News

Introducing MonkeyLearn integration with Zendesk

Raúl Garreta
by Raúl Garreta

Introducing MonkeyLearn integration with Zendesk

We’re thrilled to announce our MonkeyLearn integration with Zendesk!

Most of everyday business interactions with our customers are done through text from emails, chats, and social media channels, among other sources.

In particular, customer support involves processing huge quantities of text, in the form of support tickets that are opened by customers whenever they need help.

Usually one or multiple agents in the support team have to read through the ticket to understand the request from the customer, which takes a considerable amount of time. What if machine learning understood human language in these tickets? Wouldn’t it be great to be able to automate some of this process?

This is why we have built the MonkeyLearn integration with Zendesk, to help customer support teams use AI within their workflows. Let’s check it out!

How does it work?

Whenever a new ticket comes in, a Machine Learning model will automatically process the content (subject and body) to analyze for the following:

  • Intent
  • Topic
  • Sentiment
  • Urgency

By having those characteristics (metadata in the form of tags) automatically identified by a machine learning model, we could save huge amounts of human labor, which means saving money and providing higher quality service (faster response times and more accurate responses).

To predict each of the previous four aspects, a machine learning classifier will be trained to predict the corresponding tag based on the text input (subject + body) and desired output (intent, topic, sentiment or urgency). Our first version of this integration can train a machine learning classifier based on your Zendesk historical data (that is, tickets you received and manually tagged) to predict the corresponding topic that must be registered in a categorical field.

Here’s what the settings look like:

zendesk ticket classification
Setting up the Zendesk + MonkeyLearn integration.

Here’s what each section will control:

  • Field to classify: sets the field that you want to predict with MonkeyLearn, that is, the field that will be automatically populated according to the ticket’s subject and body.
  • Automation level: sets how much automation you want. The higher, the more automation, the lower the less, which means that only high confidence tickets will be automatically populated on Zendesk.
  • MonkeyLearn API token: is the token that is required to connect with your MonkeyLearn account. You can get your token here after you sign up to MonkeyLearn.

What do the results look like?

It’s a very subtle change to see predictions from text analysis in your Zendesk view. What changes is that MonkeyLearn will populate the metadata in the corresponding field automatically every time a new ticket is received.

zendesk ticket classification
MonkeyLearn predicting the ‘About’ field on Zendesk.

In this example, the About field will be added automatically in Zendesk: being that the ticket subject and content refer to a Shipping Problem, and the model detected it as such.

Based on this result, you can then use triggers or macros as you would do with any other field or tag.

Benefits

The following are some of the benefits of using the MonkeyLearn integration with Zendesk.

Reduce response times and increase resolution rates

By having fields like topic, intent, sentiment and urgency automatically added, customer support agents can have more data about the nature of a ticket which will help them to answer faster.

Moreover, by combining these pre-populated fields with simple triggers and macros, you can achieve:

  • Smarter ticket routing: instead of a simple round-robin strategy, you will be able to automatically route your tickets more intelligently to the most appropriate agent. For example, if a new ticket is about a billing issue, automatically route it to the side that handles billing.
  • Better prioritization of your tickets: some tickets may be more critical for your business to respond to than others, is it ideal for a customer support team to stick to a first-in-first-out strategy all the time? Text analysis can help detect urgency, angry customers, up-selling opportunities and more magic moments that can help customer’s experience.
  • Suggest automatic responses: based on the ticket metadata (topic, intent, etc) a subset of responses can be suggested from your knowledge base. Making the answer fully automated by a trigger or reducing time for a human agent to find the right response by using a macro.

More time for your agents

By reducing response times and automating simple responses, AI can free up more productive time of your agents. Instead of responding to repetitive and simple questions, they will be able to focus on more complex tasks.

Get consistent analytics

Getting consistent and trustworthy metrics is vital for managers trying to track the performance of a customer support team.

How many tickets have been received per topic? What is the overall sentiment of our customer interactions?

Those are just a few key questions managers will use to detect problems or opportunities in their products and to improve the customer experience. By using Machine Learning, those tags will be populated automatically and consistently for every single received ticket. They can even analyze your historical data, which is a huge plus.

Final words

Our goal is to make AI accessible to companies in order to start automating part of their business workflows. Customer support is one of the most time-consuming business tasks that involve text data. We believe AI should complement and empower humans instead of replacing them.

If you are a customer support manager and need assistance or have any questions, don’t hesitate to reach out our team at hello@monkeylearn.com. We’re more than happy to help you set up MonkeyLearn for your Zendesk team. We also would love to have your feedback on this integration, so please feel free to contact me personally at raul@monkeylearn.com.

Raúl Garreta

Raúl Garreta

MonkeyLearn Co-Founder & CEO. Machine Learning and Natural Language Processing expert. Author of "Learning scikit-learn: Machine Learning in Python".

Notification

Have something to say?

Text Analysis with Machine Learning

Turn tweets, emails, documents, webpages and more into actionable data. Automate
business processes and save hours of manual data processing.