Considerations for Einstein Account Management

When you set up Einstein Account Management, you must have certain settings configured in your org, and you must have certain objects to populate the dashboards correctly.

• Einstein Account Management is powered by CRM Analytics, so you must enable CRM Analytics.
• Einstein Account Management requires you to have Customer Insights version 1.5 or higher enabled and configured. By default, the auto-installed version is prioritized during setup.
• If multiple versions of Customer Insights have been created, Einstein Account Management uses the most recently installed compatible version.
• Einstein Account Management uses the same security predicate as your Customer Insights app. If you selected a custom security predicate from the wizard, it inherits this selection
• Einstein Account Management uses the same field as your Customer Insights app to track the total amount of the opportunity. By default the amount field is used.
• To create the default Einstein Discovery Account Health model, you must have 400 unique accounts that experienced a change in win rate between two specified time periods: 12-6 months ago, and 6 months ago to present. If you don’t yet have enough data, the Einstein Discovery models can’t be created initially. In this case, the app is created without Einstein Discovery models. You can go back later and recreate the app when you later have enough data. For more information about Einstein Discovery specified limits, see Einstein Discovery Limits.
• To create the default Einstein Discovery Account Upside model, you need at least 400 unique accounts that experienced a change in product count or quantity between two specified time periods. For more information about Einstein Discovery specified limits, see Einstein Discovery Limits.
• Salesforce recommends that you schedule the Einstein Account Management recipe to run on a regular cadence that is scheduled within 24 hours of your Customer Insights recipes.

Note these requirements for using Einstein Account Management:

Note When you turn on CRM Analytics for your org, a default Analytics Integration User is created. This "user" accesses the data within Salesforce when the recipe is used to bring data in as datasets to CRM Analytics. It's important that this user has the proper level of permissions to access your data. For example, the integration user needs have object level access to all the objects from which data you want to pull, and field level access to the fields within those objects. If you have trouble in creating your app or can’t view data, you sometimes must modify either object or field permissions to allow the integration user to access them.

Recipes, ED Models, Components, and Datasets

The Einstein Account Management Accounts Recipe is used in Einstein Account Management:

• This recipe pulls in Account information and adds multiple aggregated metrics across Opportunities, Tasks, Events, Products, and LineItems.

The Account Health and Account Upside ED Models are created. In addition, the Account Inspector with Predictions component is created. Einstein Account Management uses all the datasets created by Customer Insights. In addition, the Einstein Account Management - Accounts dataset is created:

Dataset Name	Contents	Source	Used by	Security Predicate	Special Requirements
Einstein Account Management- Accounts	Aggregated data about accounts	Salesforce object -> Recipe	All dashboards	The security predicate is inherited from Customer Insights. By default the predicate is: 'Owner.Role.Roles' == "$User.UserRoleId" || 'OwnerId' == "$User.Id"	None.

Note Because Einstein Account Management uses security predicates in its models, you must set it up from the Setup menu rather than Analytics Studio. This best practice ensures that all records in the dataset are used when training the models. Running setup from Analytics Studio can result in inaccurate or failed models