Skip to main content

         This documentation site is for previous versions. Visit our new documentation site for current releases.      

Deploying text predictions to the Text Analytics service

Updated on July 5, 2022
Note: This content applies only to environments with a Pega Customer Service application, for which the Voice AI feature is enabled. Pega Customer Service can process natural language in embedded mode or by using the text-as-a-service mode.

When using Voice AI in service mode, the system runs text predictions in the Text Analytics service, which is not in your application environment. You cannot run any new text predictions that you build in Prediction Studio before they are deployed to the Text Analytics service, either manually or automatically by using a job scheduler.

For any updates to natural language processing (NLP) models, all associated predictions for a given model must be deployed. Otherwise, the predictions will use the old versions of the models. Any changes to predictions, such as adding and removing models and changing the prediction settings, also require a deployment of the prediction.

After you complete changes to models and predictions, you can choose to deploy all predictions or go to the Predictions work area, check whether any predictions require deployment, and then deploy them individually.

Before you begin: Ensure that the pyAutoDeployTextPredictions job scheduler is enabled in your system and that it is configured with an access group mapped to the current application.
Note: If multiple applications use text predictions, configure the job scheduler with an access group for each application.
  1. In the navigation pane of Prediction Studio, click Predictions.
  2. Perform one of the following actions:
    • To deploy predictions in bulk, in the header of the Predictions workspace, click ActionsDeploy all stale predictions.
    • To deploy a single prediction, in the Predictions workspace, to the right of the prediction that you want to deploy, click the More icon, and then click Deploy.
      Note: You can also deploy a prediction by opening the prediction, and then by clicking Deploy now in the prediction header.
    Deployment options for text predictions in Prediction Studio
    Predictions workspace in Prediction Studio contains two predictions that require deployment.
    Result: When the prediction is deployed successfully, the status of the prediction changes to DEPLOYED. If the deployment fails, the status changes to DEPLOYMENT FAILED.
  3. If deployment fails, perform the following actions:
    1. In the navigation pane of Prediction Studio, click SettingsPrediction Studio settings.
    2. In the Text service settings section, click Test connection.
    3. If the test fails, contact your system administrator to resolve the issue, and then redeploy the prediction as in step 2.
    4. If the test is successful, redeploy the prediction as in step 2.

Have a question? Get answers now.

Visit the Support Center to ask questions, engage in discussions, share ideas, and help others.

Did you find this content helpful?

Want to help us improve this content?

We'd prefer it if you saw us at our best. is not optimized for Internet Explorer. For the optimal experience, please use:

Close Deprecation Notice
Contact us