How to feed your Chatbot

Voice assistants: Also a topic for Service and User Information?

FeedbackPublication
© panuwat / fotolia

© panuwat / fotolia

In our four-part blog series "How to feed your Chatbot", we highlight the topics of chatbots and voice assistants. In doing so, we would like to demonstrate how such assistants work in general terms and, of course, whether (and how) such assistants could be used for both service and user information.

Digital assistants such as Google Assistant, Siri, Alexa, and Cortana already support many people in their everyday lives: The next workshop appointment is virtually dictated into the calendar, the weather forecast or the population of Austria are queried, and one can also ask "his" or "her" assistant how to get to the booked hotel in a foreign city the quickest.

The terms "chatbot" and "voice assistant" are often used interchangeably. With the exception of complex speech recognition and synthesis, both concepts usually have similar logic: The software must respond to a user's request. Either an action is executed as described above or the software must prompt for further information through a query, and then a dialogue is started. A chatbot reacts only to written statements and also displays its answers as written text. The voice assistant can process natural speech, and it responds with synthesised speech – sounds similar to a human being.

Overall, the subject is a controversial topic of discussion. No Wonder: Not everyone is comfortable with placing a microphone with a direct line to the servers of Silicon Valley companies in their living room or bedroom. Reports of Alexa devices laughing at night for no reason certainly don't contribute toward building confidence. However, there are not many choices left, because the computing power that makes complex speech processing and artificial intelligence possible in the first place cannot be provided offline by our terminals. At this year's I/O developer conference, Google, with Google Duplex, impressively demonstrated the top-class performance voice assistants can already achieve today.

Worldwide, we seem open to the use of chatbots in customer service. So reported the German Customer Service Association (Kundendienst-Verband Deutschland e. V. orKVD) that already in 2017 from a study, according to which, worldwide 89% (91% Germany-wide) consider the use of chatbots in customer service as neutral or even positive. Generally, one-in-four German citizens is interested in using chatbots, according to the Digitalverband Bitkom e. V. (a Digital Association). Nils Britze, Consultant for Digital Business Processes at Bitkom e.V., also sees chatbots as a building block for automated communication processes within digital customer communication (Chatbots and the Automation of Communication Processes).

Use Cases for Service and User Information

Voice assistants also seem to be suitable for providing the service or users with relevant information. It would probably be of little help if instructions for action were to be read aloud. It might be more helpful to ask specifically for connected values or possible production parameters. Receiving an understandable "translation" for the numeric error code can also save time. In a Question-Answer scenario, quickly pinpointing and identifying a technical defect could even be a "Key Use Case", which alone, might even justify the effort of a voice assistant.

However, the Key Use Case can be unique in every company, and so each voice assistant must therefore be designed individually. How this can work, you can read shortly in the second blog of our series "How to feed your Chatbot".

If you would like to experience for yourself how a voice assistant for user information functions, we’d like to introduce "Agent Smarty" to you. With the help of the Google Assistant and a Google Account, you can easily have our Assistant on your Smartphone. How it works and which questions he can answer you can be found in our News "Try out the kothes Voice Assistant".