Chatbot at your Service!

Every organization that delivers products or services knows the issue: How can I provide maximum customer services with minimal investment. This blog will discuss how a chatbot can help improving Customer Satisfaction and lower operating costs.

Organizations have tried different approaches to lower the costs of customer service.
Don’t publish the helpdesk phone number on the website; provide support only through the website with an FAQ section and hopefully, you don’t get too many negative messages through social media. Use an Interactive Voice Response System on the helpdesk phone number to guide callers to the right person or let them being lost forever in your menu structure.

Customer Service and customer satisfaction
Introduction of paid customer support services so people have to pay for each call, trying to filter out unnecessary calls. Most of the time this results in a situation where customers are only calling you when it is too late to fix.

These efficiency improvements have an important negative side: your customer satisfaction will go down.

To measure the negative trend in customer satisfaction and Net Promotor Scores, organizations are spamming their customers with evaluation forms to get confirmation of the feeling they already had. Why would you send a survey if you already know the outcome?

Customer service is always a challenge, each customer wants to be treated with the highest priority and personal touch. You want to know when your computer is fixed at the repair shop, you need to have an answer on your mortgage request for buying your dream house or your mobile phone broke down and you want to know what to do. In all cases, you want to have the answer instantly without having to wait or do research on a website to find the right data. This is where the chatbot could come in to help out.

What is a chatbot?
According to Wikipedia: A chatbot (or a chatterbot) is a computer program that conducts a conversation via auditory or textual methods. Such programs are often designed to convincingly simulate how a human would behave as a conversational partner, thereby passing the Turing test.

Chatbots are typically used in dialog systems for example in customer service systems to have a conversation with customers by interpreting questions from customers and provide answers. Integrating natural language processing in a chatbot will increase the quality and level of the dialogs. Further improvement can be reached by connecting a repository that contains terminology used in the area where the chatbot is enrolled. So, a chatbot in the medical world has a different vocabulary as the same chatbot that is enrolled in the finance world. These repositories will be enriched over time when using Machine Learning- technology. These self-learning algorithms are able to improve themselves by using a feedback process to learn from their mistakes. The combination of these technologies has brought chatbot solutions to a level where customers sometimes will think they are chatting with a real person rather than to a self-learning computerized system.
What is the added value of a chatbot
This is all very nice but chatbot technology is still not good enough to answer all customer questions automatically. That is true but suppose a chatbot is able to process between 35 and 50% of all user calls. This means that a customer support center with 10 people answering customer calls could be reduced to a customer support center with only 5 people answering customer calls. These 5 people are usually way more specialist since the chatbot will deal with the generic questions and only forwards the questions that cannot be processed by the chatbot. These are usually difficult or very specific questions.

So you will end up with a customer support center that employs fewer people but is more specialist on the topic resulting in fewer costs and better quality of services. That sounds a bit like lower investments on customer services with higher customer satisfaction, something we thought was not possible. Don’t party too early, there is also an investment needed to realize the chatbot. It is not an out of the box app that can be installed by running ‘setup.exe’

As said, a chatbot is a combination of different technologies like Machine Learning and Natural Language Processing that needs to be combined with a repository that contains specific terminology. It will take time to implement this. A lot of configuration and tuning is required to get the chatbot to work properly. The system needs time to learn and improve itself. The configuration and tuning will take for example 1-2 months, the actual learning process can take longer. Usually, this is done by running the chatbot as a ‘shadow’ system. Customer conversations are entered into the chatbot for a certain period to provide the necessary knowledge to the chatbot. The length of this process is very depending on the volume of conversations that is available but usually can take up to a few months. So all together, setting up a high-quality chatbot is not an overnight installation and therefore requires a significant investment of time and money.

The technology behind chatbots is improving rapidly, especially in the area of text analytics and machine learning. On the long-term chatbots will become more common in the area of customer communication but for now, it offers the opportunity to be disruptive and create a competitive edge for your organization by leveraging this new technology.

Recent Posts