Tesla has good customer service

Artificial Intelligence

Considering the amount of publications and events on this topic, 2017 should have been the year in which chatbots experience their big breakthrough. Even if chatbots are not in themselves a new invention - in the past year they experienced something like their big - one might almost say: second - spring. But many pilot projects have also failed.

Chatbots are programs that are able to use text recognition to react sensibly to human utterances. This allows chatbots to simulate real conversations with customers in customer service, for example - as if the customer were in dialogue with a human conversation partner. In this context, chatbots are also able to perform simple tasks, deliver information, initiate processes and react to commands. There are now some success stories around these virtual personalities. Usually, however, these are not direct customer service applications.

Fast responses, 24/7

But what makes automated communication in customer service even successful? The essential requirements that customers place on good customer service can be summarized relatively easily: a quick response to the request, if possible around the clock, friendliness and, finally, the certainty that personal data will be handled confidentially.

When it comes to giving quick answers, being available day and night and never being in a bad mood, chatbots seem to offer the ultimate solution for customer-friendly service. Communication with chatbots is basically what consumers want. Especially if it helps them faster and shortens their waiting times.

Surveys also show that if a chatbot performs the same as a service employee from a customer perspective, 55 percent of customers prefer the bot to humans. Despite all these promising prerequisites - in practice, customer service pilot projects with chatbots often lead to failure. The following rules describe the five central points that must be considered so that the use of chatbots in your customer service is actually as successful as hoped.

  1. AI bots
    Chatbots as virtual communication robots can optimize processes, support communication with customers and partners and provide new insights into communication partners.
  2. Google Allo
    The messaging app Google Allo has an integrated chatbot, the virtual assistant Google Assistant. For example, he looks for suitable restaurants and helps with reservations.
  3. Facebook
    Chatbots are also active within Facebook Messenger. Users can subscribe to messages from chatbots there, but they can also block them.
  4. Mastercard
    Chatbots are already finding their way into the world of banking. Mastercard offers the bot KAI for banks and their customers to simplify transactions.
  5. Microsoft
    There are a number of ways to develop your own chatbots, such as the Microsoft Bot Framework.
  6. News bot Novi
    Chatbots do not always look so clearly like a robot as in the case of the Novi news bot. Often times, the images in chatbots give the impression of communicating with a person.
  7. Novi answers questions
    You can ask questions about the news to the news chatbot Novi. However, the queries are still specified. Other chatbots already work with "free" questions.
  8. Chatbots learn
    You can tell a chatbot like Novi whether a topic is interesting or not. In general, chatbots should learn from this and increasingly personalize communication.

1. Use an appropriate channel

Think of the chatbot as an active customer service agent. This means: You should be able to use the bot - like your other employees - for various contact channels. It does not achieve the goal of viewing the chatbot as just another, more or less separate contact channel. This would only result in the entire knowledge, intelligence and experience being limited to exactly this one channel.

Sure, chatbots owe their boom in particular to the fact that Facebook opened its messenger platform for the development of bots. But this does not mean that bots are only at home in chat channels. A chatbot can also be active in its own app. Basically, a suitable channel must increase the convenience for the mobile user, for example through buttons with predefined answers and images. It is therefore important to avoid excessive complexity due to open questions and typing errors. In addition, it is also worth considering voice-controlled platforms, similar to Siri or Alexa.

2. Ensure a smooth collaboration

As little as you should limit a chatbot to one channel - you should also not limit a channel to the pure support of a chatbot. Because chatbots are still a long way from being able to really understand, answer or execute all inquiries. Seamless collaboration and a smooth transition between the chatbot and the human service employee are still essential.

Also make sure that the customer clearly knows at all times whether he is dealing with a bot or a human employee. If the conversation threatens to hang or the customer is clearly dissatisfied, a human service employee must be available. Customers also find this opportunity to switch from a bot to a real employee to be very valuable.

Conversely, an employee can also have a chatbot edit a process, for example to request data or make changes. This offers the customer a quick solution and allows your employees to devote themselves to tasks that are less standardized and that therefore require people with their special skills.