This blog talks about how predictive models built in the Home org of a DC1 setup using the No-code Model builder is surfaced across all of its companion orgs basis data access
Basics of Data Cloud One
Salesforce Data Cloud One is a Data Cloud feature which facilitates businesses to seamlessly manage multi-org environments. In many organizations, data is spread across different Salesforce instances (or “orgs”), often due to mergers, acquisitions, or departmental silos. This fragmented setup can lead to inconsistent insights, missed opportunities for cross-functional collaboration, and inefficiencies in leveraging the full power of data. DC1 addresses these challenges by providing a unified platform that enables smooth data integration and collaboration across multiple orgs.
Note: To get more details on Data Cloud One, take a look at this DC1 help article.
Setups of DC-orgs
Multi-org setups in DC1 can be based on a single or a combination of factors like brands / regions / teams. Consider a business with presence in multiple countries and selling multiple brands.
Configurations can be based on:
- brand, each brand being one DC1 org, or
- country, each country being one DC1 org or
- a combination of factors and so on.
Einstein Studio in context of Multiple-orgs via DC1
Einstein Studio in Salesforce Data Cloud empowers business users to create / bring predictive models: It also allows to create retrievers.
No-code predictive models
The No-Code Model Builder offers the following advantages:
- It streamlines the entire process, from data preparation to model deployment, enabling organizations to unlock the power of AI quickly and efficiently.
- It democratizes the use of predictive AI, making it accessible to a broader range of teams across the organization, accelerating time-to-value, and driving data-driven decisions.
Note: To get started on Model Builder, go through our 101 blog Build an AI Model With Clicks In Data Cloud and more here.
Bring your Own Models (BYOM) models
The BYOM option allows users to bring and reuse custom built models from Amazon Sagemaker, Google Vertex AI and Databricks models into Data cloud.
Note: To get started on BYOM, read this help article.
Use cases on Predictive Models (Einstein-created / BYOM) in DC1 Orgs
Some use cases of predictive models that can be created in the Home org and synced to the Companion org.
- Parent E-commerce company builds a Churn prediction model; Brand-wise subsidiar orgs run predictions on demand via Screen Flows in the companion orgs
- A HR / workforce company building a Centralized model for Candidate’s offer acceptance prediction; Regional orgs view predictions as offer acceptance prediction DMO in their own companion orgs
- A CPG / FMCG business build a Order quantity prediction model; Subsidiary orgs view HQ’s models in companion orgs and run flows to write back to the Orders object.Predicted Quantity field
Benefits
- Unifies intelligence without duplicating work. Centralized domain expertise can have further reach.
- Scales predictive AI across your entire ecosystem.
- Promotes governance at the central home org
- Provides flexibility in using predictions
- Faster time to value
How Model Builder works in DC1 orgs?
Now let’s get to the setup of how Model Builder fits into the DC1 architecture.
TDLR: DC1 support in Model builder essentially means “Centralized model building in the Home org” and “Distribution and visibility of model, predictions across the companion orgs”.
The following are the capabilities of the Model Builder setup in DC1 orgs:
- ✅ Models can be built and trained in the Home org (similar to a single-org setup)
- ✅ Models can be activated in Home org only (since it’s the source org)
- Prediction Usage path 1: PERSISTED PREDICTIONS via DMOs
- ✅ Home org users can generate and persist Predictions into DMOs via Transforms / Predict jobs
- ✅ Predictions available in DMOs are auto-synced to Companion orgs via DC1 Connections.
- ✅ Home org users can use predictions as criteria / filters in Segments to group based on scores.
- ✅ Companion org users can view and use these scores as filters in Segments (Coming soon)
- Prediction Usage path 2: ON-DEMAND PREDICTIONS
- ✅ Companion org users can view the travelled model, available in read-only mode.
- ✅ Companion org users can make on-demand predictions by invoking the travelled models in Flows.
- To create automated workflows based on scores
- To display Predictions on object pages
- 🚫 Companion org users will not be able to create models, or activate a model or view training metrics (coming soon).
Models syncing from Home org to Companion org
The Einstein Studio tab in the Home org displays all the models created in that org.
When these models travel to companion org, this is how the Einstein Studio tab looks like. They are displayed in a read-only mode, where you can view some metadata about the model.
You can go ahead and use these models in Data cloud Flows, or via APIs to make inferences
Similarly, LLMs models configured in Home org are synced to the companion org and available for using to create a Prompt template and seen with a read-only icon similar to Predictive models.
You can use this travelled LLM model to create a Prompt template.
Conclusion
In this blog, we saw how models can be centrally built in the Home org of a DC1 setup. And in the companion orgs the user has the flexibility to use predictions into a DMO from the home org or make on-demand predictions.
Special thanks to Bobby Brill for his review.
For more resources on Model Builder:
- Blogs:
- Salesforce Help articles:
- YouTube:
- Podcast: What are the key features of Salesforce’s Model Builder?
Written by
