Technology links energy consumers to 24×7 instant customer service
eprimo is one of Germany’s top electricity and natural gas suppliers. As a winner of many awards, we focus on what we do well—offering excellent prices and delivering a great customer experience. Which is why we’re committed to using advanced technology to meet the needs of today’s connected consumers.
In the recent past, we interacted with customers through call center agents, email, the postal service and online forms. With customers increasingly turning to digital channels, this was no longer sufficient. Missing was a direct digital channel that could deliver the instant service customers expect. This is especially important in the fall when people in Germany look to switch energy suppliers to get a better deal.
We only have one opportunity to make a good first impression and deliver the information that customers need. If we don’t provide the best digital service experience, in a few clicks they’ll be gone.
Instant 24×7 response
Part of our solution is Sophie, a 24×7 conversational agent developed with IBM Services and powered by IBM Watson Assistant and IBM Natural Language Understanding from the IBM Cloud. Sophie—the first point of contact on our homepage—can instantly answer questions about a customer’s bill and due date, new service offerings, meter readings, and changes to customer contact information, banking details and energy services. It also provides more general information about eprimo, the energy supply and our climate change policies.
Our collaboration with IBM and the commitment of our colleagues brought Sophie to market very quickly. Applying agile, iterative methods, the developers created a prototype in just eight weeks. We tested it in the busy fall season, and within 12 weeks the assistant learned to answer nearly all frequently asked questions.
Following the first successful prototype, a dedicated internal development team continuously works to increase both the amount of services included and the quality of Sophie’s interactions. For queries too complex or ambiguous, Sophie now transfers the session to a live agent.
Popular with the customer base
Connected customers really have taken to the service. When we introduced Sophie in late 2017, it received around 700 inquiries each month. These days, it receives more than 70,000 monthly and correctly answers 55 percent on the first try. User feedback tells us that those who get a quick response are even more satisfied with the virtual agent than with a live agent.
Like any new employee, Sophie had to be trained. Watson first absorbed and analyzed unstructured textual data from previous customer chats with live agents. Then, experienced customer agents and IBM data scientists monitored the dialogues to correct wrong answers. AI machine learning and monitoring by the dedicated team will continue to boost Sophie’s accuracy over time.
A digital transformation of customer care
Thanks in large part to Sophie, our digital interaction rate has increased from 40 to 75 percent over the past two years. Agents are freed from answering routine questions and can focus on issues where customers prefer the human touch, delivering higher value to them. And the cloud-based service easily scales to meet the rush of fall and winter queries to deliver high customer satisfaction.
Another plus of this project is the AI knowledge IBM transferred to our personnel. We at eprimo will apply it to other business processes to further improve the customer experience and our competitive position in the marketplace.
Watch Jens Michael Peters discuss how eprimo has leveraged Watson to drive digital transformation: