Description

Problem

Conversational agents (e.g., chatbots, voice assistants) are still gaining popularity.
Similar to other interactive systems, conversational agents require a continuous
user-oriented evaluation to ensure that they meet users‘ requirements.
Currently, feedback is often collected via thumbs up/thumbs down features. While
this simple feature allows designers to understand where users experience
problems, it does not help them to understand why users are unsatisfied with the
chatbots’ response and how they could improve it.
So far, there is a lack of research on how to elicit user feedback during the
interaction with a chatbot.

Goal

The goal of this thesis is to develop ideas for feedback features based
on existing studies and theories. Then, these evaluation features shall be
integrated in a simple chatbot prototype, followed by an evaluation of
the effects of the individual feedback features on feedback quality and
quantity.
For a master thesis, the most promising feedback feature shall be
included in a live chatbot of one of our industry partners to collect real
user feedback and evaluate the feature in a large-scale field study.