Feedback Edit

Complete your partial data with predictions (deprecated)

  • Article
  • 5 minutes to read

Important

This feature will be deprecated as of November 5, 2021. Current implementations will continue to work until the feature is removed, but you will not be able to create new integrations using the instructions below.

Predictions lets you easily create predicted values that can enhance your understanding of a customer. On the Intelligence > Predictions page, you can select My predictions to see predictions that you've configured in other parts of Customer Insights, and allow you to further customize them.

Note

You can't use this feature if your environment uses Azure Data Lake Gen 2 storage.

The predictions feature uses automated means to evaluate data and make predictions based on that data, and therefore has the capability to be used as a method of profiling, as that term is defined by the General Data Protection Regulation ("GDPR"). Customer's use of this feature to process data may be subject to GDPR or other laws or regulations. You are responsible for ensuring that your use of Dynamics 365 Customer Insights, including predictions, complies with all applicable laws and regulations, including laws related to privacy, personal data, biometric data, data protection, and confidentiality of communications.

Prerequisites

Before your organization can use the predictions feature, the following prerequisites must be met:

  1. Your organization has an instance set up in Microsoft Dataverse and it's in the same organization as Customer Insights.

  2. Your Customer Insights environment is attached to your Dataverse instance.

For more information, see Create a new environment.

Create a prediction in the Customer entity

  1. Go to Data > Entities.

  2. Select the Customer entity.

  3. In the Customer:CustomerInsights entity, select on the Fields tab.

  4. Find the attribute name you wish to predict values for, then select the Overview icon in the Summary column.

    Overview icon.

  5. If there's a high rate of missing values for your attribute, select Predict missing values to continue with your prediction.

    Overview status with predict missing values button shown.

  6. Provide a Display name and an Output entity name for the results of the prediction.

  7. A pre-populated list of options will show where you can map the values to a predicted category. In this case, your only category options will be 0 or 1, as they map to the true/false or binary nature of the prediction. In the Category column, map the field values you'd like to be classified as "0" in the final prediction to "0", and the items you'd like to be classified as "1" in the final prediction to "1".

    Example showing mapped field values to categories.

  8. Select Done and the prediction will be processed. The processing will take some time, depending on the size and complexity of data. Results will be available in a new entity based on the Output entity name of the prediction you created.

Tip

There are statuses for tasks and processes. Most processes depend on other upstream processes, such as data sources and data profiling refreshes.

Select the status to open the Progress details pane and view the progress of the tasks. To cancel the job, select Cancel job at the bottom of the pane.

Under each task, you can select See details for more progress information, such as processing time, the last processing date, and any applicable errors and warnings associated with the task or process. Select the View system status at the bottom of the panel to see other processes in the system.

Create a prediction while creating a segment

Predicting missing values for a specific attribute of choice is also possible when creating a segment. Specifically, when you quickly create a segment based on either your unified Customer entity or Customer_Measure entity.

As part of this flow, you'll choose a specific attribute to base your segment on, such as Customer Satisfaction or Purchase Amount. Upon segment creation, the system will suggest a method for predicting any missing values for this attribute.

  1. Go to Segments and select the Profiles tile.

  2. Choose a Field to create a segment on and select an Operator, then select Review.

  3. Provide a Name and a Display name for the segment.

  4. Select Save.

  5. If the segment you created has incomplete data in the source field, you can choose to predict the missing values.

    Prediction button.

  6. Provide a Display name and an Output entity name for the results of the prediction.

  7. Select Done. Your prediction will be generated shortly in a new entity with the name you provided for the Output entity name.

View a prediction

  1. Go to Intelligence > Predictions > My predictions.

  2. Select the prediction you want to review.

  3. Select the vertical ellipsis (⋮) in the Actions column and choose View.

  4. You'll see a number of data points in the view of your prediction.

    Predictions page.

    • Predicted values shows the mapping you created during the Field value to Category mapping phase. These are the values in your dataset that have been mapped to a specific category. -Top influencers are the factors within your dataset that were most likely to influence the prediction's confidence of your Field value being mapped to a specific category.
    • Performance indicates how the predictions are doing. Select the link to learn more.
    • Preview shows samples of the output dataset from your prediction and the likelihood, or our confidence, of the predicted value where 0 is uncertain, and 1 is certain.

Update a prediction

  1. Go to Intelligence > Predictions > My predictions.

  2. Select the prediction you want to update and select the Update icon.

  3. The prediction will be scheduled for processing. You can see the time it was last updated in the Updated column of the Predictions page.

Edit a prediction

After you've created a prediction, you can customize the model in the AI Builder to increase the effectiveness of your model.

  1. Go to Intelligence > Predictions > My predictions.

  2. Select the prediction you want to edit.

  3. Select the vertical ellipsis (⋮) in the Actions column and choose View.

  4. Select Customize in AI Builder.

  5. Update your model in the AI Builder. Learn more about managing models in the AI builder.

The next run of your prediction will use the updated model you've created.

Note

New models created in AI Builder will not be displayed in Customer Insights unless the model was created from the experiences listed above.

Remove a prediction

  1. Go to Intelligence > Predictions > My predictions.

  2. Select the prediction you want to delete.

  3. Select the vertical ellipsis (⋮) in the Actions column and choose Delete.

  4. Confirm the deletion.

Troubleshooting

If you can't complete the attach Dataverse process due to an error, you can try to complete the process manually. There are two known issues that can occur in the attach process:

  • The Customer Card Add-in solution is not installed.

    1. Complete the instructions to install and configure the solution.

  • App permissions aren't granted.

    1. Go to https://admin.powerplatform.microsoft.com.
    2. Select Environments.
    3. Select the vertical ellipsis (⋮) next to the environment you want to add the permission to and select Settings.
    4. Expand Users + permissions and select Users.
    5. Select + New and select User.
    6. Select Application User if it's not already selected and enter the following information:
      • User Name: cihelp@microsoft.com
      • Application ID: 38c77d00-5fcb-4cce-9d93-af4738258e3c
      • First Name: Customer
      • Last Name: Insights
      • Primary Email: cihelp@microsoft.com
    7. Select Save & Close.
    8. Select the user you just created.
    9. Select Manage Roles in the top menu bar.
    10. Select System Administrator, then select OK.