Innovation Radar

Allow me to introduce… Mr. Chatbot, our new service team member!

Chatbots are on everyone’s lips, or should we say fingertips. Today, so-called “conversational bots” already respond to customer queries, help with purchases or collect debts. They are regarded as the big thing of the future.

“You owe your buddy Patrick two euros sixty!”, the fearsome colossus tells a frightened teenager after chasing him through his apartment. This is a video published by the Sparkasse savings bank to advertise its payment app, Kwitt. To do so, the campaign focused on a chatbot through which “creditors” can send personalized videos to a “debtor”. The chatbot therefore acts as a virtual muscleman. The success is overwhelming! The use of the app increased by 22% in just six weeks. 

“Hi, you are talking to the Terminator… how can I help you?”

The chatbot used by Sparkasse is highly entertaining. Apart from the use as a PR instrument, chatbots are already being used in customer service. There are different types of these robots and/or programs which enable customers to communicate in text form. As a general rule, there are simple and complex chatbots. 

Simple chatbots answer questions such as opening hours or the year a company was founded. They are classed as simple as most of them can be programmed by operators directly without requiring any previous knowledge. They can program questions that users are likely to ask and at the same time enter an answer. Such “question and answer” (Q&A) chatbots have rather limited functions: They can’t give any answers to questions that have not been previously programmed. 

Chatbots with the ability to learn on their own, on the other hand, are called complex chatbots. They don’t just answer questions, but create complex correlations: For example, if a user asks if they need an umbrella when leaving the house, the chatbot creates the correlation between weather and umbrella. Depending on the weather, it gives a recommendation to either take an umbrella or not. These complex chatbots also display small applications that are also executed if instructed to do so. Most of them have interfaces to other systems. They can trigger an order in a restaurant directly and forward it to a POS system, for instance. They are therefore already classed as an assistant, like Alexa or Siri. However, their voice control function is still being developed.

Both simple and complex chatbots can be used for various messaging services. Facebook Messenger is the most commonly used of them all. Chatbots can also be integrated in people’s own websites. Most of the time, they open as a small window on the side of the screen, briefly introduce themselves and ask how they can help. 

An extension of customer service for an improved experience

Many simple chatbots quickly reach their limit when customers enter a question and describe their issue. They then forward the query to a human member of the customer service team. Such chatbots can be frustrating for customers. 

A well-programmed, and therefore helpful, chatbot therefore goes a long way in customer service. The service department is therefore the first point of contact for finding out standard questions. The team member therefore no longer has to ask the customer if they have already tried to restart their computer. 

Chatbots are ideal for such standard questions. 

This significantly improves customer service as customers feel that they have been listened to and are not kept on hold for ten minutes. This gives the service team members more time for dealing with complex queries. Their “virtual colleague” therefore takes a lot of strain off them. An investment in a chatbot can therefore amortize quickly. 

The 24/7 availability of the chatbot for numerous customers at the same time is a major advantage. 

This doesn’t mean that human service has become unnecessary! The opposite is the case! The acceptance and use of chatbots clearly depends on the target group. It can also happen that a chatbot doesn’t understand a question as it reaches its limit if the question is worded slightly differently to the pre-programmed version. Humans do not have this problem. 

Team play therefore also includes the “virtual colleague”. After all, the Terminator also cooperated with humans. J

What else can the chatbot do?

Can it also order food? Yes… And book flights? Yes, even that! Make an appointment with the hairdresser? Mhm! But it doesn’t help with choosing new clothes! Actually yes, it can do that too! But it doesn’t know me! Well, if it can create a job profile based on the answers to some questions, its algorithms will surely tell it which shoes would match the dress you have selected… 

Although the way in which a chatbot answers questions gives the impression that a human is on the other side, this is obviously not the case. These human traits, or a “character”, can be programmed. But in the end, it is just a program with pre-programmed answers. 

Don’t worry, it doesn’t really know you… but is very capable of learning your preferences. 


Frightening or exciting? 

We will show you the uses we have planned for these chatbots in one of our future articles.

Leave a Reply

Your email address will not be published. Required fields are marked *

Post comment