Designing Online Virtual Advisors to Encourage Customer Self-disclosure: A Theoretical Model and an Empirical Test

ABSTRACT

Virtual advisors (VA) are tools that assist users in making decisions. Using VAs necessitates the disclosure of personal information, especially when they are employed in personalized contexts such as healthcare, where disclosure is vital to providing valid and accurate advice. Yet, extant research has largely overlooked the factors that encourage or inhibit users’ from disclosing to VAs. In contrast, this study investigates the determinants of users’ intentions to self-disclose, and examines how VAs can be designed to enhance these intentions. The results of a study in the context of skin care advice reveal that the intention to disclose to a VA is not only the product of a rational process, but that perceptions of the VA and the relationship with it are important. The results further show that a parsimonious set of design elements can be used to endow a VA with desired characteristics that enhance the willingness to disclose. The study contributes to our understanding of the factors influencing users’ intentions to provide personal information to a VA, which extend beyond the expected benefits and costs. The study further demonstrates that social exchange theory can be applied in contexts in which humans are interacting with automated VAs.

KEYWORDS:

• virtual advisor
• recommendation agent
• decision aid
• online disclosure
• online decision making

Acknowledgment

We acknowledge the financial support provided by The Natural Sciences and Engineering Research Council of Canada

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed on the publisher’s website.

Notes

1. While the results presented in the main body of the paper focus only on the “covariation of interests” dimension of interdependence, the results reported in the appendices account for all three dimensions.

2. We further propose that transparency is multi-dimensional. Its dimensions concern knowing why information is being solicited (purpose transparency), and knowing how the information solicited will be used (process transparency). While the analysis presented in the main body of the paper utilizes an overall measure of transparency, the analysis contained in the appendices consider both dimensions of transparency.

3. At this point, it is important to note that the effects of increased transparency on perceived benefits are contingent on the quality of the advisor’s reasoning. When holding everything else constant, a well-designed VA that offers additional insights into its processing will exhibit more benefits than an advisor that is less transparent.

4. The analysis presented in the main body of the paper uses an overall measure of trust. The three trusting beliefs were also measured [60; 82]. The results when using these measures are discussed in Appendix G.

5. We focus on the “covariation of interests” dimension of interdependence. However, the other dimensions of interdependence were also measured. Similarly, the analysis uses overall measures of responsiveness and transparency. However, three new scales were also developed to measure the dimensions of responsiveness (understanding, caring, and validating), and two new scales where developed to measure the dimensions of transparency (purpose and process). The results considering the sub-dimensions are described in Appendix G.

6. The same item had a loading higher than 0.7 when the measurement model was assessed in PLS, or when a factor analysis was performed in SPSS. Nevertheless, the item was dropped to be conservative.

Additional information

Notes on contributors

Sameh Al-Natour

Sameh Al-Natour is an Associate Professor of Information Technology Management at Ted Rogers School of Management at Ryerson University, Canada. He received his Ph.D.. from the University of British Columbia and his M.B.A. from Simon Fraser University, both in Canada. His research focuses on the design and evaluation of human-computer interfaces, and the adoption and use of information technology.

Ronald T. Cenfetelli is the CANFOR Chair in MIS and Professor, Accounting and Information Systems at the University of British Columbia’s Sauder School of Business. He is also a Senior Editor at MIS Quarterly. He holds a Ph.D. from the University of British Columbia and an M.B.A. from Indiana University Dr. Cenfetelli’s research interests are in e-business, the behavioral and emotional aspects of technology usage, and psychometrics.

Izak Benbasat

Izak Benbasat is Professor Emeritus at the Sauder School of Business, University of British Columbia. He received a Ph.D. from University of Minnesota and Doctorat Honoris Causa from Université de Montréal. Dr. Benbasat is a member of the Order of Canada, a Fellow of the Royal Society of Canada, and a member of the Science Academy (Bilim Akademisi), Turkey. He received the LEO Award for Lifetime Exceptional Achievements from the Association for Information Systems and was conferred the title of Distinguished Fellow by the INFORMS Information Systems Society. He served as the editor-in-chief of Information Systems Research, Editor of the IS&DSS Department of Management Science, and a Senior Editor of the MIS Quarterly.