Published Release Notes

Uploading customer responses into adaptive models is no longer available

The option to train adaptive models by uploading a static list of historical interaction records has been deprecated in Pega Platform™ 8.5.

Upgrade impact

In versions of Pega Platform earlier than 8.5, it was possible to train an adaptive model by uploading historical data of customer interaction. After the upgrade to version 8.5, this option is no longer available.

What steps are required to update the application to be compatible with this change?

Use data from a report definition to train adaptive models. For more information, see Training adaptive models.

More flexibility for Date Time predictors in adaptive models

You can configure adaptive model predictors that indicate the amount of time that has elapsed since a particular event. In versions of Pega Platform™ earlier than 8.5, you could only specify the absolute date and time, for example, the date when a subscriber last visited one of your brick-and-mortar store locations. In Pega Platform 8.5, you can also indicate the amount of time that has passed, for example, the number of days since a subscriber purchased a new service plan.

For more information, see Adding a predictor to an adaptive model.

Predictive models can drive predictions

You can use predictive models as the basis of your predictions. As a data scientist, you can now use Machine Learning Operations (MLOps) to replace models in your system. You can replace a model in a prediction with a PMML, H2O MOJO, or Pega OXL predictive model, as well as a scorecard or field, and then approve the update for deployment to a production environment. You can respond to a Prediction Studio notification that an active model does not generate enough lift and decide to replace the low-performing model with a high-accuracy model. You can also update a prediction on a regular basis, for example, whenever you develop a new churn model in an external environment.

For more information, see Replace models in predictions and migrate changes to production with MLOps.

Enhancing the performance of your Next-Best-Action strategy with globally optimized strategies

Starting in version 8.6, Pega Platform™ combines the versatility of Next-Best-Action Designer with strategy performance enhancements provided by using globally optimized strategies (GOS). Decrease the run time and memory usage of executing the Next-Best-Action strategy in batch or real-time scenarios by using the globally optimized strategies generated by Next Best Action Designer.

GOS is supported by Pega Platform's standard business change management process. GOS rules are automatically included in relevant revision packages.

For more information, see Enhance the performance of your Next-Best-Action strategy with globally optimized strategies (8.6).

Support for picklists with parameterized data pages in App Studio in Cosmos React UI

You can now use data pages with parameters to populate a property of the picklist type with filtered results in App Studio. For example, in a survey case type, you can use a parameterized data page to configure cascading drop-down controls in which the values in a secondary drop-down list are based on the value that the user selects in the primary drop-down list. With dynamically-sourced picklists, you get greater flexibility in configuring picklists, and users see more accurate values.

For more information, see link Dropdown control Properties — General tab.

Default value of the threadpoolsize agent affects batch indexing

After you patch Pega Platform to version 8.5.2 or higher, the system changes the default value of the threadpoolsize agent, which controls the number of concurrent activities (threads) in the system, from 5 to 15. Batch indexing in Pega Platform™ does not require all 15 threads, so you can change the agent value to increase system performance by managing the indexing/distributed/batch/numworkers dynamic system setting.
If your deployment does not support that setting, and batch indexing does not use Queue Processors, the system uses the threadpoolsize value for batch indexing instead.

For more information, see Editing a dynamic system setting.

Easier customer record management in Customer Profile Designer (early preview)

The new Customer Profile Designer module of Pega Customer Decision Hub™ makes it possible for marketing analysts and strategy designers to define the associated data for each customer context directly in the Pega Customer Decision Hub portal. It is also possible to define more complex associated data structures that use a custom data flow, or define associated data of different types, such as RDBMS and Cassandra, for the same customer context.

Customer Profile Designer is available in Pega Customer Decision Hub 8.6 version as an early preview version. The functionality will be further expanded in future releases.

For more information, see Manage customer records in Customer Profile Designer (8.6).

Low-code authoring of case pages

Mobile authoring for Cosmos React applications now supports the configuration of native mobile pages that include case details. Your mobile channel automatically generates these pages for each case type in your application. This enhancement improves the user experience in your app by providing users with native pages that you can adjust to best suit your business process. For example, you can highlight specific views on a page by modifying which tabs appear on the page, or add a floating action button to help your users edit the case.

For more information, see Authoring mobile case pages for Cosmos React applications.

In Pega Platform™ version 8.6, mobile features in Cosmos React applications are available only as a preview, with limited support for various interface elements and features, such as offline mode. The recommended production design system for Pega Mobile in Pega Platform 8.6 is Theme Cosmos.

For more information about the Cosmos React design system, see Cosmos React and UI architecture comparison.

Text predictions simplify the configuration of text analytics for conversational channels

Enable text analytics for your conversational channels, such as email and chatbot, by configuring text predictions that manage the text models for your channels. This new type of prediction in Prediction Studio consolidates the AI for analyzing the messages in your conversational channels in one place and replaces the text analyzer rule in Dev Studio.

Through text predictions, you can efficiently configure the outcomes that you want to predict by analyzing the text in your channels:

• Topics (ticket booking, subscription cancellation, support request)
• Sentiments (positive, neutral, negative)
• Entities (people, organizations, airport codes)
• Languages

You can train and build the models that predict these outcomes through an intuitive process, and then monitor the outcomes through user-friendly charts.

For more information, see Predict customer needs and behaviors by using text predictions in your conversational channels.

Upgrade impact

Channels that you configured with text analyzers in the previous version of your system continue to work in the same manner after the upgrade to the current version. When you edit and save the configuration of an existing channel, the text analyzer rule is automatically upgraded to a text prediction. The associated text prediction is now an object where you can manage and monitor the text analytics for your channel. When you create a new channel in the upgraded system, the system automatically creates a text prediction for that channel.

What steps are required to update the application to be compatible with this change?

1. Enable the asynchronous model building and reporting in text predictions through job schedulers that use the System Runtime Context (SRC) by adding your application to the SRC.
   For more information, see Automating the runtime context management of background processes.
2. Enable model building in text predictions by configuring background processing nodes.
   For more information, see Assigning decision management node types to Pega Platform nodes.

Automatic packaging of cases for offline use

Newly created mobile channels now support automatic packaging of cases for use in offline mode. The packaging process includes case data from all native list pages that you create from offline-enabled case types and add to the navigation of your mobile app. Previously, automatic packaging was limited to assignments in the work list (and their related work items), and cases that were listed in the data page that you manually designated for packaging.

This enhancement improves the user experience for mobile app developers by ensuring that the data that is referenced in list pages for offline-enabled case types is automatically available for the users of your offline app.