Towards an Empathetic Chatbot: Effects of a Chatbot Responding to User Frustration on Customer Service Interactions
Keywords
Loading...
Authors
Issue Date
2019-07-05
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
The purpose of the current study was to test whether an empathetic chatbot, which can
recognize user frustration, would increase the number of words participants use during the
interaction. Different studies have provided solutions on analyzing user emotion and other
studies have provided insights into communication styles that are beneficial for chatbot
interactions, however there is a lack of studies on a combination of matching communication
to user emotion in the context of customer service bots. A software which analyzes
frustration from written text was validated with a pilot study. This software was implemented
into a new chatbot which sends either an agreeable response or mirrors users’ punctuation
style based on the level of frustration. 48 participants (19 male, 11 female, 1 not specified)
with an average age of 23.4 played a frustration inducing game, after which they interacted
either with the empathetic bot or a control bot. Results indicate that participants did not use
more words during an interaction with the empathetic chatbot compared to the control bot,
probably due to a lack of statistical power. Explorative analysis indicates that there is no
difference on the intention to use the bot again, but that frustration was decreased more in
users interacting with the empathetic bot.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen