A high-quality customer experience in the age of digitalisation is practically indispensable in all customer scenarios. This also includes the service promise, where companies are more and more frequently relying on smart technologies, and where chatbots with text and speech-based digital systems are now at the forefront. A representative survey on behalf of the digital association Bitkom, found that every fourth German resident can already picture themselves communicating with a chatbot. Alexa, Cortana and Siri are also well-used speech bots that simplify our every-day lives.
But if you thought that chatbots could only be used to automatically answer FAQs or to interactively replace FAQ lists, you’d be wrong. Bots can also trigger background processes, including making personal appointments, online shopping, and even sending flowers to a certain person.
A growing number of companies therefore uses this technology for their customer service. Through coordinated collaboration with their human colleagues, bots, armed with AI, can independently handle service requests and answer them accurately. And the biggest benefit is that chatbots aren’t only always available, their software also lowers the huge amount of customer requests coming in over the phone and by e-mail.
IT service desk resources continue to be mostly tied up with standard questions along the lines of:
If just these questions could be automatically answered, then huge amounts of customer service resources would be freed up to deal with more complex problems. But it’s also clear that chatbots aren’t able to have conversations about all topics. Each bot is actually a piece of software that relies on reworking the expert knowledge of human experts. An essential and challenging aspect of every chatbot project is: just how does the knowledge get into the system?
For development and implementation, the right expectation management plays a decisive role. Although the most important use cases represent the greatest benefit for companies, it makes sense to start with a simple use case and add functionality over time. It’s therefore a good idea to start with a proof of concept and seeing whether using a bot is actually beneficial. Once this has been carried out, needs analysis, entity definition, content creation, and bot training can be carried out, in addition to technical integration. Because chatbots use a diluted form of artificial intelligence, it is also useful to train them within a narrow scope of duties, since the larger the scope, the easier it is for the bot to confuse the user’s intent.
Accordingly, we in the system house have also started enthusiastically to develop a chatbot with the aim of identifying possible potential for use in a corporate context. In cooperation with our service management, we have now succeeded in mapping the first use cases with our chatbot. Please contact us if you would like to learn more!