From user feedback to requirements using chatbots

Abstract

This thesis studies the opportunities of involving users in the requirements engineering process using chatbots. First a semi-structured literature study explores the current field of Crowd Requirements engineering, user feedback and chatbots in general. This literature study leads to a number of initial requirements for using chatbots in the requirements engineering process. Complementary requirements are gathered using a semi-structured interview with the stakeholders at the case study company, leading to an initial requirements list regarding chatbots in the field of requirements engineering. Furthermore, chatbots can help in guiding the users through the bug reporting process by controlling the flow of the conversation and using follow-up questions to further specify information. Therefore, we have designed and evaluated a conversation flow specific for this case. However, we have found that the 5W’s used in journalism can be applied to specify the details of a bug report. Finally, we have implemented automatic entity detection in the prototype, this way we were able to further pinpoint the specific bugs and turn them into proto-requirements. After the evaluation the subjects were asked to rate the chatbot using a usability scale, and were interviewed to find out their opinions about using a chatbot for requirements engineering and identifying the strengths and weaknesses of the chatbot. Finally, the resulting bug reports generated by the chatbot were discussed with the developers, they rated the bug reports generated by the chatbot higher or equal to the original bug reports. The most recurring reason for this higher score is the structure that the generated bug reports provide the developers with