To diagnose issues with API-generated tickets, you can create a view or use search functions in Zendesk.
Check if all tickets from a specific channel are affected or only a subset. Use views to list API-generated tickets and add columns for intent and sentiment. Alternatively, use the 'via:api' search to identify problematic tickets and determine if a common trait can be used to filter them.
If your ticket's intent or sentiment prediction seems off, it might be due to language support or non-human content. Intelligent triage predictions can be inaccurate if the ticket's language isn't supported or if non-human content is added by…
If a ticket's language isn't supported, it can lead to incorrect intent or sentiment predictions. To handle unsupported languages, you can create a trigger to remove the incorrect sentiment and intent values before your business rules process the…
Creating a trigger to remove incorrect predictions involves setting specific conditions and actions in Zendesk's Admin Center. First, navigate to Business rules > Triggers and add a new trigger. Set conditions to identify new tickets without…
Non-human content can confuse intelligent triage because it's trained on human-submitted information. Integrations that add metadata or system-generated messages can lead to unexpected predictions. If you notice consistent issues, consider…
If only some API tickets have incorrect predictions, identify a common trait to filter them. Look for consistent tags, ticket fields, or text in the subject line. Once identified, create a trigger to remove incorrect sentiment and intent values for…