WhatsApp, Facebook Messenger and Telegram – chat programmes have become the new standard means of communication. But many people not only want to chat with their family and friends, but also to communicate with companies and organisations using the same method. Here, chatbots provide support for customer service, as a marketing tool or for knowledge transfer. For effective communication, however, they must be equipped with the sufficient knowledge. Most chatbots are modelled manually – an expensive solution with limited options. This is why Nina and Dirk Wenig from the University of Bremen developed the intelligent chatbot IDA Bot. It is based on machine learning and independently acquires domain-specific knowledge from websites, manuals, newspaper articles and other sources. This knowledge can be managed and adapted via user interfaces. “Chatbots are of particular interest to companies because they are, for example, able to respond to customer enquiries around the clock” explains Nina Wenig. “Artificial intelligence is an exciting phenomenon for many people, but it can also be intimidating as regards the perceived loss of control. This is why we are also thinking about control possibilities.” The process used makes it possible for companies of any size to quickly deploy an intelligent chatbot without having to rely on extensive knowledge modelling. A prototype of the IDA-Bot is currently being tested with companies in order to be able to further accommodate actual customer interests.