The subject of messengers in customer service is taking off again. WhatsApp - the most widely used messenger in Germany - provides an official interface (API) with “WhatsApp Business”, which enables companies to use the messenger for customer communication. OTTO is the first online retailer to use WhatsApp for this in customer service. From now on customers can direct questions about orders, cancellations or advice on products and services via WhatsApp.
For many people today the messenger is their number 1 communication tool. But does anyone want to communicate with a company that way? Daniel Lipinski, an expert in customer service at OTTO, says: “Customers primarily use the communication channel that they like to use day to day to get in touch with companies. Anyone who is at home in the social media world will prefer to use it to contact companies. Someone who writes dozens of emails every day will generally choose this way.” Put simply, customers with an affinity for messengers will use this as a means of contact, others will prefer to give us a call.
A “WhatsApp” is quick to write - and quickly replied to
Arne Flechner is the Head of Customer Service at OTTO. He and parts of his team do what teenagers spend a lot of their free time doing - answering messages from messengers and social media. “The communicational rules of the customer service game are different on these channels than in emails or classic letters,” explains Arne. And that means? “Communication takes place in a dialogue, similarly to in chats with friends or acquaintances. So customers and service employees quickly resolve an issue by exchanging several short messages. The level of communication is therefore more direct and personal,” says Arne. Another thing we find is that lots of customers react to a reply over messenger very quickly. But according to Arne the inquiries and subjects are the same as service employees deal with over the phone or via email.
Chatbots require patience
If you listen to a few futurologists, then bots will soon take over a large share of customer communication. But, reality looks a little different. “When we consider AI and machine learning we are all quick to think of a chess computer, which is so clever that it defeats rows of chess professionals,” explains Björn Spielmann, Senior Project Manager for Service & Customer Experience at OTTO, “but in the end that relies on logical mathematics.” When it comes to chatbots, language plays a central role. This makes things significantly more
complicated. Because though they are often glorified as such, chatbots aren’t Mozarts or Sheldon Coopers, which quickly teach themselves everything like geniuses. They are actually similar to small children, who must patiently clarify lots of things. They need continuous training to improve.
Language, a very complex subject
“Today a spelling mistake can be enough to get an incorrect or meaningless answer from a chatbot,” explains Björn. To this end, language should always be seen in context. So “thank you for being a great support” may be meant as a friendly farewell or as an ironic parting statement. A person recognises that but a bot today can’t yet do the same. Sure, chatbots are already used in customer service and answer simple questions quite well. But until they can really resolve complex customer issues, developers, service employees and editors will continue spending a lot of time training the bots.
And then customer centres find this: “A good chatbot supports people by resolving simple issues. That relieves employees of work and enables them to deal with complicated customer issues.” That means a chatbot filters the customer issues presented to them and either answers the query itself or passes it on to a service employee. So in the future it won’t be a case of “humans vs. bots”, but “humans with bots.”