To use the Multifunction Calling using Microsoft Teams AI Agent recipe, the following prerequisites must be met:
•Create an Azure application to use Microsoft Teams connector.
•Generate JWT token to configure Snowflake Cortex AI connector.
Create an Azure application to use Microsoft Teams connector
To use Microsoft Teams connector, you must create an Azure application for a Microsoft Teams connector.
1Open the Azure portal.
2Navigate to Microsoft Entra ID.
3Go to Manage > App registration > New registration and enter a name for your application.
4For supported account types, choose the appropriate option based on your application needs. For redirect URI, select Web and enter your redirect URL.
5Click Register.
6Go to the newly registered application's overview page.
7Go to Manage > Authentication and add the redirect URIs if you haven't already.
8In the Implicit grant section, enable Access tokens and ID tokens.
9Go to Manage > API permissions > Add a permission and select Microsoft Graph.
10Choose Delegated permissions and select necessary permissions, such as User.Read, email, offline\_access, openid, and profile and click Add permissions.
11On the application's overview page, copy the Application (client) ID.
12Go to Manage > Certificates & secrets > Client secrets and click New client secret as shown in the following image:
13Enter a description and select an expiration date and click Add.
14Copy the value of the new client secret. It will not be displayed again.
15Use the client ID and secret to connect your Microsoft Teams connector application to Azure.
Generate JWT token to configure Snowflake Cortex AI connector
To generate the JWT token for configuring the Snowflake Cortex AI connector, the following prerequisites must be met:
•Generate the private key and the public key
- To generate a private key, run the following command in your terminal window:
To get the public key fingerprint, run the following query in the Snowflake Web UI:
DESC USER [USER_NAME]
The following image shows the RSA public key fingerprint:
•Use the cross-region inference parameter
By default, the cross-region inference parameter is set to DISABLED, which means the inference requests are only processed in the default region.
The following examples show how to set the cross-region parameter for various use cases:
Any region
To allow any of the Snowflake regions that support cross-region inference requests to process your requests, set the parameter to 'ANY_REGION'.
•ALTER ACCOUNT SET CORTEX_ENABLED_CROSS_REGION = 'ANY_REGION';
Default region only
To process inference requests only in the default region, set this parameter to 'DISABLED'.
ALTER ACCOUNT SET CORTEX_ENABLED_CROSS_REGION = 'DISABLED';
Specify regions
To allow only specified regions to process your requests, set this parameter to the regions separated by commas. For a full list of regions, see CORTEX_ENABLED_CROSS_REGION.
The following example specifies AWS_US and AWS_EU regions to process your inference requests:
•ALTER ACCOUNT SET CORTEX_ENABLED_CROSS_REGION = 'AWS_US,AWS_EU';