Document Processors Managed Agent in Copilot Studio: A Complete Step-by-Step Guide

The Document Processors Managed Agent in Copilot Studio marks a major leap forward in intelligent document automation within the Power Platform ecosystem. Designed to handle structured and unstructured data extraction, this managed agent integrates seamlessly with your Copilot environment to interpret, process, and deliver insights from business documents — all without requiring complex custom code.

Whether you’re dealing with invoices, contracts, resumes, or forms, this feature allows organizations to connect AI-powered document understanding directly into their Copilot experiences. In this article, we’ll explore what the Document Processors Managed Agent is, how it works, and how you can install, configure, and use it efficiently.

Note: For a broader understanding of all Managed Agents in Copilot Studio, you can read our first article: Managed Agents in Microsoft Copilot Studio.

What is the Document Processors Managed Agent?

The Document Processors Managed Agent is a Microsoft-managed runtime service that connects Copilot Studio with document processing capabilities powered by Azure AI and Power Platform services. It simplifies the process of extracting structured information from PDFs, images, Word documents, and other file types by leveraging AI-based document understanding models.

Key Highlights:

• No code or minimal code setup: Ideal for business users and citizen developers.
• Azure-hosted and managed: Reduces the need for manual agent maintenance or infrastructure provisioning.
• Integrated with Copilot Studio: Enables custom Copilots to interpret and process documents dynamically within a conversation.
• Supports AI models: Works with prebuilt or custom AI document models.

This integration bridges the gap between document intelligence and conversational AI, allowing users to interact with data-rich documents directly inside Copilot environments.

How Document Processors Managed Agent Works

Understanding how this managed agent works helps developers design smarter Copilots.

Here’s a simplified view of the document processing workflow:

1. Document Submission:
   A user uploads or references a document (PDF, DOCX, image, etc.) through Copilot.
2. Agent Activation:
   The managed agent in Copilot Studio triggers and routes the document to a preconfigured AI document processing model.
3. Document Understanding:
   The document is analyzed using machine learning models hosted on Azure AI Document Intelligence or Power Platform Document Automation services.
4. Data Extraction:
   Structured outputs such as invoice totals, names, dates, or key-value pairs are extracted.
5. Copilot Response Generation:
   The extracted data is returned to Copilot Studio, where it can be displayed to the user, stored in Dataverse, or used in Power Automate workflows for further actions.

This entire cycle happens within seconds, providing a conversational and interactive document processing experience inside your Copilot.

Architecture Overview

The Document Processors Managed Agent architecture is designed for scalability and integration flexibility. Below are the main components:

• Copilot Studio: The platform where you build and configure your AI-powered Copilots.
• Managed Agent Runtime: A Microsoft-managed connector layer that facilitates secure communication between Copilot Studio and external AI services.
• Power Platform Environment: Hosts connectors, flows, and AI models used for document processing.
• Azure AI Document Intelligence (Form Recognizer): The underlying service responsible for extracting structured information from files.
• Dataverse or SharePoint: Used to store and manage processed data outputs.

This architecture ensures that sensitive information remains secure, using Microsoft Entra ID authentication and role-based access control (RBAC).

Installing the Managed Agent in Copilot Studio

Setting up the Document Processors Managed Agent is a straightforward process within Copilot Studio. Here’s a step-by-step guide:

Step 1: Open Copilot Studio

Sign in to Microsoft Copilot Studio using your organizational account. Navigate to the Manage → Agents section.

Step 2: Add a Managed Agent

Click Install Agent and select Document Processors Managed Agent from the available list. This option is visible under the “Managed Agents (Preview)” section.

Step 3: Configure Connection Settings

Provide configuration details:

• Environment: Choose your Power Platform environment.

• Authentication: Use Microsoft Entra ID for secure connection.

• Permissions: Ensure your Copilot has the required roles (Environment Maker or Admin).

Step 4: Connect AI Model

Select or link an AI model built with Power Automate AI Builder or Azure AI Document Intelligence. You can use either a prebuilt model (e.g., Invoice, Receipt, Identity Document) or a custom-trained model.

Step 5: Validate and Test

Once configured, validate the connection. Upload a sample document to verify that the agent can process and return structured output.

Once successful, the agent becomes available as a callable action from your Copilot topics.

Install Document Processors Managed Agent in Copilot Studio: Screenshots

Below are the step-by-step screenshots of how to install the document processor agent in Copilot Studio.

Install Document Processors agent

Log in to your Copilot Studio portal.

On the home page you will see the “Install a managed agent” section; from there, click on the “Document processors (Preview)” agent, and then click on the install button. As I have already installed it, you can see it is showing as “Installed.”

This process will go through the installation wizard.

Upload sample documents

First you need to upload a sample structured document that you want to process. For example, the invoice document.

Configure Document Validation Rules

Here you can configure document validation rules. For example, currency in the document should follow the USA currency format, date, time, etc. You can configure these types of rules. This configuration is optional. However, it is recommended to configure it for your production data processing.

 

Configure Document Reviewer

In this section you need to configure your document reviewer email ID and approval message.

Configure Document Source Trigger

In this screen, you need to configure your document source triggers. As of today there are three triggers available:

• When a new e-mail arrives in the shared mailbox: You can configure your shared mailbox.
• When a new email arrives: You can configure individual email IDs.
• When an item is created: As of today this is disabled.

This configuration explains that, based on your configuration, if any document comes to the shared mailbox or individual email, your invoice document will be processed and the extracted document attributes will be stored in the Dataverse table.

Finally, click on the Save and close button. Your Document Processors managed agent installation is completed.

Verify the installation

Check your Power Platform solutions; there you will see two new solution files have been created with this installation:

• Document Processors (Preview): Empty solution file
• Document Processors Core: With all agent related objects

Document processors solution files

Open your Document Processor’s Core solution file, and you will see all your agent objects like Agent components, Agents, AI models, Apps, Cards, Cloud flows, Connection references, Custom connectors, Data workspaces, Environment variables, and Tables, etc.

Integrating Document Processor Agent with Copilot Topics

After the installation, you can call the agent directly within a Copilot topic:

1. Go to Authoring → Topics.

2. Choose an existing topic or create a new one.

3. In the authoring canvas, add a new action node.

4. From the list of actions, select the Document Processors Managed Agent.

5. Pass parameters such as:

   • Document URL or file ID

   • Model ID

   • Extraction type

Once the topic is published, users interacting with the Copilot can upload or reference a document and instantly receive analyzed data in a conversational format.

Real-World Use Cases

Here are some practical applications of the Document Processors Managed Agent:

1. Invoice Processing Automation:
   Extract invoice numbers, amounts, and vendor names directly within a Copilot chat.

2. Resume Parsing for HR Automation:
   Integrate with Power Automate to screen resumes and store parsed results in Dataverse.

3. Contract Review & Summary:
   Allow Copilot to summarize contract terms or identify key clauses for legal teams.

4. Purchase Order Validation:
   Automatically extract order numbers and totals for procurement workflows.

5. Customer Document Verification:
   Validate identity documents in onboarding scenarios using AI document recognition.

Each of these use cases demonstrates how Copilot and document AI together can reduce manual workload, enhance accuracy, and improve productivity.

Best Practices for Implementation

To get the most out of the Document Processors Managed Agent, follow these best practices:

• Use Prebuilt Models When Possible: For invoices, receipts, and identity cards — they’re faster to deploy and optimized by Microsoft.
• Train Custom Models for Domain-Specific Documents: When working with custom formats or internal forms.
• Combine with Power Automate: Automate follow-up actions such as email notifications, record creation, or approvals.
• Monitor Using Analytics: Track usage and performance in Copilot Studio’s analytics dashboard.
• Secure Access: Use Entra ID and environment-level security to control access to document data.

These practices ensure performance efficiency, data security, and maintainability of your solution.

Benefits of Using Managed Agents in Copilot Studio

The introduction of managed agents — especially the Document Processors Agent — is part of Microsoft’s broader vision for AI-powered automation in enterprise environments.

1. Simplified AI Integration

Developers can use complex AI features without writing backend integration code.

2. Unified Experience

Integrates Power Platform, Azure AI, and Microsoft 365 under one Copilot interface.

3. Scalable and Secure

Microsoft manages the runtime, ensuring automatic updates and enterprise-grade compliance.

4. Extensible

You can extend the agent with Power Automate or Dataverse connectors.

5. Conversational Data Intelligence

Turns document insights into real-time conversational interactions, driving better decision-making.

YouTube Video Demo: Document processors managed

Watch the detailed walk-through with the live demo right from installing the document processor’s agent to configuring and testing real-time use case data.

Conclusion

The Document Processors Managed Agent in Microsoft Copilot Studio is a game-changing feature that empowers organizations to bring AI-driven document understanding directly into their conversational Copilots. With minimal setup and no-code integration, users can automate data extraction, streamline business processes, and enable smarter document-based interactions.

As Microsoft continues expanding Copilot Studio’s ecosystem, managed agents like this one will redefine how enterprises integrate AI, automation, and data intelligence into their day-to-day workflows.

If you haven’t tried it yet — now is the perfect time to experiment and see how document automation meets conversational AI inside Microsoft Copilot Studio.

References

• Document Processors Managed Agent – Power Platform Architecture
• Install Managed Agent – Microsoft Copilot Studio

See Also

• Explore Our Complete Microsoft Copilot Studio Articles Library

 

About Post Author