Customer support teams are faced with frustrating and repetitive daily tasks, and one of their biggest concerns is spending too much time on tagging tickets:

• Ticket tagging is tedious work.
• Agents don’t see the value of using tags.
• Every agent tags tickets using their own criteria.

while tags are important to your business, agent time is becoming more and more valubale, so you need to find a way to automate processes that involve ticket tagging.

Enter keyword extraction, an AI tool that can be used to tag and route huge amounts of tickets automatically. before we dive into this keyword extraction, let's take a look at why ticket tagging is important.

Why is tagging tickets important?

Tags allow you and your team to manage and customize your support workflows. Some of the things that you can achieve include:

• Use tags in your business rules (automations, macros, and triggers) to create custom workflows.
• Create views based on tags.
• Generate reports by tags, for example, you can quickly understand support request trends using the top 100 tag cloud that Zendesk provides.

All of this can help your team save up time as well as give managers the tools they need to analyze and improve their processes.

Zendesk already has a solution in place

You can use the automatic tag feature that Zendesk provides. Which “scans new incoming ticket descriptions looking for words longer than two characters and then compares those words to the tags that have already been used. The top three matches are added to the ticket”

Very useful but it comes with its own set of disclaimers:

• It will only work provided you have already tagged tickets before.
• “If you inadvertently add tags for high-frequency words such as 'and' and 'the' these words will generate tags in your tickets”

This means that if you are starting this process from scratch, it won't work, and it will require a big effort from your team to set in place. Last but not least, if you are already constrained by the size of your team, this will create overheads that you should consider.

With MonkeyLearn you can jumpstart this process

A new idea that some of our customers have implemented is starting to use MonkeyLearn's online keyword extractor to automatically add tags to their tickets machine learning. This is becoming a popular way of using MonkeyLearn with Zendesk since it offers a lot of benefits:

• No previously tagged tickets are needed.
• You rely on unified criteria that you can control and improve over time (soon to be added custom extraction to MonkeyLearn).
• You don’t need to worry about the frequency of words (it already does it for you)
• You’ll be using Machine Learning algorithms, not relying on string matching.
• Agents won't deal with anymore with this frustrating and repetitive task.
• Agents will have more free time on their hand to serve your customers better.
• Managers will be able to rely on this tags for reporting and analysis.

Setting up the process:

To do this you can either follow the steps or just use this template that we have prepared. To set this up you will require:

• Zendesk Account.
• Zapier Account.
• MonkeyLearn Account.

The following steps will help you create a zap that will populate your new incoming tickets with its most relevant keywords as tags:

1. Choose Trigger App: Zendesk

1.1. Select Trigger: New Ticket

1.2. Choose your account

1.3. Select the View from Zendesk

1.4. Test the step

2. Choose Action App: MonkeyLearn

2.1. Select Action: Extract Text

2.2. Choose your account

2.3. Select Extractor: Keywords

2.4. Select Text:

2.4.1. Add Subject from Zendesk Trigger

2.4.2. Add Description from Zendesk Trigger

2.5. Test this step

3. Select Action App: Zendesk

3.1. Select Action: Search: Find a Ticket

3.2. Choose your account

3.3. Select Query: New Ticket ID

4. Select Action App: Zendesk

4.1. Select Action: Add Tag(s) to Ticket

4.2. Choose your account

4.3. Select Custom Value for Ticket ID: Step 3 Ticket ID

4.4. Select Tags: Extract Text: Keywords

4.5. Test

5. Turn on your Zap

From now on, each new ticket that comes into the selected view will get tags automatically added.

Final words

In the background is running the MonkeyLearn keyword extraction model, that will return the most relevant keywords of a text. This means that when you analyze the most frequent tags, you’ll be seeing the most frequent relevant keywords.

You can also use tags as triggers to automate your workflows and processes as well as filters to analyze in depth.

Last but not least you could easily use other machine learning models to add information in other Zendesk ticket fields like:

• Entity extraction (names of persons, locations, organizations),
• Price extraction,
• Urgency detection classifier (classifies customer support messages into "urgent" and "not urgent"),
• Needs human action classifier (detects if human action is needed, for example, to respond to a question or request, deal with a complaint, or respond to negative feedback),
• Or even your own custom extraction or classification model.

We would love to hear more about what you are using tags for, shoot us an email if you have any ideas you would like to implement. Or better yet, just go ahead and create a free MonkeyLearn account and start automating.