Salesforce Einstein: Global Model Opt-Out Process

Salesforce Einstein: Global Model Opt-Out Process

Einstein intelligence is built on data. And the more data Einstein has, the more powerful the predictions are. But not everyone has enough data to build a predictive model of their own, and in many cases, no one customer does. To give customers more meaningful AI, we’re building global predictive models that all customers can use. Global models look for aggregate, anonymous trends across many Salesforce customers, so the first step is to gather data from multiple Salesforce orgs. We’ll use the data to build global models.

How do I opt out of the Global Models for Salesforce Einstein?
If the following are true, global models may include your data:

• Your Salesforce organization has at least one Einstein license
• Your organization has signed an Order Form allowing us to use your data submitted to Einstein
• Your organization has enabled at least one Einstein feature.

If you want to exclude your data from Salesforce Einstein global models, you can opt out. If you opt out, we stop using your data to build global models. Additionally, Salesforce will not use your Customer Data for research and development of new services or features, and no Salesforce employee will view your Customer Data outside of a support case, pilot, or as otherwise described in your legal agreements.

This is now an Administrator controlled setting. To opt out, you can follow the steps in Manage Salesforce Access to Customer Data

If this setting is not visible to you once Einstein is enabled, please submit a support Case with the Activations Team for further assistance. 

In the Case Subject please include: Global Model opt-out request.
You do not need to include a Business Case or justification for the opt-out request.

NOTE: If Support assistance is needed, you need to submit a case for each org that you wish to opt out. Sandbox orgs aren't impacted by this change as they are already excluded from the global model.

Additional Information

What are global models?
Currently, most Salesforce Einstein features build a unique predictive model based on each customer’s own Salesforce data. This single-customer model is great for applications that don’t require a lot of data, but some models require more data than any single customer can provide. For example, query classification and ranking models used by Einstein Search require data sets that are only achievable through global models. Global models combine data from multiple Salesforce orgs to create a predictive model that can be used by many. Even if you have enough data to create a single-customer model, there are times when a global model provides better predictions.

Opt out applies to data use in the following features and services:

• Einstein Activity Capture
• Einstein Article Recommendations
• Einstein Bots
• Einstein Generative AI for Service
• Einstein Generative AI for Sales
• Einstein Generative AI for Commerce
• Einstein Case Classification / Case Wrap-up 
• Einstein Conversation Mining
• Einstein Reply Recommendations
• Einstein Deal Insights
• Einstein Prediction Builder (both as a Service and as a feature of Tableau CRM Plus)
• Einstein Opportunity Scoring
• Einstein Recommendation Builder
• Einstein Search
• Einstein Search for Knowledge
• Einstein Vision & Language
• High Velocity Sales
• Marketing Cloud Account Engagement Einstein/ Marketing Cloud Account Engagement Advanced
• Sales Cloud Einstein
• Salesforce Inbox
• Service Cloud Einstein
• Services and features branded as Agentforce
• Services and features that use Hybrid Search or Search Actions

How do global models impact my data?
If your Salesforce data is included in global models, rest assured that nothing changes in terms of security, privacy, or data residency. Your data stays in our Trusted AWS environment, described in the Einstein Trust and Compliance Documentation. When applied to Einstein features, global models never cause any of your data, including confidential information, to be shared with other customers. How can we be so sure? Because global models work by looking for aggregate, anonymous trends from multiple sources.

As an example, for Einstein Opportunity Scoring, a global model looks at data from multiple Salesforce customers and detects different trends. By analyzing a large volume of activities, the global model might, for instance, detect correlation between the timing of an activity and the likelihood to win. In another example, the global model identifies a pattern where opportunities that haven’t had any activity over the last two weeks are 50 percent less likely to close. Such trends are then applied to each Customer’s Data and used to make Customer-specific predictions about how likely an Opportunity is to close. The trends encoded in the model don’t contain any individually identifiable data, and the output, which in this example is the Customer-specific prediction on an Opportunity, is only ever shared with the Customer.