Skip to main content

One drive User Connector Setup Guide

This guide provides step-by-step instructions for creating and configuring a One drive connector with Abacus.AI's ChatLLM platform.

An Entra admin must set this connector up first

Microsoft blocks ordinary users from consenting to the Abacus.AI app on their own. Before anyone else in your tenant can connect OneDrive, a Microsoft Entra admin must create the connector once and tick the "Consent on behalf of your organization" checkbox on Microsoft's Permissions requested screen, then click Accept.

This is not the same as the Grant admin consent button in the Entra admin center. Once the admin has ticked the checkbox, every user in that tenant can sign in and create their own connector freely — until then Microsoft stops them with "Need admin approval".

See Microsoft Entra Tenant Consent for the full walkthrough.

Use OneDrive directly in chat — no Custom Chatbot required

Once OneDrive is connected from Profile → User Connectors, it works straight away in ChatLLM Teams chats with RouteLLM selected, and with the Abacus AI Agent.

  1. In ChatLLM Teams, click Connectors on the home page (or go to Profile → User Connectors) and connect OneDrive.
  2. Open a new chat and select RouteLLM (route-llm) from the model selector.
  3. Ask a question about your OneDrive data — for example "Find my OneDrive files updated this week". The first time you query it you may be asked to authorize the connector; complete the sign-in once and the request continues automatically.

The Custom Chatbot steps below are only needed when you want a purpose-built bot with its own instructions and a restricted set of tools. See the User Connectors guide for the full walkthrough.

Prerequisites​

Before starting, ensure you have access to:

  • Abacus.AI platform
  • A valid One drive account
  • Administrative access to create connectors and projects

Custom OAuth (optional)​

By default we request access using our app to read data. If you want to use your own app, please follow following steps to create the app.

  1. In the Abacus.AI interface, click on your profile picture in the top right corner.
  2. Select "Manage Connectors" from the dropdown menu.
  3. Click on "Add New Connector", choose "Onedrive" from the list.
  4. Turn the 'Config Only' toggle on.
Onedrive_user

Note: Please enable the Config Only toggle; otherwise, users will end up gaining access to other users onedrive.

  1. Copy the app’s Client ID, Client Secret and Tenant ID, then enter them into the CLIENT ID, CLIENT SECRET and TENANT ID fields respectively. Note: Please refer to microsft documentaiton to create the app and add necessary scopes to app. Add the redirect URI https://abacus.ai/oauth/callback to your OAuth app so that Abacus.AI can complete the OAuth authorization flow.
  2. Click create. ** Note: The configuration process for the connector does not require a OneDrive connection. If you are prompted to connect to OneDrive, please review the preceding steps to ensure the configuration is set up correctly.**

Step-by-Step Instructions​

1. Navigate to Projects Page​

Navigate to the projects page by clicking on the ABACUS.AI logo.

2. Create New Project​

Click on "Create a new project" to begin setting up your ChatLLM project.

3. Select Project Type​

Choose GenAI --> Custom Chatbot option from the available project types.

4. Configure Project Name​

Enter a descriptive name for your project that reflects its purpose.

5. Skip to Dashboard​

Select "Skip to project dashboard" to proceed directly to the project configuration.

6. Access Model Training​

  1. Click on the Model option in the left toolbar
  2. Select "Train Model" in the top right corner of the page

7. Configure Advanced Options​

Navigate through the following menu structure:

  • Select Advanced Option
  • Choose Tool Use
  • Select One_Drive_tool
  • (Optional) When using your own app, navigate to Config Connectors and pick the connector configured in the Custom Auth step.

8. Initiate Model Training​

Click on "Train Model" to begin the training process.

9. Access Trained Model​

Once training is complete:

  1. Click on Models in the navigation
  2. Select your newly trained model

10. Create Deployment​

Click on "Create a new deployment" to make your model available for use.

11. Configure Deployment Type​

  1. Select "Offline Batch + Realtime" deployment option
  2. Click "Next" to proceed

12. Name Your Deployment​

Enter a user-friendly name that your end users will see, then click "Deploy".

13. Monitor Deployment Status​

  1. Return to your model by clicking Models → [Your Model Name]
  2. Wait for the deployment status to show as "Active"
  3. Click on the deployment name once active

14. Access Prediction Dashboard​

Click on the "Prediction Dash" option under the Deployment section.

15. Test Your Bot​

Click on "Go to Abacus.AI Chat" to access the testing interface.

16. Start Using the Chatbot​

  • Enter your questions or prompts in the chat interface
  • Note: Each user will be prompted to log in to One drive once for authentication

Permissions and Access​

When setting up the connector, the following permissions are requested. Please note that all the permissions are delegated permissions:

  • Files.Read: Read user files
  • Files.Read.All: Read all files that user can access
  • Files.ReadWrite.All: Have full access to all files user can access
  • offline_access: Maintain access to data you have given it access to
  • openid: Sign users in
  • profile: View users' basic profile
  • SensitivityLabel.Read: Read the user's sensitivity labels
  • User.Read: Sign in and read user profile

These permissions allow the connector to pull necessary data.

Important Notes​

  • Authentication: Users will need to authenticate with One drive on their first interaction
  • Permissions: Ensure proper One drive permissions are configured
  • Sensitivity Labels: The connector supports Microsoft Purview sensitivity labels. If configured, files matching specified sensitivity labels can be excluded during data ingestion. Ensure the SensitivityLabel.Read delegated permission is granted.
  • Testing: Thoroughly test the connector with various One drive operations before deploying to end users

Troubleshooting​

If you encounter issues:

  1. Check deployment status is "Active"
  2. Ensure proper authentication credentials
  3. Review One drive permissions and access rights