Recipient Organization
UNIV OF HAWAII
3190 MAILE WAY
HONOLULU,HI 96822
Performing Department
Family & Consumer Sciences
Non Technical Summary
Artificial intelligence (AI) refers to intelligence displayed by machines and is capability of a computer or computer-controlled robot to accomplish works related to intelligent beings (Copeland, 2018). Current AI applications in marketing include search, recommendation engines, programmatic advertising, marketing forecasting, and speech/text recognition (e.g., Amazon Echo and Facebook messenger) (Faggella, 2018). Chatbots, which are alternatively known as chatter robots and AI bots, are conversational agents with AI that offer appropriate responses and answers to users based on pre-compiled data (Wailthare et al., 2018). One of the leading fashion brands has launched a chatbot on Facebook messenger to promote their new collection (Arthur, 2017) and has integrated chatbots with their out-stream video advertisements using custom AI (Williams, 2017). AI technology has become a solution for enhancing real-time customer service, brand reputation, and brand loyalty by transforming customer interactions (Maruti Techlabs, 2018). However, little research has examined customer experiences with AI-based customer service and content. This project will use a mixed methods research design. As a qualitative study, this project will examine customer experiences with and the uses and gratifications of AI-based customer service and content (i.e., video ad chatbots; chatbots on social media messenger) [Study 1]. As a quantitative study, this project will investigate the influence of customers' perceptions of quality (i.e., reliability; automated responsiveness/interactivity; assurance/competence; empathy; diversion; diagnosticity/Information source) of AI-based customer service and content (i.e., video ad chatbots; chatbots on social media messenger) on customer experiences, brand engagement, and brand loyalty [Study 2].?
Animal Health Component
0%
Research Effort Categories
Basic
50%
Applied
50%
Developmental
0%
Goals / Objectives
Artificial intelligence (AI) refers to intelligence displayed by machines and is capability of a computer or computer-controlled robot to accomplish works related to intelligent beings (Copeland, 2018). Current AI applications in marketing include search, recommendation engines, programmatic advertising, marketing forecasting, and speech/text recognition (e.g., Amazon Echo and Facebook messenger) (Faggella, 2018). For instance, the use of AI applications with Facebook messenger aims to shape the "online to offline" approach of chat-based purchases; it also enables consumers to order products via chat alone (Faggella, 2018).Chatbots, which are alternatively known as chatter robots and AI bots, are conversational agents with artificial intelligence that offer appropriate responses and answers to users based on pre-compiled data (Wailthare et al., 2018). AI-powered and real-time messaging bots for customer service could offer an opportunity for firms to connect with new and existing customers, provide real-time customer support with solutions and recommended products, and reduce customer abandonment rates and complaints (Maruti Techlabs, 2018; Williams, 2017).An increasing number of leading fashion brands (i.e., American Eagle Outfitters, H&M, Burberry, Levi's) have launched chatbots. For instance, Tommy Hilfiger, has launched a chatbot on Facebook messenger to promote their new collection (Arthur, 2016) and has integrated chatbots with their out-stream video advertisements using custom AI (Williams, 2017). Specifically, the chatbot or the AI bot invites viewers to become involved by clicking a time-sensitive greeting (Williams, 2017). Then, viewers are able to navigate the brand's latest collections while obtaining assistance from the AI bot. The AI bot asks a series of questions along with offering pre-set answers about a customer's taste and size (Williams, 2017). Next, viewers choose products that are suggested by the AI bot. Viewers are transferred to the brand's website to complete their purchases (Williams, 2017). Thus, AI for customer service makes self-service interfaces more intuitive and interactive to optimize personalized customer experiences (Maruti Techlabs, 2018; Williams, 2017).Artificial intelligence (AI) technology has become a solution for enhancing real-time customer service, brand reputation, and brand loyalty by transforming customer interactions (Maruti Techlabs, 2018). According to Maruti Techlabs (2018), more than 85 percent of all customer support communications will be performed without contacting customer service representatives by 2020.Previous researchers examined the effects of different tones on user experience in the context of chatbots for customer care on social media (Hu et al., 2018), the direction of creating affect and emotional resources for more human-like conversational agents (Banchs, 2017). However, little research has examined the influence of customers' perceptions of quality regarding AI-based customer service and content on customers' brand engagement and brand loyalty. In addition, apparel retailers in Hawaii do not employ AI technology yet.To provide managerial implications and strategies for AI-powered digital solutions, this project will use a mixed methods research design. As a qualitative study, this project will examine customer experiences with and the uses and gratifications of AI-based customer service and content (i.e., video ad chatbots; chatbots on social media messenger) [Study 1]. Uses and gratifications refer to the "how and why" of media use as well as the motivations of specific uses and the satisfaction people obtain from those usages (Joinson, 2008). As a quantitative study, this project will investigate the influence of customers' perceptions of quality (i.e., reliability; automated responsiveness/interactivity; assurance/competence; empathy; diversion; diagnosticity/Information source) of AI-based customer service and content (i.e., video ad chatbots; chatbots on social media messenger) on customer experiences, brand engagement, and brand loyalty [Study 2].
Project Methods
In the first year, before the main survey for Study 1, a pilot test will be conducted with online apparel consumers in a computer lab. Participants will be asked to identify any items that they have problems understanding or responding to. Based on participants' feedback, the wording of the questionnaire items will be revised.Concerning the main survey for Study 1, participants will be invited to participate in the research through e-mail. They will be asked to complete a self-administered questionnaire online. They (n = 25) will interact with AI-based customer service and content (i.e., video ad chatbots; chatbots on social media messenger) and then respond to an online survey with open-ended questions.After reading over all the participants' responses, the line-by-line method will be used to identify major concepts. Similar responses will be grouped together. Once all of the responses are grouped into a category, each category will be analyzed to determine the underlying themes. A textual description of the experiences of the persons (what participants experienced) will be developed. Two coders will be used for data analysis. To calculate inter-coder reliability, the following formula will be used:Inter-coder reliability = (Number of agreements-Number of disagreements) / Number of agreements.In the second year, prior to collecting data for the main study for Study 2, an online data collection for a pilot test will take place with online apparel consumers in a computer lab. The purpose of this pilot test is to ensure that questions are clear and understandable. Based on participants' feedback, the wording of the questionnaire items will be altered.Data will be obtained from consumer panels of a market research company specializing in consumer survey methodology. The market research company has partnered with numerous sponsors that have large emailable databases of customers to create their consumer panels. The company will be contracted to recruit participants, so the PI will not be involved with recruitment of participants and data collection. At the time of this project, their panel is composed of three million members that mirror the U.S. population. The company will recruit panel members through an e-mail invitation-only process. An online survey will be created by the PI. Online panel members will be invited to participate in this online survey if they view themselves as meeting the stated qualifications.Considering the general response rate for online surveys (approximately 10%), the company initially will choose 5,000 individuals at random from their consumer panels. The company will then send out 5,000 invitations that will explain the qualifications of participation, provide a survey link, and a unique ID to targeted panelists. This ID will enable the company to track participants. The market research company will provide a "virtual currency" incentive to participants to complete the survey. This currency can be accumulated over time and then redeemed for items such as gift cards and travel vouchers.A voluntary informed consent form will be provided on the first page of the online survey. Before completing this online survey, respondents will try to interact with AI-based customer service and content (i.e., video ad chatbots; chatbots on social media messenger). Then, respondents will complete a self-administered questionnaire online. It will take 7-10 minutes for respondents to complete this survey. Online apparel consumers in the U.S. (n = 500) will be asked about perceptions of quality, customer experiences, brand engagement, and brand loyalty via an online survey. Scales used in data collection will be adapted from existing reliable measures. Participants will respond to items using 5-point Likert scales ranging from 1 = strongly disagree to 5 = strongly agree.After the data collection, the PI will obtain an Excel file containing participants' responses from the market research company. Then the PI will analyze data using SPSS and AMOS. Description, percent and frequencies will be used to examine demographic information. A confirmatory factor analysis will be conducted to assess composite reliability and construct validity (i.e., convergent and discriminant validity). Structural Equation Modeling via AMOS will be used to analyze the measurement and structural relationships among variables. According to a 95% (p < .05) significance rate, the hypotheses will be supported or rejected. Convergent validity will be supported by the following: (1) all loadings are significant (p < .001), (2) the composite reliability for each construct exceeds the recommended level of .70, and (3) the average variance extracted for each construct fulfills the recommended benchmark of .50 (Hair et al., 1998). As evidence of discriminant validity of the scales, none of the confidence intervals of the phi estimates will include 1.00.A research abstract and manuscript will be developed and submitted to one of the professional conferences. A research presentation will be developed for that conference. The manuscript will be submitted to and will be published in one of the peer-reviewed journals. This journal article will be disseminated to apparel retailers in Hawaii.