Designing Hybrid Customer Service Agent Systems

Problem space

Conversational agents, e.g. in the form of chatbots, are being used more and more in the service sector and are also finding greater acceptance among the population thanks to user-centric research. The use of conversational agents is particularly worthwhile in industries with a high volume of standardized and repetitive processes and customer inquiries, such as banks, insurance companies, and especially the energy industry (hsag).  With a dialog-oriented approach, these agents can provide information, answer customer questions and thus process them conclusively, perform data entry or record data and forward it accordingly. Customer service is thus relieved and can devote itself to more specific and complex concerns. However, since the machine data processing of conversational agents still has linguistic, logical as well as data integration limitations, not every customer concern can be processed to satisfaction. In order to take advantage of machine processing, while at the same time not creating disappointed customers in the event that AI fails to address the customer concern, a close integration of human service agents and machine conversational agents is needed - a hybrid customer service agent system.

Project Goals

Against the background of this problem, the goal of the doctoral project is to develop a design that ensures a high level of customer satisfaction when dealing with conversational agents and, at the same time, also takes into account the needs of employees in a hybrid solution. The project is therefore at the interface of research on conversational agents, customer service and human-machine collaboration. Divided into several subprojects, methods and concrete artifacts are to be developed to estimate the quality of dialogs with AI conversational agents, to identify handover points, and to design the handover from machine to human and, in general, the collaboration between machine and human in a customer- and employee-friendly way. Since design principles play a role at many points in this process, such as the form in which a customer is notified of or executed a switch to a live chat with a human agent, the project is build as a design science research project. In an iterative process of development and experimentation, a design for hybrid customer service agent systems is to be developed, which in its applicability simultaneously enriches the practice in customer service and in its generality and novelty also the research world. The contribution of the research project should be able to answer questions such as the following:

  • Which errors occur in conversations with intelligent conversational agents in the context of customer service?
  • At which points does it make sense to use human support and how can these points be determined?
  • How can human-machine collaboration be designed? Specifically: How is the hybrid customer service agent system solution shown to the customer, e.g., in the form of the interface or waiting times, and how is it tailored to human customer service agents and their workflows?
  • ...


Further information on hsag

The Heidelberger Services AG is a medium-sized service company specializing in the processes and needs of the energy industry. With the areas of service, consulting, business process outsourcing, marketing and software products, hsag offers its energy industry customers a holistic spectrum of support for their operational business.

Please find additional information on Heidelberger Services AG here.

Furthermore, feel free to contact Daniel Schloß to learn more about the project, hsag, or consulting and services in the energy industry context in general.