SUBSCRIBE

Chatbot Use Case: I am Horst, the digital concierge at Deutsche Telekom

Horst is a real star in the Telekom network. Almost everyone likes him, even though he is one of those annoying types who asks for a favor after responding to a request. Would you mind answering a few short questions? To help us ensure customer satisfaction. We all know the routine. We all know it’s an irritation. But Horst is pleasant to deal with. Read in this blog post how Deutsche Telekom developed a chatbot with its own personality and language and has been using it successfully for customer service since 2018.

Smart Customer Service with Chatbots

72 hours after a completed service inquiry, the customer receives a personalized message from Horst, written in a relaxed and fun style. While somewhat unexpected coming from such a large company, it has been both effective and well received. Horst is responsible for all customer satisfaction surveys. The Telekom’s chatbot is a further innovative building block in their overall customer service vision and strategy which has been systematically expanded over the years. In addition to the chatbot, their multifaceted customer care plan uses Knowledge Connect, a social media customer service tool to centrally manage and reply to requests from Facebook, Twitter, forums, blogs and the Telekom’s internal support community. Requests are analyzed and forwarded automatically to the relevant customer service team.

To strike the right balance between creating value with automation but still respecting customers and avoiding impersonal mass communication, the Telekom analyzed which classic service processes could be partially automated. In order to maintain a high level of person-to-person service, it was decided that only an isolated and independent process should be automated. Therefore, they selected a uniform customer satisfaction survey across all social media channels. To automate this process to the greatest degree possible, the Telekom decided for a chatbot that is integrated with Knowledge Connect. It checks in with customers after they’ve made a service inquiry or request to see how happy they were with the service received.

Horst Works as Part of a Team

Although the chatbot is an automated solution, it is just one part of the overall customer service process. Therefore, it works as part of a team and follows built-in escalation points when a customer should be handed off to a human agent.

It analyzes customer responses to decide whether clarification is required and what next actions to take. Satisfied customers are offered additional products and services whereas dissatisfied customers are forwarded to an agent for individual communication. Horst the chatbot solves a key challenge for the Telekom in that it normally marks services requests as complete once they have a result or solution. However, the customer in question may still be unhappy. Using the chatbot, it can quickly identify and forward cases that still require a follow-up leading to increased long-term customer satisfaction.

 

The personalized and individual aspects of customer service are extremely important to maintaining our best-in-class position in the market. With Unymira’s chatbot, we’ve succeeded in establishing a new service channel with its own unique personality that has dramatically boosted response rates compared to conventional surveys.

Oliver Nissen, Deutsche Telekom AG

 

hausmeister-horst

5 Star Chatbot

Horst’s personality stands out in particular, being built with its own unique character and language style. This makes him both recognizable and more relatable for customers. He also identifies himself immediately as a bot and yet still has an acceptance rate of 75%! Classic customer satisfaction surveys via email have an average answer rate of 3% to 5%. Horst’s response rate is a whopping 35%, between 7 and 10 times higher than usual. Within just a few months of launch, it generated over 5,000 customer responses.

Based on the positive customer reactions and excellent results, additional specialized chatbots are planned for the future. For example, a chatbot for sales is being developed that will initiate contact with customers, however only following strictly defined criteria. Moreover, only those who have previously communicated with the bot and found it positive will be contacted.

Chatbotcon 2020 - Kostenlos anmelden!

 

Share article:
Henry Michel

Henry Michel

After completing his studies as an industrial engineer with a focus on technical logistics, Henry Michel was responsible for complex software as a product manager and project manager. At unymira, he brings his accumulated experience as a product manager to bear especially in the areas of chatbot and self-service. His expertise covers both IT self-service and customer self-service.

Transform your knowledge in customer service!

More interesting articles