Architectural Overview
The PardotEinsteinSettings metadata type controls machine learning and predictive analytics features connecting Salesforce CRM with Marketing Cloud Account Engagement (formerly Pardot). It governs features such as Einstein Behavior Scoring, Einstein Campaign Insights, and Einstein Attribution.
From an application architecture perspective, activating these models provisions background data sync jobs that analyze prospect engagement signals—such as email opens, form submissions, and video interactions—to generate predictive scores on Lead and Contact records. CI/CD pipelines deploying custom FlexiPage layouts with Einstein scoring widgets must ensure PardotEinsteinSettings is deployed first to avoid component dependency errors.
Directory and File Conventions
In SFDX source format, PardotEinsteinSettings is stored in the settings/ directory and referenced using the Settings type with member PardotEinstein.
| Property | Value |
|---|---|
| Metadata Type Name | Settings (member: PardotEinstein) / PardotEinsteinSettings |
| SFDX Directory | force-app/main/default/settings/ |
| File Suffix | .settings-meta.xml |
| Full File Path | force-app/main/default/settings/PardotEinstein.settings-meta.xml |
| Scratch Org Setting Key | pardotEinsteinSettings |
Key Metadata Fields
The XML schema of PardotEinsteinSettings configures individual AI model activations:
| Field Name | Field Type | Required | Architectural Description |
|---|---|---|---|
enableEinsteinBehaviorScoring |
boolean |
No | Enables machine learning evaluation of prospect actions to produce engagement behavior scores. |
enableEinsteinCampaignInsights |
boolean |
No | Analyzes campaign recipient demographics and engagement to identify key trends. |
enableEinsteinAttribution |
boolean |
No | Activates multi-touch AI campaign influence modeling on Opportunity records. |
enableEinsteinKeyAccounts |
boolean |
No | Enables Account-Based Marketing (ABM) predictive scoring on target account hierarchies. |
Note
Training Einstein Behavior Scoring models requires a minimum baseline of historical prospect engagement records. In new scratch orgs or sandboxes without data, the models remain idle until sufficient activity history is generated.
Declarative XML and Package Manifest Examples
Below is an enterprise configuration for activating Pardot Einstein AI features:
<?xml version="1.0" encoding="UTF-8"?>
<PardotEinsteinSettings xmlns="http://soap.sforce.com/2006/04/metadata">
<enableEinsteinAttribution>true</enableEinsteinAttribution>
<enableEinsteinBehaviorScoring>true</enableEinsteinBehaviorScoring>
<enableEinsteinCampaignInsights>true</enableEinsteinCampaignInsights>
<enableEinsteinKeyAccounts>true</enableEinsteinKeyAccounts>
</PardotEinsteinSettings>Scratch Org Definition Configuration
When creating scratch orgs for testing Account Engagement components:
{
"orgName": "Pardot Einstein Testing Org",
"edition": "Enterprise",
"features": ["Pardot", "EinsteinAnalytics"],
"settings": {
"pardotEinsteinSettings": {
"enableEinsteinBehaviorScoring": true,
"enableEinsteinCampaignInsights": true,
"enableEinsteinAttribution": true
}
}
}Package Manifest (package.xml)
<?xml version="1.0" encoding="UTF-8"?>
<Package xmlns="http://soap.sforce.com/2006/04/metadata">
<types>
<members>PardotEinstein</members>
<name>Settings</name>
</types>
<version>68.0</version>
</Package>CLI Retrieval and Deployment Commands
Manage PardotEinsteinSettings via the Salesforce CLI (sf):
# Retrieve PardotEinsteinSettings from the target org
sf project retrieve start --metadata Settings:PardotEinstein
# Deploy PardotEinsteinSettings to synchronize AI configurations
sf project deploy start --metadata Settings:PardotEinstein
# Validate deployment into a target sandbox
sf project deploy validate --metadata Settings:PardotEinsteinArchitect Best Practices and Gotchas
- Feature License Requirements: Deploying
PardotEinsteinSettingsinto an environment lacking Marketing Cloud Account Engagement Plus, Advanced, or Premium editions will fail with license entitlement errors. - Lightning Page Dependencies: Do not add the Einstein Behavior Scoring Lightning component to Lead or Contact record pages in packages without bundling
PardotEinsteinSettingsas a pre-requisite. - Data Refresh Lag: When
enableEinsteinBehaviorScoringis enabled, initial model scoring cycles run asynchronously every few hours. Do not expect instantaneous score recalculation upon record creation.
Warning
Disabling enableEinsteinBehaviorScoring wipes model state and historical scoring variance tracking, requiring a complete retraining cycle upon reactivation.
Frequently asked questions
What capabilities are activated by PardotEinsteinSettings?
It activates Einstein Behavior Scoring, Einstein Campaign Insights, and Einstein Attribution for Marketing Cloud Account Engagement.
Where is PardotEinsteinSettings stored in an SFDX directory?
It is stored under force-app/main/default/settings/ as PardotEinstein.settings-meta.xml and retrieved as Settings:PardotEinstein.
Why is PardotEinsteinSettings required for Lightning page deployments?
Lightning pages containing the Einstein Behavior Scoring or Campaign Insights components fail deployment unless PardotEinsteinSettings is active.
Can PardotEinsteinSettings be enabled in all scratch orgs?
Enabling PardotEinsteinSettings in scratch orgs requires Account Engagement feature licenses and Pardot Einstein entitlements on the Dev Hub.
Comments
Questions, corrections and tips are welcome. Comments are GitHub Discussions, so you sign in with GitHub.