What is PrediCX?
PrediCX is an AI classifying machine that allows you to automatically tag your data with meaningful labels. It’s based on a combination of Machine Learning and a ‘human-in-the-loop’ labelling incoming data, minimizing the human effort and maximizing the accuracy and the scope of the process.
Before getting into its many uses, we'll show you how to integrate PrediCX into your Zendesk account.
How to set up within Zendesk?
In your Zendesk Admin bar, there’s an option for Zendesk Marketplace (Admin > App > Marketplace). This is a resource to look for partners and apps to improve your Zendesk experience.
You can use the search bar to find PrediCX, and then follow the next steps:
1. Install the PrediCX app on your account.
2. Click on the PrediCX icon on the Navigation Bar to access the Settings page.
3. Click on the PrediCX Credentials tab to provide the URL and the application key to be used in the classification requests.
4. Click on the Classification Rules tab and on the New Rule button to start adding rules and classifying your tickets.
And that’s it, you’re set to go!
How to keep customers satisfied
Here’s how to use this tool to help you improve your workflow and customer experience:
When the Zendesk / PrediCX integration is completed, the tags created from PrediCX will be automatically applied to any ticket created or updated on Zendesk. They can then be used to create ticket views and conditions to build Triggers that best respond to your personal needs.
As an example, let’s imagine a customer service contact center from “MyBank”.
Customers want to get solutions anytime of the day, but their customer service isn’t open 24 hours.
Since they don’t want to neglect those incoming tickets, they set up a trigger which conditions are based on the label applied by PrediCX, and which action will be to send an automatic reply containing a self-serve article to the client.
This is how it would look like:
How to enhance and maintain your agents’ performance
Another use you can give to those same labels, e.g. during the customer services working hours, is to employ them to build Triggers that will suggest pre-set Macros for agents to respond to certain types of common queries and complaints. This way the agents become more efficient and the AHT reduces significantly.
Your Macro for the same problem would look like this:
How to improve your Workflow
PrediCX can also be used to improve your workflow and make prioritisation more efficient, saving time and ensuring the right queries and tickets are in the right way by the right people.
Automatic labelling will allow your customer service representatives to:
Get early warnings of issues.
Fast track any complaints or urgent enquiries.
Allocate each ticket to the team that will best deal with the problem.
For example, they will be able to create Triggers to escalate the ticket to a manager when it meets the conditions of problematic labels such as “Churn” or “Several interactions needed” and “Problem not solved”. This way they can act swiftly prioritizing outbound calls and customer retention strategies.
A trigger like that might look like this:
Since customers who threaten to leave tend to have previously endured a string of bad experiences with the company, it would be advisable to submit an instant automatic message to let them know that their problem is taken into account and that they will receive a call in the next 2 hours.
The same trigger that sends that reply to the customer can also route this ticket to a special group in charge responding to urgent complaints.
This trigger would look something like this:
Another possible utilization would be to swiftly notify the software development team about the presence of app or web glitches, or also about new software needs expressed by their users to grasp new product opportunities.
There are several different ways to use automatic labels to create triggers that will route requests to the appropriate department that can better deal with the issue.
How to bring more insight to your Zendesk Views
Views are the way to organize your tickets by grouping them into lists based on a certain criteria.
There’s a number of pre-set Zendesk Views, and these can be modified and extended to your own needs. This flexibility, enables you to use the PrediCX labels of your choice to create more meaningful views that will help you determine what tickets need attention from you or your team and plan accordingly.
So, going back to the original example, a Views menu from the App & Web support team could look like this:
By now hopefully you can imagine many other scenarios where PrediCX would help you streamline your workflows, prioritize and arrange your tickets and overall enhance your customer service center performance.