More and more companies are using chatbots as the first touchpoint in customer service. Very few companies can afford (and dare) to have fully speech-based dialogue systems interact with their customers – Alexa and Siri are market-leading examples. In most cases, these are text-based dialogue systems via standard and already familiar messenger interfaces. A lot has happened since the 20-year-old assistance system “Karl Klammer” from Microsoft. Today, around 35,000 company chatbots are already active in the Facebook Messenger to take care of the inquiries and concerns of users and customers.
Many chatbots are, technically speaking, more like a full-text search than artificial intelligence. For a first customer contact, they are quite suitable: they inquire and verify customer data, limit the search ranges, and supply cross-references to appropriate subpages and concrete offers of the enterprise. However, as soon as it comes to advise or recommendations off the beaten track, they quickly reach their limits. Repeated inquiries or not goal-directed references lead to frustration and negative experiences with the customers – and produce the exact opposite effect, which one wanted to achieve.
The results of a survey in Germany show that users attach great importance to the professional competence and user-friendliness of a chatbot “conversation partner.” The desire for transparency is also found in the top ranking, i.e., you want to know whether you are communicating with a human or automated counterpart. Customer benefit is seen in the fact that chatbots are always available and responsive – and that for a large number of interlocutors at the same time. However, around ¾ of the interviewees state that it is essential for them to be able to speak personally with customer service employees despite a broad digital offering. The reason for this is the lack of experience. In German-speaking countries, relatively few companies include chatbots on their websites.
The implementation of chatbots is a strategic direction. In order not to waste capital unnecessarily and to annoy customers, essential questions must be answered in advance, e.g.:
• How can a chatbot communicate and bring our services/products closer together?
• How can the conversion rate be increased?
• Can the conversion rate also be grown organically (i.e., without a chatbot)?
• Does the online touchpoint have emotional elements that cannot be covered by a chatbot?
If these (or other) central questions cannot be answered positively, companies run the risk of frivolously jeopardizing customer satisfaction and trust. At worst, the digital environment turns against the supposedly good thing: Screenshots with strange or inappropriate answers from the chatbots are quickly made and distributed – and the next shitstorm comes up.
The use of chatbots has potential! This technology must be appropriately used to become part of brand loyalty, re-purchasing, and sharing positive experiences. It may be sufficient to install a simple callback request on the website. Maybe it just needs a hint that the chatbot is used to refer to suitable subpages and offers. In any case, it is essential to communicate the fact and the technical possibilities of the chatbot to the users in advance. The alternative is to wait and see. Because artificial intelligence in conjunction with deep learning will make it possible in the future for chatbots to conduct real customer dialogues without getting caught up in loops and repetitions of content.
We are already looking forward to corresponding discussions about this … no matter with whom!