It is fundamental to understand your customers and what they need your help with to roll out a successful customer care strategy. Not only does it put a purpose behind your actions, but it also enables you to track progress and whether you genuinely helped them in the long run. In the previous blog, we covered four steps of how Natural Language Processing can improve your customer care. This blog goes into more detail about step 1: Understanding Your Customers. Let’s discuss the various ways to know what annoys or pushes away your customers. After all, knowing the problem is half the solution.

How are customers contacting you?

If customers are unhappy, they will (hopefully) make it known, and the way they choose to do so can already tell a lot. Call centers, online reviews, Twitter, FAQs, e-mails or even traditional paper letters are channels your customers may use to communicate with you. Having a clear overview of all your communication channels matters for many reasons:

  • Different customers use different channels. Calling often means that they expect direct help. Sending e-mails can indicate that they want to share complex information. If they use Twitter, they may wish to express their frustration or expect others to chime in.

  • Each channel generates a specific type of data. Call centers generate audio, which is rich in information but less straightforward to handle than plain text. Tweets can contain images or GIFs, and letters need to be scanned and handwritten, etc.

  • Different solutions can be used for other channels. While it may be possible to create a chatbot that sends text messages instead of a person, it isn’t easy to replace an actual human being in a call center conversation.

Even though it is essential to try to meet your customers where they currently are, it may be wise to consider whether providing a different channel can help users that otherwise wouldn’t have contacted you at all.

When are customers contacting you?

Knowing when customers try to contact you can say a lot. Peaks in communication can indicate outages, issues with the website, or that your bills raise questions. Mapping peak times to events relevant to your business gives you a chance to fix the problem at its root.

Understanding when peak hours are can also help you better schedule your customer care staff and increase the speed at which you handle customer inquiries. Each channel may have a specific behavior. For instance, people might use Twitter during their commute but may prefer to call over lunch.

Which customers are contacting you?

Each business attracts and tries to reach different demographics. Understanding who faces what problems is therefore crucial. Besides by providing customized help for an individual, you can better address the issues at their core. Let’s say that you receive a lot of questions regarding bills and what they mean. If your entire customer base asks these questions, you may want to redesign how the information is presented. However, suppose you notice that it’s not your entire customer base but only a subset that purchases particular products. In that case, it is wise to examine whether the pricing information for these products is adequate. Lastly, the more critical customer to help may be the one who isn’t reaching out. By solely focusing your efforts on a potentially small but vocal subset of your customer base, you miss the chance to address issues faced by other customers who choose to remain quiet.

What are customers contacting you about?

While you may have a good idea of what issues your customers face, it is critical to set up systems that validate these beliefs. Natural language processing allows you to do just that. It is unlikely that a human can keep track of every call center call, tweet, e-mail, or online review, but a computer can.

Using techniques such as topic extraction, sentiment analysis, speech-to-text, or even machine translation allows you to keep an overview of what problems exist and how prevalent they are. This will enable you to choose the solution that will create the most impact. It may also highlight issues that you didn’t expect, offering you a chance to take action.


Customer care starts by understanding your customers. And Natural Language Processing can help you do just that. Extracting key topics from hours of call center conversations, translating reviews into a single language, identifying the tone and sentiment of an interaction can provide you with the tools to focus your efforts on the solutions that will truly make an impact and improve your business.

Happen to miss the other relevant NLP blogs? Don’t worry! You can check them out here:

This blog is written by Valentin Calomme – AI Engineer & Business Line Manager at Mediaan