LinkedIn
Copied!

Table of Contents

Troubleshooting email routing conditions

Version:

Only available versions of this content are shown in the dropdown

During testing of the routing conditions that you define for intelligent routing, Pega Email Bot™ might not always generate the correct outcome. To solve this issue, there are several guidelines that you can follow while testing your routing conditions in the system. Troubleshooting issues and verifying all of the expected actions that trigger during intelligent routing ensures that the email bot works as expected and improves the availability of the system in a production environment.

  1. Check the email bot settings for natural language processing (NLP):

    1. Configure one or more advanced text analyzers in the system, for example, the iNLP text analyzer. For more information, see Adding a text analyzer for an email bot.

    2. Define the topics (the subject matter of emails), that the system will detect during intelligent routing. For more information, see Defining topics for text analysis for an email bot.

    3. If you want the email bot to also analyze attachments during intelligent routing, enable email attachments analysis in the system. For more information, see Enabling email attachments analysis during email triage.

    4. If you want the email bot to support multiple languages, enable automatic language detection during intelligent routing. For more information, see Enabling automatic language detection for text analysis.

  2. Verify the expected email bot behavior for each routing condition:

    1. Add all suggested cases to start in the system as a result of intelligent routing. For more information, see Adding suggested cases for an Email channel.

    2. If applicable, ensure that the system can route emails to the correct operator or work queue.

  3. Continue to train the text analytics model for the email bot:

    1. Create enough instances of sample training data to trigger the action for each defined routing condition. For more information, see Creating training data manually for an email bot.

    2. Modify the training data to detect the correct topic, sentiment, language, and entities that are used in the routing conditions. For more information, see Correcting training data in an email bot.

    3. After you rebuild the model, test the routing conditions once again until they trigger the correct action. Repeat this process until you have verified the correct outcomes for each of the routing conditions. For more information, see Testing email routing conditions.

Did you find this content helpful?

Have a question? Get answers now.

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