Problem: An omnichannel-chatbot is to be introduced in a company, which requires a knowledge database as an information source.
Objective: The objective of the work was to model a knowledge database for a chatbot and to derive measures and strategies for the conceptual design of this knowledge database.
Research question: What measures can be derived from the conceptualization of a knowledge database for an omnichannel-chatbot to be able to answer 80 percent of customer and candidate inquiries?
Method: The research method used to address the research question was design science research, which involved adhering to the guidelines and framework established by A. Hevner. To accomplish this, a variety of steps were taken, including conducting a thorough literature review, engaging in brainstorming sessions with peers, creating a relational model, and conducting interviews with reception assistants. The objective of these steps was to define and assess concrete use cases for the test phase.
Results: A relational model of a knowledge database and a catalogue of measures for the creation of the knowledge database were created. The relational model is available as a MySQL Workbench file, and the catalogue of measures contains the steps for selecting the database model, creating the requirements catalogue, identifying the data sources, describing the technical measures, and creating the relation model.