Many small and medium-sized energy suppliers now have chatbots in use. Especially in the service area this can help to handle recurring questions and simple problems more efficiently and save costs. In addition, chat bots are intended to reduce the workload of employees and increase customer satisfaction. Initial experience shows, however, that it is difficult, especially for medium-sized companies, to analyze existing data from the operation of their chat bots. Therefore, the present research project aims at developing an automated alternative for the analysis of existing chat data. Using operational data from a customer of Heidelberger Services AG and the platform of the Smart Data Innovation Lab, general methods for the analysis of chat processes are to be developed from qualitative and quantitative data. With the help of these, indicators and predictions will be made possible to identify customer sentiments and possible handover situations. This can also be used to improve the product in operation and thus increase customer satisfaction.