1. Introduction

To better connect customers and improve brand communication, the fashion industry is using Visual Generative AI more and more in personalized marketing. Visual Generative AI offers a novel approach by creating tailored content that resonates with individual preferences. Fashion, as a means of personal expression [1], taps into the fact that consumers want to stand out today. Personalization improves this process by providing tailored content. The increasing importance of personalisation is evidenced by the 5%–15% increase in revenue and the 10%–30% increase in marketing efficiency within a single system [2].

This paper thus aims to explore the effectiveness of Visual Generative AI in personalized marketing during the brand-consumer communication process, exploring its impact on consumer satisfaction. Through analyzing existing literature reviews, and quantitative research and linking it to the key theory, this effectiveness is assessed through consumer satisfaction.

1.1. Definition

This section first explains brand communication, its scope, and its impact on consumer decisions by highlighting personalized marketing as a crucial strategy, defining personalization, and demonstrating its effectiveness. According to Ateke, B. W. [3], brand communication is a comprehensive strategy that combines all communication channels to manage the brand and consistently engage with target audiences through marketing, public relations, advertising, and personal contacts. During this process, brand communication builds a brand image and strengthens brand recognition by highlighting fundamental values. The ultimate goal is to boost consumer trust and brand influence, influencing consumer decisions and encouraging positive brand behavior [4]. Delivering customised material based on recipients' unique interests is known as personalization, the aim of reaching the right individual at the right moment with the correct message [5]. To deliver a better experience, this strategy makes use of insights from consumers' behavioral and personal data. It improves consumer engagement through social media interactions, reviews, and purchase information [2]. As a kind of brand communication, personalized marketing consistently provides targeted audiences with relevant material. Through content creation tools like OpenArt, Dalle-2, and Hugging Face—which generate new, meaningful visual content using image/shape recognition and natural language processing—visual generative artificial intelligence (AI) improves this personalization process in the fashion industry by improving brand communication and customer satisfaction [6]. Brands may use VIsual Generative AI to provide highly customized visual content that speaks to specific consumers, increasing the effectiveness of personalization in promoting engagement and satisfaction.

Generative AI has the ability to increase revenues in the luxury, fashion, and apparel industries by $275 billion over the next three to five years, according to McKinsey [7]. This paper attempts to understand how effective it is to provide new personalised content [8] that Visual Generative AI aids in the brand communication process by focusing on this rising trend and leveraging the fact that clothes and apparel are a known necessity and are needed on a day-to-day basis worldwide.

1.2. Research question

This paper examines the effectiveness of Visual Generative AI in personalized marketing within the fashion industry, proposing that by tailoring fashion systems to better align with customer preferences through an understanding of underlying algorithms, brands can enhance communication and ultimately improve consumer satisfaction.

2. Literature review

Comprehensive reviews have evaluated the effectiveness of Generative AI in enhancing brand communication through personalized marketing across various industries. While some highlight its positive impact, others question its effectiveness.

2.1. Visual Generative AI is an effective tool to aid in personalized marketing

In the fashion sector, this article suggests that Visual Generative AI works better than Text Generative AI. According to Bergin [9], when exchanging clear and unambiguous data, text-based media with a low media richness was favoured, based on the media richness theory (MRT). Text can explain style, but it cannot capture visual intricacies. Viewing a piece on a virtual model—possibly even themselves—benefits the consumer. This instantaneous visual feedback captures the appearance and feel of a fashion item and helps shoppers make decisions. Because of this, this paper examines how well Visual Generative AI supports personalised marketing, recognising that it can be applied in various fashion industry processes, such as from designers to brands during the design phase or from brands to consumers during the personalisation phase.

The AI-driven approach has two modules for the design process: rendering and sketch generation. The rendering module uses a multi-conditional characteristic interaction module to synthesise several sketches created by the sketch-generation module with configurable textures to create ready-to-wear products [1]. The added value of Visual Generative AI is that it can speed up the process and broaden the designer's perspective by ingesting a lot of data, such as trends, previous collections, and customer interests, to generate ideas for designers [7].

Moving on to the personalization process. The use of Visual Generative AI in brand communication by numerous top businesses is evidence of its growing popularity and efficacy as a personalised marketing tool. There are two types of personalised marketing: creation (where Nike uses AI to scan athletes' feet and analyse their data to create personalised designs that improve performance) and recommendation (where Stitch Fix uses AI algorithms to analyse customer profiles and preferences to curate personalised selections) [10].

Acknowledging the various Visual Generative AI tools present, these tools serve different purposes during this personalization process, tailoring to different stages in the fashion industry.

Through understanding the diverse Visual Generative AI tools, aided by examples of brands’ utilization, this paper proposes that Visual Generative AI is thus an effective tool that aids in personalized marketing.

2.2. By understanding the algorithms, brands can use Visual Generative AI to develop fashion systems tailored to their target audience

Though acknowledging the diverse Visual Generative AI tools present, brands have different aims and target audiences; no single tool can specifically cater to their brand communication. To maximize the usefulness of these tools, brands need to understand the backend algorithms so that they can create personalized fashion systems specifically tailored for their target audiences, achieving their brand objectives.

Low-level pixel computing, mid-level fashion comprehension, and high-level fashion analysis comprise the three stages of the Visual Generative AI tools' fundamental idea [8].

Beginning with low-level pixel computation, which includes human segmentation, landmark identification, and posture estimation [8]. In order to efficiently classify clothing products, styles, and patterns, this entails the automated study and recognition of fashion aspects inside photographs. By recognising body shape and dimensions, landmark detection improves the virtual try-on experience [13]. This stage creates the virtual user figure and sets the stage for personalization.

The analysis proceeds to the fashion items once the user's virtual self has been created. In order to extract detailed information about clothing and fashion elements, such as garment types, collars, necklines, patterns, and textures, one must analyse and segment images or videos. This is the focus of mid-level fashion comprehension [8]. Detecting and adding fashion features to the system's database that may be virtually tried on and recommended for consumers in the last step of high-level fashion research; this improves the brand's comprehension of fashion data. Trend predictions and tailored recommendations are part of this phase [8]. Item retrieval uses visual content, trends, body shape analysis, and postures to search for visually comparable things in order to accommodate a wide range of consumer preferences [13]. This process yields more accurate insights for personalised style and recommendation systems.

Notwithstanding these successes, problems persist. Since many of them are not open-source, it is still challenging to guarantee access to high-quality, diverse datasets, which makes model inclusivity and resource sharing more difficult. For real-time applications, fashion synthesis models need to be more efficient, which calls for algorithm optimisation. Improving multimodal integration will facilitate smooth communication between text, images, and sketches in design and recommender systems, resulting in outputs that are more precise and customised [13]. Creating interpretable systems will boost confidence and enable improved comprehension and customisation of produced outputs as AI models grow more complicated.

Generative AI has proven the ability for new possibilities in the design process and speed up content production processes [14]. Brands can therefore increase their chances of creating a distinctive fashion system and staying ahead of the curve by learning how the algorithms operate.

2.3. Personalization enhances brand communication through the brand addressing consumers' individual needs and consumers' increased engagement with the brand, increasing consumer satisfaction

A successful fashion system, as mentioned above, will only be effective if consumers are satisfied. As brand communication is two-way, this paper proposes that personalization increases the chance of this success through increased consumer interaction with the brand, as well as increased understanding of consumers’ individual needs by brands. Through this interactive communication process, satisfaction thus arises as consumers feel valued.

When customers see that their problems are being addressed and when they have an easy time making decisions, they feel appreciated. Customers often struggle with decision-making during the purchasing process and experience remorse over their choices. As a result, they look for knowledge on their own from a variety of platforms to help them handle these problems, which might result in information overload and poor decision-making [15]. Therefore, brands' ability to effectively communicate with customers helps to mitigate this unhappiness, and personalisation makes this possible. Personalisation expedites the information-seeking process by attending to individual requirements and interests. A communication strategy that shows an awareness of the unique demands of each customer is known as personalised marketing [16]. This understanding is a type of effective communication since it influences client engagement behaviours in a positive way [15]. Consumers are more likely to feel appreciated and understood by a brand when marketing messages and services are personalised, and this can increase customer satisfaction [17].

The above addresses the communication from brand to consumer. Consumer-to-brand communication, or engagement [15], can strengthen the bond between a company and its customers. Engagement refers to the mindset, actions, and degree of kinship that exists between consumers and brands. One way that customers engage with this customised fashion system is through interaction. When technology functions on a human level, it forges a charming connection with the consumers since AI has transformed expert systems from being dependent on rules to being data-driven and based on deep learning [15]. Customers who engage actively with information through personalisation are more likely to exhibit positive attitudes and behaviours, engage at higher levels, and feel more in control of their experience.

One of the main factors impacting customer satisfaction is effective brand-consumer communication. According to Chhabria, S. G., Gupta, S., & Gupta, H. [17], consumer satisfaction is the degree to which a product or service fulfils or goes beyond the expectations of the consumer. It is essential because it creates and preserves a foundation of devoted clients. Personalised marketing has grown in significance as more companies use digital platforms to engage with their clientele. This is because customised material may create a stronger emotional bond between customers and brands. Brands can more effectively customize their marketing messages and increase customer satisfaction by getting to know each client's tastes and demands. This also increases the chances that customers will interact with the brand [17]. Using this concept back to the fashion industry, by providing personalization on clothing through Visual Generative AI, it provides a more engaging experience as consumers can personally interact with the screen and potential models of clothes based on their virtual selves.

3. Theoretical framework

The HII Theory of Interactive Media Effects (TIME) examines how media influence users’ attitudes, behaviors, and interactions. The many mechanisms behind the action route are explained by the agency model of customization, the motivational model, and the interaction effects model [18].

Virtual try-ons, customised recommendations, and dynamic content creation are examples of AI medium features that improve user experience by offering interactive and customised interactions that traditional media cannot [19]. Visual cues, such as graphics and interactive features, affect users' perceptions and behaviours and provide pertinent information to help them make decisions [20]. When users are able to personalise their fashion choices and obtain recommendations that are tailored to their preferences, they feel more empowered. After that, users rely on cognitive heuristics, like the availability heuristic, which makes them believe that goods that are frequently recommended are better.

Users' perception and behaviours are greatly influenced by how they view the AI medium. Users are more likely to engage with information and trust recommendations from AI when it is perceived as trustworthy and capable of understanding preferences [21], which increases user pleasure. Perceptions like this impact activities like virtual catalogue browsing, virtual try-ons, and fashion item customisation [22]. A high level of engagement strengthens customer pleasure and brand affinity, strengthening the bond between the two.

Mutual augmentation is the term used to describe how interaction can synergistically improve system and user capabilities [23]. AI systems improve user capabilities in the fashion sector by providing sophisticated tools for personalisation and fashion discovery. As a result, there is a two-way communication process where user input and interactions enhance the AI's capacity to make pertinent and correct recommendations. Users interact with AI systems to make personalised recommendations and purchase decisions; hence, the effectiveness of these interactions is critical to user engagement and happiness. An AI system that is responsive and easy to use improves user satisfaction by encouraging a social dialogue in which suggestions are customised based on user feedback. Because they feel that the brand understands them, this relationship fosters trust and brand loyalty.

When consumers can tailor their wardrobe selections and get recommendations that suit their tastes, their sense of agency is increased. High levels of engagement are generated by these interactive experiences, which entice consumers to stay on the platform longer, look at a wider variety of goods, and develop a stronger bond with the brand. Users who feel more empowered have higher levels of satisfaction and a stronger bond with the company.

4. Methodology

Quantitative research, a survey is used for this study to analyze the users' cognition, behaviors, attitudes, and needs of consumers on the role of Visual Generative AI in personalized marketing in the fashion industry.

4.1. Sampling procedure

The study targeted consumers aged below 18 to 65, including both professionals and non-professionals in the fashion industry. Convenience and snowball sampling, in particular, were two non-probability sampling techniques used. The survey link and QR code were shared within the researchers' WeChat friend circles, and researchers also distributed the survey to their parents and friends, encouraging them to share it further, generating 291 responses.

4.2. Measures

The questionnaire items and scales were adapted from previous research, as referenced in the Applied Economic Research and Trends (AERT) report [24]. This survey has four sections: the first gathered demographic information (age, gender, occupation, education level); the second explored participants' views on Visual Generative AI; the third examined personalized marketing and its impact on consumer satisfaction; and the fourth assessed consumer engagement, focusing on specific features and information privacy.

A research framework is thus created based on the HII Theory of Interactive Media Effects (TIME), along the main line of "cognition-behavior-attitude-need," to capture comprehensive insights, seeking to understand

1. Cognition: Awareness and understanding of participants on AI participation in apparel customization services

2. Behavior: Frequency and purposes participants use visual generation AI in apparel customization.

3. Attitude: Acceptance, willingness to pay and share personal information (e.g., body shape, fashion preferences) with the AI

4. Needs: Desired features in AI-powered fashion personalization tools

4.3. Findings

Visual Generative AI is widely recognized, with 86.6% of participants aware of it, yet its usage is limited, with only 30% having utilized it. Despite this, over 70% of participants believe Visual Generative AI will be valuable in the fashion industry, and three-quarters expect personalized fashion services in the future. Currently, only 20.96% have used it, with the majority citing high costs as a concern.

Respondents anticipate features like virtual fitting and preference-based style suggestions. While 80% are willing to pay extra for personalized services, most prefer additional costs to be under 20%. However, consumers remain cautious about sharing personal information, with less than 40% willing to do so without reservations.

5. Discussion

The results indicate that users hold positive expectations for Visual Generative AI in personalized marketing within the fashion industry.

Aligned with the HII Theory of Interactive Media Effects (TIME), Visual Generative AI enhances the interaction between consumers and brands by offering key affordances such as action and interaction. This technology enables brands to generate personalized content, which improves communication, increases engagement, and strengthens the brand-audience connection. TIME illustrates how Visual Generative AI supports personalization and real-time interaction in fashion, offering tailored content that enhances user satisfaction and brand involvement.

The integration of Visual Generative AI with personalized marketing aligns with fashion’s visual nature, where personal taste is crucial. By using image recognition and natural language processing, AI creates customized content that reflects personal preferences. This allows consumers to visualize how items look on virtual models or themselves, providing immediate feedback that aids decision-making. Personalized marketing thus enhances users' sense of being valued by the brand, boosting satisfaction and showcasing the effectiveness of Visual Generative AI.

TIME theory underscores the importance of personalization and user control in media effects. It is anticipated that Visual Generative AI will become increasingly prevalent in fashion as brands leverage this technology for personalized recommendations and virtual try-ons. This approach enables brands to align their communication strategies with consumer preferences, fostering satisfaction and positive brand behavior.

Adoption of AI depends on trust [25], and purchase intentions can be positively influenced by favourable attitudes towards technology [26]. Even though more than half of the participants were willing to share their preferences, 10.5% of non-professionals expressed hesitancy due to privacy concerns. Companies should ensure transparency about data usage, give customers control over their information, and establish strong data protection measures to increase technology familiarity and support the adoption of AI in fashion personalisation [27].

6. Conclusion

This study demonstrates the potential of Visual Generative AI in advancing personalized marketing strategies within the fashion industry. The findings reveal that while there is substantial awareness and positive consumer sentiment toward the technology, actual adoption remains constrained by challenges related to cost and data privacy. Nevertheless, the ability of Visual Generative AI to produce tailored, consumer-specific content suggests a significant opportunity for brands to enhance engagement and satisfaction. Future research should focus on overcoming the barriers to adoption and further exploring the ethical implications of data usage in personalized marketing. By addressing these challenges, the fashion industry can fully harness the capabilities of Visual Generative AI to create more personalized and impactful consumer experiences.

Acknowledgement

Runrun Zhang, Xuanheng Pan, Yanze Shi and Kaijie Gui contributed equally to this work and should be considered co-first authors.