Set Up Your Data in Vertex AI

Access Salesforce data in a Google Cloud Vertex AI notebook using a Python connector. The user-friendly interface lets you quickly query your data in Data 360 so you can extract the data that you need into Python.

1. In Vertex AI Workbench, install the CDP Python Connector to access Data 360 objects in and ingest Salesforce data into Vertex AI.
2. As part of the Python Connector configuration, create an external client app or connected app.
   

   Important Creating connected apps is restricted as of Spring ‘26. You can continue to use existing connected apps during and after Spring ‘26. However, we recommend using external client apps instead. If you must continue creating connected apps, contact Salesforce Support.

   See New connected apps can no longer be created in Spring ‘26 for more details.

3. Authenticate to securely access your data using one of these methods.
   • Google Service Account Credential
   • API Key
4. In Data 360, connect to your Salesforce data source in Vertex AI to query the dataset. For example, to get your email engagement data, access the Email Engagement object in Marketing Cloud Engagement.
5. To consume the inferences in Data 360, upload the Python file to your notebook.
6. To get your model endpoint, go to the Vertex AI dashboard in the Online Prediction section. Note the URL as you need these details to connect your model with Salesforce.