Table of Contents

Process emails in multiple languages in an email bot (8.4)

To ensure that your Pega Platform™ application can handle user emails in different languages, set up multiple language support in Pega Email Bot™. Configure your email bot to detect topics and entities from emails in additional languages, so that the system suggests the correct business case and then provides an email response in the user's own language. By training the machine learning models in different languages for a single Pega Platform application, you increase efficiency by increasing the number of automatically triaged emails. For more information about the topics and entities detected by email bots using text analysis, see Exploring text analyzers.

You specify the languages that the email bot supports during Email channel configuration so that the system knows when to perform text analysis and respond in a different language. The following figure shows where you configure language support for an Email channel:

Email channel interface - Configuration tab
Email channel interface - Configuration tab
Email channel interface - Configuration tab

The languages supported for email text analysis and processing include English, Dutch, French, German, and Spanish, in any combination. The default language for the email bot is English but you can override this setting by editing the pyGetDefaultLanguage data transform rule for your application. For example, you can configure text analysis of topics and set up suggested email reply templates individually for both English and German users of the same application. For a complete list of additional languages used with for the email bot, see Text analytics features in Pega Platform. To set up multiple-language text analysis for suggested cases in an Email channel, you use the Add suggested case window:

Add suggested case window - Text analysis tab
Add suggested case window - Text analysis tab
Add suggested case window - Text analysis tab

When training an email bot with data, you can filter training data records by a language with one click. In addition, you can now choose the language in which to create a new training record. When reviewing training data for a specific language, you can also map the text to any of the existing entities in the system, or create a new entity. After you apply your changes to the text analytics model using the reviewed training data, the system builds models in all of the languages that you defined. If you upgrade from a Pega Platform version prior to 8.4, as a one-time action, you use the Map languages window to select existing reply templates and map them to specific languages:

Map existing templates for upgrade compatibility
Map languages window used for upgrade compatibility
Map existing templates for upgrade compatibility

For more information, see Selecting languages for an Email channel, Overriding the default language for an Email channel, and Enabling automatic language detection for text analysis.

Suggest Edit

Have a question? Get answers now.

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