training.
• Responses are saved automatically in an Excel file on OneDrive.

The autonomous agent in Copilot Studio will detect when the form is completed, pull the data from Excel, and create a task in Microsoft Planner. Managers can then assign instructors and schedule classes accordingly. Notifications can also be sent to relevant team members to review the new request.

Building the Agent in Copilot Studio

Amelia walks through creating a new agent in Copilot Studio called Class Assignment. The process includes defining the purpose of the agent and enabling settings such as Generative AI orchestration and multi-agent connections. These settings allow the agent to operate dynamically and potentially collaborate with other agents in future workflows.

Step 1: Setting Up the Trigger

The trigger is configured to activate when a Microsoft Form is submitted. By customizing the trigger name (e.g., “When Class Picker Form is Completed”), Amelia ensures that multiple triggers can be managed within the same agent without confusion. Authentication with both Microsoft Forms and Copilot Studio may be required during setup.

Step 2: Configuring Tools (Actions)

Copilot Studio uses “tools” (formerly called actions) to execute tasks after a trigger. Amelia adds several tools to process the form data and interact with Microsoft Planner:

1. List Rows in Excel: Pulls responses from the linked Excel file.
2. Get a Row: Extracts specific data such as class choice and date preference.
3. List Plans for a Group: Identifies the correct Planner group to update.
4. Create a Task: Generates a Planner task using class name as the title and date preference as the start date.

By naming each tool specifically, Amelia ensures clarity and easy reference when writing instructions later.

Step 3: Writing Instructions

Once tools are configured, instructions tell the agent how to use them after the trigger fires. Amelia writes step-by-step instructions that:

• Run the Excel tools to gather the most recent submission data.
• Extract the class choice and date preference.
• Format the date properly (MM/DD/YYYY).
• Create a Planner task with the extracted details.

This structured approach ensures the agent behaves predictably and processes each new form submission correctly.

Testing the Workflow

Amelia demonstrates how to test the agent by submitting a new form. The Copilot Studio activity map confirms each step is executed—listing rows, retrieving data, listing plans, and creating the task. She then verifies in Microsoft Planner that the new task (e.g., “Fabric”) has been created with the expected details.

While the start date did not fully align in the first test, this highlighted the importance of verifying results in the connected applications and fine-tuning the workflow as needed.

Key Takeaways

• Autonomous agents in Copilot Studio can fully automate repetitive workflows.
• Using Microsoft Forms, Excel, and Planner together ensures seamless integration of responses into actionable tasks.
• Clear naming and structured instructions are essential for managing complex agents.
• Testing and validation help confirm the automation works correctly across all tools.

Don't forget to check out the Pragmatic Works' on-demand learning platform for more insightful content and training sessions on Copilot Studio and other Microsoft applications. Be sure to subscribe to the Pragmatic Works YouTube channel to stay up-to-date on the latest tips and tricks.