top of page

How to Connect Qlik to Microsoft Teams with MCP

12 hours ago
6 min read

In this Data Voyagers tutorial, I show how to connect Qlik to Microsoft Teams using MCP and Microsoft Copilot Studio, including how to resolve an OAuth registration error encountered during setup. Watch the video below to follow the process on screen, then find the written step-by-step guide right after it.



You can use Microsoft Teams to ask an AI agent questions about the Qlik content you can access. The connection runs through Microsoft Copilot Studio, where you create an agent, attach Qlik’s Model Context Protocol (MCP) server, and publish the agent to Teams.

In this tutorial, we’ll follow the setup I tested for Data Voyagers, including an OAuth registration error we encountered and the setting that resolved it.


What we’re building


The setup has three parts:


  • Microsoft Teams provides the chat interface.

  • Microsoft Copilot Studio hosts the agent and connects it to tools.

  • Qlik’s MCP server provides access to Qlik Cloud capabilities.


Our first test will be simple: ask the agent to list the shared spaces available through its Qlik connection. That gives us a clear way to verify that the agent is retrieving real tenant information.


Before you start


Have the following ready:


  • A Qlik Cloud tenant with MCP access.

  • Qlik tenant administrator access to configure OAuth registration and approve the initial connection.

  • Access to Microsoft Copilot Studio with permission to create and publish agents.

  • A Microsoft Teams account in your organization.

  • Permission to install the published agent in Teams.


Your organization’s Microsoft and Power Platform policies must also permit the connection and deployment.


The screenshots in the video above show the interface used during this setup. Labels and menu locations may change as the products evolve.


Step 1: Prepare OAuth registration in Qlik Cloud


We’ll use OAuth Dynamic Client Registration, or DCR. This allows a compatible client to register its OAuth application with Qlik during connection setup.


  1. Sign in to Qlik Cloud as a tenant administrator.

  2. Open the Administration activity center.

  3. Go to Settings → Security.

  4. Find OAuth dynamic client registration (DCR).

  5. Turn on Enable dynamic client registration.

  6. Under Allowed authentication methods, select Client secret + None.

  7. Save the change if prompted.


The authentication-method selection matters. In our test, Copilot Studio requested the public-client method none. Qlik initially rejected it because the tenant permitted only clients using a client secret.


Here, None refers to client authentication at the OAuth token endpoint. Users still sign in to Qlik. Qlik documents that public clients use authorization code flow with PKCE.

This is a tenant-level registration policy, so configure it according to your organization’s requirements.



Step 2: Create your agent in Copilot Studio


  1. Open Microsoft Copilot Studio.

  2. Select the environment where you want to build the agent.

  3. Create a new agent.

  4. Give it a clear name. In our example, we used Qlik Data Voyagers Assistant.

  5. Add a description explaining its purpose.

  6. Save the agent.


For example, the description could be:

An assistant that uses Qlik tools to help users find and explore content available through their Qlik Cloud connection.

Step 3: Add the Qlik MCP server


Open the agent’s tool configuration. Depending on your Copilot Studio interface, use the + beside Tools, or follow Tools → Add a tool → New tool.


  1. Select Model Context Protocol.

  2. Enter a server name, such as Qlik Data Voyagers.

  3. Add a description explaining what the connection provides.

  4. Enter your Qlik MCP server URL.


The URL follows this pattern:

https://<your-tenant>.<region>.qlikcloud.com/api/ai/mcp

For the Data Voyagers tenant, we used:

Use your own tenant hostname and region.


Step 4: Authenticate and attach the tool


  1. Choose OAuth 2.0.

  2. Select Dynamic discovery.

  3. Create the MCP tool configuration.

  4. When asked to select a connection, choose Create a new connection.

  5. Enter a display name if desired.

  6. Click Create.

  7. Complete the Qlik sign-in and administrator approval.

  8. Select Add to agent when prompted.


Confirm that the Qlik connection appears under the agent’s Tools section.

For the initial setup, use a Qlik tenant administrator account so the registration can be approved. Qlik automatically removes unapproved DCR registrations after one hour.


Step 5: Give the agent clear instructions

The connection provides tools. Instructions help the agent decide when and how to use them.


The following is a suggested starting point:

You help users find and explore Qlik Cloud content through the connected Qlik MCP tools.Use Qlik tools for questions about tenant content, spaces, applications, fields, measures, and data. Retrieve current information before answering these questions.Never invent resource names, identifiers, or results. If a tool fails, explain the failure and identify what information is needed to continue.Ask for clarification when multiple applications match the user’s request.For analytical calculations, use Qlik’s calculation capabilities and report the returned results.Before making changes to Qlik content, explain the proposed action and ask the user to confirm.Keep answers clear and concise.

These instructions guide behavior. Access restrictions should be enforced through the connection configuration and Qlik permissions.


Save your changes.


Step 6: Test the Qlik connection


Before publishing, use the agent’s preview or test chat.

Send this request:

Use the Qlik Data Voyagers tool to list the shared spaces I can access.

Replace the tool name if you chose a different one.


Check that the response includes actual space names from your tenant. A general explanation of how to find spaces does not verify the connection.


You can follow with a more specific request:

Find the applications in the Demonstrations space.

Use a space that exists in your tenant.


Step 7: Add the Teams channel


In the interface shown in our video:


  1. Find Channels in the agent’s configuration panel.

  2. Click the + button.

  3. Select Teams + Microsoft 365.

  4. Complete the channel configuration.


If a separate option appears to make the agent available in Microsoft 365 Copilot, select it according to where you want users to access the agent.


Step 8: Publish the agent


  1. Save your changes.

  2. Click Publish.

  3. Complete any required review or confirmation.

  4. Wait for the success message.


In our test, Copilot Studio displayed:


Your agent published successfully

The dialog also provided a View in Teams + Microsoft 365 link.


Publishing makes the current agent configuration available through its configured channels.


It does not, by itself, prove that Qlik tool calls will work in the published chat.


Step 9: Open and test the agent in Teams


  1. Click View in Teams + Microsoft 365.

  2. Open Teams when prompted, or continue in the browser.

  3. Select Add or Open on the agent page.

  4. Start a conversation with the agent.

  5. Ask it to list your shared spaces.

  6. Complete any Qlik connection or sign-in prompt.

  7. Repeat the question if necessary.


Compare the response with the content you can access in Qlik Cloud.


Step 10: Verify access before sharing


Once your own test works, test with a colleague.


Check which credentials the published Qlik tool uses. A successful test with the maker’s account does not establish how another user’s connection will behave.


Have your colleague sign in and ask for content they are permitted to access. Verify that the returned information matches their Qlik permissions before distributing the agent more widely.


Then use Copilot Studio’s sharing options. Organization-wide availability may require Microsoft administrator approval.


Troubleshooting: “token_endpoint_auth_method ‘none’ is not allowed”

During our setup, connection creation failed with:

invalid_client_metadata
token_endpoint_auth_method "none" is not allowed for this tenant

DCR was already enabled, and the account was a tenant administrator. The problem was the allowed authentication method.


Copilot Studio requested none, while Qlik allowed only Client secret.


To resolve it:


  1. Open Qlik Cloud → Administration → Settings → Security.

  2. Find the DCR settings.

  3. Change Allowed authentication methods to Client secret + None.

  4. Save if prompted.

  5. Return to Copilot Studio.

  6. Cancel the failed connection dialog.

  7. Create the connection again.

  8. Complete sign-in and approval.


This resolved the registration error in our test.


How to read a truncated connection error


Copilot Studio may display only part of the error in its banner. Before changing settings, retrieve the complete message.


In Microsoft Edge:


  1. Right-click the error text.

  2. Select Inspect.

  3. In the Elements panel, locate the selected element containing the message.

  4. Right-click that element.

  5. Select Copy → Copy outerHTML.

  6. Paste it into a text editor and read the full error.


This was how we found the authentication-method rejection.


Also, a Network request can return HTTP 200 while a separate internal operation fails. The absence of a red request does not establish that the connection succeeded.


What to try next


After the shared-space test works, try requests such as:


  • “Find applications related to sales.”

  • “Describe the data model in this application.”

  • “List the master measures available in this app.”

  • “Use Qlik to calculate sales by region.”


Available actions depend on the tools exposed by the MCP server and the permissions associated with the connection. Start with a known application and compare results against Qlik before expanding the agent’s use.


Watch the accompanying Data Voyagers video to follow the setup on screen, including the registration error and its resolution.

 
 
 

Comments


© 2026 Data Voyagers

  • Youtube
  • LinkedIn
bottom of page