Today, Jörg Heinemann talks to Sophie about the use of chatbots and the Google Assistant at OTTO. In this podcast episode, you can find out when OTTO started using chatbots, how the first use case was defined and who Clara actually is.

About Jörg Heinemann

Jörg Heinemann is Principal Innovation & Digitalization at OTTO. Every day, he works to ensure that customers can communicate with OTTO in completely new ways, whether via chat or voicebot. He is also active in the smart home sector and is working on household appliances that in future will trigger a coffee order without any communication with the user when the coffee in the machine is running low, for example.

And when was your first time?

“When I heard about the first Echo devices at the beginning of 2016, I wasn’t convinced that this technology would catch on here,” says Jörg Heinemann five years later. On a business trip to the USA, he was to be proven wrong. There he had the opportunity to meet some device manufacturers in person and was overwhelmed by the sheer number of devices. He quickly realized that it was only a matter of time before the appliances found their way into our household. From now on, he will support the individual initiatives at OTTO when it comes to chatbots.

Where is my order?

In 2016, OTTO began using the Google Assistant on various devices. Initially, users could only inquire about the delivery status of their shipment by voice or text. Gradually, however, the service has been expanded to include additional service questions. Nowadays, you can also be informed about the deal of the day and even order it directly. Devices with a screen in particular offer great potential here, so that the user can get a better picture of the potential purchase object, says Jörg Heinemann.

In addition, users of OTTO’s own app now mainly chat with Clara, a chatbot. Only when Clara is unable to answer the questions to the customer’s satisfaction does an employee at OTTO’s Relation Center take over. In the future, Clara will also be used in WhatsApp. Although OTTO is also already represented here, customers only chat with natural persons and not with a chatbot.

The first use case

OTTO took a very pragmatic approach to finding out the first use case for the Google Assistant. First of all, the question of what the most frequent questions from customers are was clarified. To this end, we worked closely with our colleagues from the Service and Relation Centers, as they process thousands of inquiries every day and therefore have the best overview. Once the most frequent customer inquiries had been listed, the next step was to find out which questions could be implemented technically and in a timely manner. In addition, attention was paid to creating real added value through automation, so that employees and the Relation Centers have more time for other things elsewhere.

Finally, the first use case identified was shipment tracking, as described above. It was discovered that the majority of customers used to log into their customer account or call the service center. Automation has created a win-win situation here, reports Jörg Heinemann. Customers receive a quick and uncomplicated response and OTTO as a company saves resources with the bot, which can be used more effectively elsewhere. Following the shipment tracking, further questions from customers were automated in the bot, which can now answer several hundred questions. This has significantly increased the interaction that takes place exclusively with the bot. In April/May of this year, a quarter of inquiries were received, but by June/July this figure had risen to around a third.

The personality of Clara

First of all, so far Clara has only written and not spoken. Nonetheless, OTTO has put a lot of effort into the tonality and was able to draw on existing know-how within the company. The teams from the Corporate Identity department, which deal with OTTO’s brand language among other things, played a key role in this. As a result, Clara writes in an approachable, simple, direct and personal way and is on first-name terms with users. Their language is based on that of the employees from the Relation Centers. Even if it is always made clear that the user is writing with a bot, ideally there should be no recognizable difference between the bot and a natural person purely in terms of interaction, says Heinemann.

Learnings

1. just do it: You don’t have to automate the entire range of customer questions, the main thing is to get started.

2. when defining the use case, you should look at where the biggest pain is that can be solved most quickly by automation.

3. the personality of the chatbot should be based on that of a real person from the Relation Center. However, it should be made clear that the user is chatting with a chatbot.

4. if a company is active on different channels (e.g. website, WhatsApp) then these channels should use a uniform technical basis. In this way, the bot learns in the background regardless of the channel through which a request arrives and users can expect a consistent level of service.

It’s best to listen to the podcast episode with Sophie Hundertmark and Jörg Heinemann for yourself. Have fun!

And if you don’t want to miss any more podcast episodes, then sign up for the mailing list at the bottom of this page.