Setting up Gemini in Deskpro is done in two main parts, which can be summarised as:
-
Setup of a Google Cloud project with billing, enabling Vertex AI APIs, and configuring a service account with a key file
-
Create the Google cloud project
-
Enable billing
-
Enable the Vertex AI API
-
Add a Service account in IAM
-
Add a key to the service account with the necessary permissions to use Vertex AI
-
Download the key file
-
-
Adding the Gemini AI connection to Deskpro to use the Google Cloud project created and using the service account key file
-
Select to add the connection
-
Select Gemini as the provider
-
Enter the Google Cloud project ID
-
Add the service account key file
-
Select the prompting and/or embedding models to use
-
Verify the connection settings
-
Create the connection
-
Setup Google Cloud project
To use Gemini with Deskpro, you’ll need access to Google Cloud. Deskpro uses the Google Cloud Vertex AI service to access provisioned Gemini models provided by Google.
Create the Google Cloud project. Ensure billing is configured for the new project.
Once enabled, open the project and load the APIs and Services product
Check if the Vertex AI API is enabled in the list. If it isn’t, click Enable APIs and Services button
Search for the Vertex AI API and select it, click Enable to activate the API
Go to the IAM and Admin service
Click Service Accounts in the left side menu
Give the service account a name and optionally a description and click Create and Continue
Add the roles “Vertex AI User” and “Vertex AI Viewer” and click Done
Open the newly created service account and click the Keys tab
Click Add Key and click Create new Key
Keep the option for a JSON key selected and click Create. The key file will be automatically downloaded by the browser. Retain this file for creating the AI connection in Deskpro.
Add Gemini AI Connection to Deskpro
Go to the New AI connections screen. Click New connection in the right
In the New Connection drawer, select “Google Gemini (Vertex AI)” from the provider drop down:
The form will change. Now, copy the Google Cloud project ID to the project ID field
Click Browse under the Service Account key file field and select the JSON Key file of the Google Cloud service account that will be used
Select a text generation and/or embedding model to use for the connection
Click Verify settings to confirm Deskpro is able to communicate with Google Cloud for the project ID given using the service account key file.
You should see the message that the Vertex AI connection is valid. If you do not get this, check the details you have entered in the form:
The last steps are to:
- Tick the box to accept terms and conditions
- Finally, click Create
If you have not set up your AI features continue to manage further productivity features.