Loading
Salesforce now sends email only from verified domains. Read More
AI Accelerator and Scoring Framework
Table of Contents
Filter by (0)Add
Select Filters

          No results
          No results
          Here are some search tips

          Check the spelling of your keywords.
          Use more general search terms.
          Select fewer filters to broaden your search.

            Search all of Salesforce Help
            Search all of Salesforce Help
            Show AI Accelerator Predictions and NBA Recommendations on Lightning Experience Record Pages

            You are here:

            1. Salesforce Help
            2. Docs
            3. AI Accelerator and Scoring Framework

            Show AI Accelerator Predictions and NBA Recommendations on Lightning Experience Record Pages

            Use predictions that the machine learning model generates for your use case to get Einstein Next Best Action (NBA) recommendations. Show these predictions and NBA recommendations in one view on a Lightning Experience record page by adding the AI Accelerator—Einstein Predictions & Recommendations component to the page.

            Required Editions

            Available in: Lightning Experience
            User Permissions Needed
            To create and save pages in Lightning App Builder: Customize Applicatiom

            Before you add the AI Accelerator—Einstein Predictions & Recommendations component to a page, find the feature data type specified on the AI Accelerator Setup page.

            1. Open a record page of the primary response object selected in the AI Accelerator Setup page.
            2. Click Setup icon, and then select Edit Page.
            3. In Lightning App Builder, drag the AI Accelerator—Einstein Predictions & Recommendations component onto your preferred location on the page.
            4. In the component’s configuration panel, select a card.
              You can show either predictions or NBA recommendations, or both.
            5. If you select Einstein Predictions & NBA Recommendations or Einstein Predictions as the card to show, perform these steps in the Einstein Predictions section of the component’s configuration panel:
              1. In the Use Case Definition field, select the use case whose predictions you want to embed in the page.
              2. In the Prediction field, select the model that you configured for your use case in the AI Accelerator Setup page.
                If you configured multiple models for your use case, select the model that best suits your requirements.
              3. For the model to use features available in the specified record ID, select Disable feature extraction.
              4. If you want to use computed features, select the ID of a feature extractor.
                Important
                Important To use the default feature extractor or the feature extractor in the use case configuration file, leave the Feature Extractor field blank.
              5. Select a feature data type.
                • If the feature data type on the AI Accelerator Setup page is Real-Time Input or Batch Input, select Actual Feature Data.
                • If the feature data type on the AI Accelerator Setup page is Sample Input, select Sample Feature Data.
                Note
                Note If Disable feature extraction is selected, the feature data type isn’t considered.
              6. Configure these settings as necessary.
                Property Description
                Show Prediction Label Shows the prediction label.
                Title A descriptive label for the prediction.
                Show title Shows the title of the prediction.
                Prediction score unit

                A unit of measure for the prediction. For example:

                • Currency: For money-based predictions.
                • Percentage: For probability-based predictions.
                • Time interval: For time-based predictions, such as hours, days, and weeks.
                Unit precedes score Shows units preceding the prediction, such as currency symbols. By default, units follow the prediction.
                Positive prediction label A label for the prediction when the prediction is higher than the threshold in a binary classification model. For example, Win, Retain, and Success.
                Negative prediction label A label to show when the prediction is lower than the threshold in a binary classification model. For example, Loss, Churn, and Fail.
                Show top predictors Shows the predictors with the highest impact on the prediction.
                Collapse details Shows top predictors but hides other details.
                Show ranked probabilities Shows ranked probabilities for outcome values. Applies only to predictions associated with multiclass classification models.
                Include link to model card Shows the Learn about this model link that users can click to see the model card associated with this prediction. For more information, see Configure the Model Card for a Prediction Definition.
                Show improvements Shows Einstein Discovery suggestions on how to improve the predicted outcome. Applies only to predictions associated with binary classification or numeric models.
                Number of improvements to show The maximum number of improvements returned on the prediction card. Applies only to predictions associated with binary classification or numeric models.
                Improvement threshold percentage A number to show only improvements that impact the predicted outcome by this percentage or higher. This number applies only to predictions associated with binary classification or numeric models.
                Show values for predictors Shows the impact value next to predictors and improvements.
                Show prediction warnings

                Shows user warnings. User warnings include:

                • Missing fields that are required for the model
                • Field values that are outside the valid range that's used to build the predictive model
                Calculate prediction as a date Shows time-based predictions as based on a relative start date and an offset value (the original prediction result).
                Start calculating date prediction from The date from when you want to start generating predictions.
                Time period

                The unit for the offset value—original prediction result—that's used to calculate the predicted date:

                • Minutes
                • Hours
                • Days
                • Weeks
                • Months
                Predicted date format

                The predicted date format:

                • Time Remaining: Formats the prediction according to the selected period.
                • Specific Date: Formats the prediction as a date according to the logged-in user’s locale settings.
                Show prediction history starting from

                A reference point to show score history, such as:

                • Last change
                • Last week
                Set Component Visibility Component visibility rules for a device, record field, or other field.
            6. If you select Einstein Predictions & NBA Recommendations or NBA Recommendations as the card to show, perform these steps in the NBA Recommendations section of the component’s configuration panel:
              1. In the Title field, enter a title for the recommendations card.
              2. In the Strategy Source field, select the source of the strategy that’s used to build the strategy.
              3. In the Action Strategy field, select the strategy that you want to show.
                The Action Strategy value is required to generate recommendations.
              4. In the Maximum Recommendations Displayed field, enter the maximum number of recommendations to show.
                You can display up to 4 recommendations.
              5. Configure these settings as necessary.
                • Hide Empty Component: Shows the recommendations component only when recommendations exist.
                • Launch Recommended Action In: Specify whether recommendations open in a display window or a new browser window.
                • Show recommendation title: Shows the title of the recommendation.
                • Show Image: Shows images associated with each shown recommendation. If no images exist, a placeholder is shown.
                • Show Description: Shows each recommendation’s description.
                • Show Reject Option: Shows the reject option.
                • Launch Flow on Rejection: Launches a flow when a user rejects a recommendation. The flow then performs an action, such as updating a case or sending an email.
                • Set Component Visibility: Adds filter conditions and logic to the component properties to enable Dynamic Lightning Pages.
            7. Save your work, and then activate the page.

            Users can view the predictions and NBA recommendations for the use case on the Lightning Experience record page in real time.

             
            Loading
            Salesforce Help | Article