The increasing urgency of sustainable consumption, particularly in clothing manufacturing, necessitates understanding how to influence consumer choices. This study focuses on the online purchase of sustainable clothing, exploring the roles of visual information and voice assistants in shaping consumer behavior. Existing research highlights various factors influencing sustainable clothing purchases, including design, production, product information, consumer knowledge, attitudes, and purchase behavior. While consumers are attracted to eco-friendly, ethical, and organic clothing, sustainable attributes alone may not always drive purchasing decisions. The COVID-19 pandemic further accelerated the shift to online shopping, increasing demand for sustainable products. Social media and peer influence also play a role. The study investigates multisensory perception's impact on purchase behavior, leveraging the capabilities of artificial intelligence (AI) in the form of voice assistants. These assistants offer personalized service, quick responses, and hands-free convenience, potentially enhancing the shopping experience and influencing purchase intent. However, consumer comfort levels with AI-driven shopping interfaces and trust in virtual assistants remain a concern. This study aims to determine the influence of visual and voice cues on purchase behavior, contributing to multisensory marketing and consumer behavior research. It further explores the moderating effect of consumer knowledge on sustainability issues and identifies potential challenges for online retailers in implementing these technologies.

The literature review examines the impact of AI and multisensory marketing on consumer behavior. AI, through visual information and voice assistants, enhances personalized services and influences purchase intent. Research shows that packaging design, symbols, imagery, and sounds significantly affect consumer attitudes, emotions, and behavior. Multisensory perception influences purchase behavior and apparel evaluation. Voice assistants provide unique shopping experiences and meet consumer needs. However, consumer comfort with AI interactions is variable, with concerns about affective and cognitive issues arising from incongruent conversations. The role of attitudes toward sustainable clothing is also explored, showing a significant impact on purchase intention and behavior. Positive attitudes influence willingness to pay a premium for sustainable products. However, a gap often exists between attitudes and actual purchasing behavior due to factors like cost and lack of knowledge. Visual information is a crucial factor, influencing purchase intention and behavior through intrinsic and extrinsic cues. Social media further influences perception and purchasing behavior. The perceived intelligence of voice assistants is also examined, with research indicating their potential to positively influence consumer attitudes and behavior. However, limitations exist in the acceptance of AI due to algorithmic limitations. Consumer knowledge about sustainability issues is a crucial factor that influences purchasing behavior. Consumers with greater knowledge are more likely to have positive attitudes and intentions, but lack of information can impede purchase decisions.

Data was collected online using the Sojump platform, with 2656 valid responses from individuals in Shanghai, China, who had completed higher education. The study employed a five-point Likert scale to measure constructs, including visual information (VI), perceived intelligence of voice assistants (PIVA), customer knowledge of sustainability issues (KSI), positive attitude towards sustainable clothing (ATSC), and purchase behavior (PB). The questionnaire included items adapted and revised from previous studies. Data analysis involved several steps using SPSS Statistic 23.0 and Amos 23.0 software. First, reliability and validity were assessed using Cronbach's Alpha, factor loadings, composite reliability (CR), average variance extracted (AVE), and Kaiser-Meyer-Olkin (KMO) measures. Convergent validity and model fit were evaluated using confirmatory factor analysis. Multiple regression analysis examined relationships between variables, with variance inflation factor (VIF) values used to check for multicollinearity. Mediation analysis using Process SPSS investigated the mediating role of ATSC, and moderation analysis examined the moderating role of KSI on the relationship between VI/PIVA and PB, mediated by ATSC. The study’s sample characteristics included demographics such as gender, age, education level, occupation, and monthly clothing expenditure (detailed in Table 1). The reliability and validity of the measuring instruments were assessed (Table 2 and Table 3), ensuring the data's suitability for analysis.

The study’s key findings are based on the results of reliability and validity tests, multiple regression analysis, mediation analysis, and moderation analysis. All Cronbach’s alpha values exceeded 0.8, demonstrating excellent internal consistency, and confirmatory factor analysis indicated a good model fit (Table 2 and Table 3). Multiple regression analysis revealed significant positive relationships as hypothesized: (H1) ATSC positively impacts PB (β = 0.230, p < 0.001); (H2) VI positively impacts ATSC (β = 0.334, p < 0.001); (H3) VI positively impacts PB (β = 0.295, p < 0.001); (H4) PIVA positively impacts ATSC (β = 0.291, p < 0.001); and (H5) PIVA positively impacts PB (β = 0.185, p < 0.001) (Table 4). Mediation analysis confirmed that ATSC mediates the relationships between VI and PB (t = 15.126, p < 0.001), and between PIVA and PB (t = 17.159, p < 0.001) (Table 5 and Table 6). Moderation analysis showed that KSI moderates the relationships between VI and PB (LLCI = −0.087, ULCI = −0.018, p < 0.001), and between PIVA and PB (LLCI = −0.103, ULCI = −0.033, p < 0.001), both mediated by ATSC, supporting H6 and H7 (Table 7). The indirect effect (VI → ATSC → PB) was stronger with higher KSI levels (Table 8), and similarly, the indirect effect (PIVA → ATSC → PB) was also significant across KSI levels. The results consistently demonstrate the significant influence of visual information and perceived intelligence of voice assistants on purchase behavior for sustainable clothing, mediated by positive attitudes and moderated by consumer knowledge of sustainability issues.

The findings strongly support the hypotheses, indicating that both visual information and the perceived intelligence of voice assistants significantly influence consumer attitudes and purchase behavior toward sustainable clothing online. The mediating role of positive attitudes highlights the importance of creating positive perceptions of sustainable clothing. The moderating role of knowledge underscores the need for effective communication of sustainability information to consumers. The results align with previous research emphasizing the influence of multisensory marketing and AI-driven interactions on consumer behavior. The study’s implications are significant for online retailers and brands, suggesting that investing in high-quality visual content and developing intelligent voice assistants can enhance consumer engagement and drive purchases. The significant mediating role of ATSC emphasizes the importance of influencing consumer attitudes to promote sustainable purchasing behavior. The findings also highlight the need for providing consumers with sufficient knowledge and information to bridge the intention-behavior gap. By addressing these factors, retailers can effectively promote sustainable fashion choices in the online market.

This study provides valuable insights into the impact of visual information and voice assistants on consumer behavior in the online market for sustainable clothing. The significant influence of both factors, mediated by positive attitudes and moderated by knowledge, offers practical recommendations for brands seeking to improve consumer engagement and drive purchases of sustainable products. Future research could explore other sensory modalities, expand the geographic scope beyond China, and investigate age and gender differences in consumer responses. Additionally, further research could delve deeper into the design and implementation of voice assistants and visual presentation elements to maximize their positive impact on promoting sustainable fashion.

The study’s primary limitation is its geographic focus on China, potentially limiting the generalizability of the findings to other cultures. The sample's age range (predominantly under 41 years old) and gender distribution (slightly more female) might also influence results. The study focuses only on visual and auditory cues, neglecting other sensory modalities that could be impactful. Future research should address these limitations by replicating the study in different cultures, focusing on broader age and gender demographics, and examining the roles of other sensory stimuli in shaping purchase behaviors.