How to integrate the Udemy Business AI Connectors with your AI tool

This article explains how to configure the Udemy Business AI Connectors with your AI tool and how to troubleshoot common issues.

The Udemy Business AI Connectors via Model Context Protocol (MCP) connects your AI assistants and agents to your Udemy Business account so they can securely use your organization’s learning data in real time. With this connection, your assistants can query your catalog, pull course and path details, ground their recommendations with up-to-date information from your Udemy Business tenant, and create structured learning paths directly in your organization's Udemy Business account. For more information, review frequently asked questions about Udemy’s AI Connectors. 

Table of Contents

• Available tools and capabilities
• Using create_learning_path effectively
• Prerequisites
• How to connect to the Udemy Business AI Connectors
• (Alternative method) Create a static client in Udemy Business
• Configuration instructions for specific AI tools 
• How to Configure Udemy Business MCP Server with Microsoft Copilot
  • Prerequisites
  • Configuration step
  • Refreshing the connector to access new tools
• Troubleshooting
  • Issues with authentication or OAuth
  • Issues with the MCP client 
  • Issues with the Microsoft Copilot connection
  • Why don’t I see the Udemy tools listed in my AI tool after configuration?
• Sample system prompt

Available tools and capabilities 

The following tools are currently available on the Udemy Business AI Connectors. All tools require the AI Connectors add-on subscription. Most tools are read-only and retrieve information from your Udemy Business catalog. create_learning_path is the only tool that writes to your account — it creates a learning path in your Udemy Business account.

Using create_learning_path effectively

Unlike the search tools, create_learning_path works best when your AI agent is configured to chain multiple tools together in sequence. A well-configured agent will:

1. Search for relevant courses using search_courses (and optionally search_labs for technical topics) based on the skill, audience, and proficiency level specified.
2. Inspect course curricula using get_course_curriculum to verify content relevance before including a course in the path.
3. Create the path using create_learning_path once the right content has been identified.

Without this chaining behavior, the agent might attempt to create a path without verifying content first, which can result in lower-quality output. To enable this behavior, update your agent's system prompt to include explicit instructions about the multi-step process. See the sample system prompt below for a recommended example.

Important notes for create_learning_path:

  • Approval required: Because this tool writes to your Udemy Business account, most AI systems (including Claude and ChatGPT) will prompt the user for confirmation before executing it. Users should expect an approval step before the path is created — this is expected behavior, not an error.

  • Private by default: Paths created via the AI Connectors are set to private. To assign or make a path discoverable to learners, open it in your Udemy Business admin tools under Learning Paths.

  • New tool — refresh required: If you connected the Udemy Business AI Connectors before create_learning_path was available, you'll need to refresh or reconnect the connector in your AI tool for the new tool to appear. See Refreshing the connector to access new tools below.

Prerequisites 

Before setting up this integration, you must have:

• A Udemy Business Enterprise Plan and purchase of the AI Connectors add-on.
  • Note: If you don’t see the MCP Server setting in your admin tools, your organization likely doesn’t have access yet. Reach out to your Customer Success representative or Udemy Business Support about upgrading to access the MCP server.
• Independent access to an AI tool like ChatGPT Enterprise, Claude AI Enterprise, Cursor, or a custom AI agent that supports MCP connectors. 
  • Note: The users for whom the AI Connector licenses are purchased should also have access to an MCP supported AI tool.

How to connect to the Udemy Business AI Connectors

The actual steps in your AI tool to configure a connection to the AI Connectors will vary by AI tool. In general, the steps will involve the following, if your tool supports Dynamic Client Registration (DCR):

1. Specify the MCP Sever endpoint: https://api.udemy.com/mcp
2. Choose the Transport (if available) as Streamable http.
3. Choose the Authentication as OAuth/OAuth 2.1
4. Click Use Dynamic Client Registration (DCR).
   • If you want to use Static Client Credentials instead and your AI tool supports it, see instructions for creating a client in Udemy Business, then paste the Client ID and Client Secret.
5. Accept any “Terms” for using connectors in your AI tool.
6. Click Connect and this will initiate a sign in to your Udemy Business account via SSO or your email and password, depending on your organization. Once signed in, you'll be required you to provide consent to allow the AI tool to access your Udemy Business account. Click Accept when asked to provide consent. 
   • Note: Ensure you trust the AI tools from which you are connecting to the Udemy Business AI Connectors.

(Alternative method) Create a static client in Udemy Business

If you want to use Static Client Credentials to connect your tool instead of Dynamic Client Registration, follow these steps to create an MCP client in Udemy Business to generate your Client ID and Client Secret.

1. In Udemy Business, navigate to Admin tools > Settings > MCP Server.

2. Click Create new client.

3. From the drop-down menu, select your client application, or select Other to enter a custom AI tool name. 

4. In the Callback URLs field, enter the OAuth callback/redirect URLs for your application. Check with your AI tool provider to get the callback/redirect URLs. For some of the standard AI Tools, you’ll find one or more callback/redirect URL(s) already entered. Note: Explicitly specifying the callback/redirect URLs ensures that only the authorized application handles the redirect, thus maintaining the integrity and identity of the applications involved.

5. After you click Create Client, you’ll be provided with a Client ID and Client Secret. These credentials allow your app to connect to the Udemy Business MCP server. 

Important security note: Save the Client ID and Client Secret in a secure location. This will be needed when configuring the AI tool to connect to the AI Connectors. For security reasons, you won’t be able to view the Client Secret again through your admin account. If you lose the client secret, you’ll need to delete the client and create a new client to restart the setup.

6. After safely saving your Client ID and Client Secret in a secure location, click Done. 

Configuration instructions for specific AI tools 

This section provides instructions for connecting to the AI Connectors for a sample of AI tools.

Note: These configuration instructions are for products external from Udemy and are provided here as examples. If the user interface for the AI tool differs from these instructions, please follow the prompts and documentation provided by the applicable AI tools instead. 

ChatGPT

1. Navigate to Settings > Apps & Connectors > Advanced Settings and enable ChatGPT developer mode

•  Note: This may need to be done by the ChatGPT admin.

2. In the Settings -> Apps & Connectors section in ChatGPT, click Create. 

3. Provide a name for the AI Connectors, example: Udemy Business MCP.

4. Enter the AI Connectors endpoint: https://api.udemy.com/mcp

5. Choose Dynamic Client Registration (DCR), then click Connect.

6. This will initiate a sign in to your Udemy Business account via SSO or your email and password, depending on your organization. Once signed in, you'll be required to provide consent to allow the AI tool to access your Udemy Business account. Click “Accept” when asked to provide consent.

7. Once connected, you’ll see a list of tools available as part of the AI Connectors.

8. As a ChatGPT admin you can then publish the connector to all your ChatGPT users in your Enterprise. They will then see the Udemy Business AI Connectors in their ChatGPT settings page, and they will just need to click Connect which will take them through Step 6 above.

Claude 

1. Navigate to claude.ai or open Claude desktop app.

2. Go to Organization settings > Connectors > Add (All Available) and search for Udemy Business. Add the connector to your organization in Claude.

3. Once the connector is enabled for the organization, individual users can go to their Settings > Connectors > Udemy Business > Connect.

Cursor

1. Go to Settings > Tools and Integrations > MCP Tools.

2. Click New MCP Server and add the following configuration under mcpServers:

"udemy-mcp": {
     "url": "https://api.udemy.com/mcp",
     "type": "http"
   }

3. Return to the Tools and Integration tab and authenticate the integration.

How to Configure Udemy Business MCP Server with Microsoft Copilot

Microsoft Copilot can access Udemy Business learning content through an agent created in Microsoft Copilot Studio. This requires a separate Microsoft Copilot Studio license (distinct from the standard Microsoft Copilot license as per Microsoft's SKU structure). 

The agent connects to the Udemy Business MCP server and can then be published to your Microsoft tenant, making it available through channels like Microsoft Copilot Chat, Microsoft Teams, and other Microsoft 365 applications.

Prerequisites

Before configuring the Udemy Business AI Connectors with Microsoft Copilot, you’ll need:

• A Microsoft Copilot Studio license
• Admin access to Microsoft Copilot Studio
• Admin access to your organization's Udemy Business account
• Udemy Business Enterprise Plan with AI Connectors add-on subscription

Configuration Steps

Step 1: Create an Agent in Microsoft Copilot Studio

1. Navigate to Microsoft Copilot Studio and create a new agent.
2. Provide a name for your agent. Example: UB Skills Discovery Agent
3. Enter system instructions that define how the agent should behave, and how you’d want the tool results to be formatted.  
   • See a sample system prompt.

Step 2: Add Udemy Business MCP Server as a Tool

1. Navigate to the Tools tab in your agent
2. Click Add a tool.
3. Select the Model Context Protocol card from the available tool options.

Step 3: Configure the MCP server connection

In the MCP server configuration pop-up, fill in the following details:

1. Server name: Enter a descriptive name (Udemy MCP)
2. Server description: Add a description
3. Server URL: https://api.udemy.com/mcp
4. Authentication: Select OAuth > Dynamic discovery
5. Click Create.

.

Step 4: Create and authenticate the connection

1. Once the tool is created, you'll see a Connection pop-up.

2. Click the connection icon and select Create new connection. 

   

3. Click Create.
4. You'll be redirected to the Udemy Business login screen.
5. Sign into your Udemy Business instance using your credentials.
6. Provide consent for Microsoft Copilot Studio to access your Udemy Business account.

Step 5: Verify connected tools

After establishing the connection, you should see the available Udemy Business tools listed under the connected MCP server.

Step 6: Publish the agent

1. Once the agent is created and tools are successfully configured, click Publish. 

2. Configure the appropriate Share settings to make the agent available to your tenant users.

3. Under the Channels tab, select the channels where you want this agent to be available (for example, Microsoft Copilot Chat, Microsoft Teams, SharePoint, etc.)

Step 7: Access the agent

Once published, your Copilot Chat users can interact with the agent by:

• Typing @agentName in Copilot Chat (replace agentName with your agent's name).

  

Example of a chat with the agent:

Important Authentication Note

The Udemy Business MCP server requires OAuth authentication. End users will need to sign in and provide consent for the agent's tools to function properly. Check with your IT team to understand how authentication is configured for users in your Microsoft tenant when accessing OAuth connectors with agents created in Copilot Studio.

Refreshing the connector to access new tools

If you connected the Udemy Business AI Connectors before a new tool was added, your AI tool may or may not automatically detect the updated tool list. Most AI tools allow you to refresh the tools list to fetch new tools or they might do this automatically when you start a new conversation or sign out and sign back in.

After reconnecting, verify that create_learning_path appears in your tool list alongside the existing search tools.

Troubleshooting 

This section provides troubleshooting tips for some common issues with the AI Connectors. If you’re unable to resolve your issue, please contact Udemy Business Support.

Note: If you’re experiencing issues during configuration, ensure your organization, your admin account, and your chosen AI tool meet the prerequisites set out at the beginning of this article. 

Issues with authentication or OAuth

• Double check the Client ID and Client Secret.
• Disconnect the AI Connectors connection, then go through the OAuth flow again to reauthenticate.
• Log out of your Udemy Business account and then sign back in. 
• After making configuration changes, restarting desktop applications might be necessary for the updates to take effect.

Issues with the AI Connectors client 

• If the Udemy Business AI Connectors tool isn’t working in Cursor, it could be caused by expired or incorrect authentication tokens. To fix this:
  1. In the Command Palette, search for and select Clear all MCP tokens.
  2. Reauthenticate your accounts.
• When queries fail due to timeouts, consider simplifying or breaking larger requests into smaller, more manageable inputs. 
• When you encounter errors related to token limits, consider starting a new conversation thread or adjusting the token limit in your MCP client settings. 
• If tools fail to load, restarting the MCP client application and logging back in should resolve the issue.
• create_learning_path shows a confirmation prompt before running.
  • This is expected behaviour. Because this tool creates content in your Udemy Business account, your AI tool will require you to approve the action before it executes. Click Allow or Confirm when prompted to proceed with path creation.

Issues with the Microsoft Copilot connection

• If the agent's tools aren't working, verify that end users have completed the OAuth sign-in and consent flow.
• Check with your IT team about OAuth connector policies in your Microsoft tenant.
• Ensure your Microsoft Copilot Studio license is active and properly configured.

Why don’t I see the Udemy tools listed in my AI tool after configuration?

• Confirm connection to the correct AI Connectors endpoint.
•  All AI Connector tools require an active AI Connectors add-on subscription. If you recently connected and don't see all tools listed, try refreshing or reconnecting the connector in your AI tool — new tools added after your initial connection will not appear automatically. If you still see significant differences after reconnecting, please reach out to our support team.

Sample system prompt

Role: You are an Enterprise Learning Assistant powered by Udemy Business. Help users discover learning content to close skill gaps and advance their careers.

Your Tools:

- search_courses - Full courses (2-40 hours) for comprehensive skill development

- search_learning_paths - Curated multi-course journeys for role-based learning

- search_lectures - Quick 5-15 min videos for just-in-time learning

- `search_labs` - Hands-on cloud practice environments (AWS, Azure, DevOps)

- `get_course_curriculum` - Detailed breakdown of any course content

- `create_learning_path` - Build and deploy a structured learning path in Udemy Business based on a skill, audience, and proficiency gap

How to Help:

1. Understand the need - Ask about their role, skill gap, or learning goal

2. Search smartly - Use appropriate tools and filters (level, duration, ratings)

3. Recommend with context - Explain WHY each suggestion fits their need

4. Keep it simple - Show 2-4 top recommendations, not everything

5. Build when asked - If the user wants a deployable program, chain search_courses → get_course_curriculum → create_learning_path to assemble and save a structured path directly in Udemy Business. Always inspect curricula before including a course.

Tool Chaining — Learning Path Creation

When a user asks to build, create, or deploy a learning path, follow these steps in order. Do not skip steps.

1. SEARCH: Use search_courses to find relevant courses matching the skill, audience, and proficiency level. If the topic is technical or cloud-based, also use search_labs to find a relevant hands-on lab.

2. INSPECT: For each course you plan to include, call get_course_curriculum to review its sections and confirm relevance before committing it to the path. Do not include a course without inspecting it first.

3. CREATE: Once you have confirmed the right content, call create_learning_path exactly once with the assembled course and lab selections.

4. CONFIRM: After the path is created, share the path title and link with the user and remind them that the path is private by default. Direct them to their Udemy Business admin tools to assign it or make it discoverable.

Important: create_learning_path writes to the user's Udemy Business account. The user's AI tool will ask for confirmation before this step executes — this is expected behavior.

Output Format:

CRITICAL: Format ALL course/lecture/path/lab titles as bold clickable links using this exact syntax:

Title Text

Present recommendations using this structure:

```

[Brief 1-2 sentence answer to their question]

Recommended Learning:

📚 Course Title  

- ⏱️ Duration: [time] | ⭐ Rating: [stars] ([reviews])

- 💡 Why: [1 sentence explaining fit]

📚 Course Title  

⏱️ Duration: [time] | ⭐ Rating: [stars] ([reviews])  

💡 Why: [1 sentence explaining fit]

Next Steps:

- [Suggested action 1]

- [Suggested action 2]

```

Icon Prefix Guide:

- 📚 for Courses

- 🎯 for Learning Paths  

- 🎓 for Lectures

- 🔬 for Labs

Best Practices:

- Match difficulty level to user's experience

- For technical skills, combine courses + labs

- For quick answers, use lectures

- For role transitions, use learning paths

- Always format titles as Title - never show raw URLs

- Highlight quality signals (4.5+ ratings, bestseller badges)

Example:

User: "I need to learn Python for data analysis"

Response:

```

Python is essential for data analysis. Here are targeted resources to get you started:

Recommended Learning:

📚 Python for Data Science and Machine Learning Bootcamp  

⏱️ Duration: 25 hours | ⭐ Rating: 4.6 (135K reviews)  

💡 Why: Comprehensive coverage of pandas, NumPy, matplotlib - exactly what analysts 

need

🎓 Introduction to Pandas DataFrames  

⏱️ Duration: 12 minutes | ⭐ Rating: 4.5  

💡 Why: Quick-start lecture if you need to work with data immediately

Next Steps:

- Want hands-on practice? Ask me to find Python labs

- Need a full learning path? I can search role-based programs for data analysts

```

User: "Build a learning path for our cloud engineers moving from intermediate to 

advanced AWS"

Response:

Here's what I'll build — a structured path for cloud engineers progressing from 

intermediate to advanced AWS skills:

  🎯 AWS Cloud Engineering: Intermediate to Advanced

  - Section 1: Core AWS Services Deep Dive (Course — 2h 30m)

  - Section 2: AWS Architecture and Best Practices (Course — 3h 15m)

  - Section 3: Hands-On AWS Infrastructure Lab (Lab — 2h)
 

The path has been created in your Udemy Business account and is currently set to 

private. You can assign it to your team or make it discoverable from your Learning 

Paths admin tools.

Next Steps:

- Ready to assign this path to your team? Open it in Udemy Business admin tools.

---

FORMATTING RULES - MUST FOLLOW:

1.  Format links as:  Title  with emoji icon prefix

2.  Never show bare URLs or "Link:" labels

3.  List duration, rating, and "why" as an unordered list using "-" bullets under 

each content item

4.  Add TWO blank lines between any two content items

5.  Use CAPS for section headers (RECOMMENDED LEARNING, NEXT STEPS)

6.  Use emoji bullets in the unordered list (⏱️ for duration, ⭐ for rating, 💡 for 

why)

7.  Keep "why" explanation to one concise sentence

Keep responses concise, helpful, and visually clean.