Zendesk – NLP processing of ticket content

How can Zendesk ticket content be processed via NLP (natural language processing) to perform actions on tickets?

Example of ticket-mood detection

We will parse the following ticket comment of a Zendesk ticket through NLP (natural language processing):

Hello!
Every time I order in your shop I have following problems. I can’t understand why my experience is so awful!
The parcel is delivered too late – it is damaged every second time and the content seems to be used by some other person already.
When I try to return it you tell me It’s not possible!??

Which information can be retrieved using NLP in Zendesk tickets?

  • Sentiment analysis – retrieves the mood of a text.
    It analyses the text and counts the number of sentiments like positive, negative, neutral.
  • The above example has negative mood
  • The mood can be stored inside a custom ticket field called “mood”

Zendesk NLP processing example for detecting ticket mood
automatically processed Zendesk ticket content through NLP

Solution

When a new tickets gets created the content will be processed through the Leafworks Zendesk AddOn Api Middleware.
The mood will be stored in a custom ticket field called “ticket-mood“.
You can then make use of Zendesk triggers to prioritize incoming tickets based on the mood of the customers. Happy customers will be fine when receiving a response later – un-happy customers could get the response faster!

zendes-parse-subject-via-middleware
processing through the Zendesk AddOns API Middleware