I created a view using Zendesk Copilot's sentiment analysis
Hello, I'm Subaru.
This time, I'll focus on utilizing Copilot, an add-on for Zendesk.
Introduction
Zendesk has an AI feature called Copilot as an add-on. This feature provides various functions in addition to normal Zendesk capabilities, such as auto-assist using Zendesk AI, initial response suggestions, and ticket integration recommendations. In this article, I'd like to focus on utilizing sentiment analysis.
Setup
Copilot
First, enable the sentiment analysis feature. Go to Admin Center > AI > Intelligent Triage > Sentiment, and confirm that "Detect Sentiment" is enabled. Also, sentiment detection can be selected for each channel, so choose which channels you want to detect.
When sentiment analysis is enabled, it detects "Very Positive," "Positive," "Neutral," "Negative," or "Very Negative" from the ticket content. For Talk, sentiment can also be detected from transcribed content.
When checking a ticket with detected sentiment, you can see the sentiment results displayed along with the purpose and overview.
Views
Now we'll create views based on sentiment detection results. Normally, "Sentiment" doesn't appear as an option when setting view conditions. However, by using tags, we can set conditions based on sentiment detection results. Looking at the ticket earlier, we can confirm that tags are automatically added according to the sentiment detection results.
Go to Admin Center > Workspace > Agent Tools > Views, and add a new view. Add a condition that includes tags for sentiment detection results. (For Neutral, you would add the tag sentiment__neutral)
Verification
Let's check the results. When examining tickets assigned to the created view, we can confirm that items with neutral detection results are properly sorted.
Summary
In this article, I created views utilizing sentiment analysis, one of Copilot's features. While I focused on creating views, I believe the same method can be used with automations, triggers, and other features.
Zendesk AI features continue to emerge, so let's keep utilizing them in the future.
I hope this blog was helpful.
