Use the customizable Interaction API for text analytics with your email bot to set up routing conditions based on a decisioning result, to determine the outcome of the interaction strategy. The pyInteractionDF data flow for the Interaction API performs natural language processing through the iNLP text analyzer for the email bot and feeds the outcome to the interaction strategy. The system captures the strategy outcome in the pyDecisionResult property for the Data-pxStrategyResult class. For more information, see Customizable Interaction API for text analytics.

For relevant training materials and to test this feature on a virtual Pega instance, see the Routing email based on content module and the Routing email based on content challenge on Pega Academy.

1. In the navigation pane of App Studio, click Channels.
2. In the Current channel interfaces section, click the icon that represents your existing Email channel.
3. In the Email channel, click the Behavior tab.
4. In the Intelligent routing section, in the Process list, select a processing mode:
   • To trigger only the first action in the list that matches the conditions, select First matching action.

     All other actions in the list with matching conditions are not triggered.

   • To trigger every create case and automatic reply action in the list that matches the conditions, and skip all route to actions, select All matching actions.

     If no create case or automatic reply action in the list match the conditions, the system triggers only the first route to action that matches the conditions.

   Selecting the routing processing mode enables the system to trigger only the first action or every create case and automatic reply action in the list with matching conditions.
5. In the Intelligent routing section, click Add condition.
6. In the Action list, select an action that you want to run automatically for email routing.
   For example: Create case
7. In the Value field, enter an email template, case name, operator, or work queue.
   For example: For the Create case action, enter a case type name: Insurance Claim
8. Define the validation conditions for a field on the form as the criteria for carrying out the routing action:

   Choices	Actions
   Detect email header content	In the When list, select Header and the name of the email field that you want to validate, for example: Subject. In the comparator list, select the test that you want to perform on the field, for example: Contains. In the value field, enter or select a value to compare against the user input, for example: price.
   Detect an entity - a word or phrase in an email	In the When list, select NLPEntity, and an entity name, for example: #Email. In the comparator list, select the test that you want to perform on the field, for example: Is equal. In the value field, enter a value to compare against the user input, for example: [email protected].
   Detect the topic - the subject matter of an email	In the When list, select NLPTopic. In the comparator list, select the test that you want to perform on the field, for example: Is equal. In the value field, enter a value to compare against the user input, for example: Insurance Quote.
   Detect a decisioning result - the outcome of the interaction strategy	In the When list, select NLPDecisioning result. In the comparator list, select the test that you want to perform on the field, for example: Is equal. In the value field, enter a value to compare against the user input. The email bot can check whether the decisioning result is empty, is not empty, or contains a specified value as the outcome of the interaction strategy, for example, a postal code, or a company or product name. For more information, see Customizable Interaction API for text analytics.
   Detect the positive, negative, or neutral sentiment of an email	In the When list, select NLPSentiment. In the comparator list, select the test that you want to perform on the field, for example: Is equal. In the value list, select what sentiment to compare against the user input, for example: Positive.
   Detect the language of an email	In the When list, select NLPLanguage. In the comparator list, select the test that you want to perform on the field, for example: Is equal. In the value list, select a language to compare against the user input, for example: French.
   Use a when rule	In the When list, select Condition. In the field to the right of the When list, enter or select a when rule that you want to use as the condition for carrying out the action. For example, you can configure a routing condition to route the email to an operator Andrew Sullivan when the pyIsInWorkList rule is true.

   Each routing condition specifies a single way that the email bot automatically responds to an email.
9. Optional: To add more automatic responses based on email header content, NLP text analysis of the email, or a when rule, repeat step 5 through 8 for other routing conditions.
10. Specify how your email bot responds when you do not define any criteria or when the email content does not meet any of the criteria that you specified in step 8 by defining the default action:
    1. In the Otherwise area, in the Action list, click an action to perform.
    2. In the Value field, enter an email template, case name, operator name, or work queue.
11. Click Save.
    For example:

    

    

12. Verify that the intelligent email routing works as expected.
    For more information, see Testing email routing conditions.