This thesis project examines the opportunities for manufacturing operators to use root cause analysis directly at the line with the support of a cognitive advisor. The project was part of the EU Horizon COALA project and uses a specific business case at a detergent company as a base for the research and the prototype development. The analysis of the business case showed that currently operators do not use methodological root cause analysis at the line.

The report describes the research, development and showcases a prototype of a textual cognitive advisor that has incorporated the “5 Why?” root cause analysis method. This prototype was developed making use of multiple sprints adding new functionalities over time. The prototype was built making use of the open source RASA framework, GraphQL and NEO4J to develop and demonstrate the technical feasibility of the project.

The final prototype is able to have a conversation with the operator, support the operator with the “5 Why?” from problem statement till root cause by implementing best practices, and save each connected step into a graph database.

Furthermore a validation of a voice-chat-enabled cognitive advisor based on the business case is researched by conducting a (lab based) scientific experiment with 20 participants. The results of the experiment showed that there are interesting opportunities for cognitive advisors for operators in manufacturing in terms of interaction time and data quality for quantitative analysis use, but more research is needed for this topic.