Alexa, it’s not just us! Voice commerce through the lens of service providers and consumers

Recent advances in artificial intelligence (AI) and natural language processing (NLP) have fueled the emergence of novel channels in electronic commerce (e-commerce). One prominent channel is conversational commerce, a subset of e-commerce characterized by interactions that occur through natural conversation—either by voice or text (Lim et al., 2022). Within conversational commerce, voice commerce specifically refers to interactions facilitated by voice-based technologies, such as voice assistants like Alexa, Google Assistant, and Siri (Mahr and Huh, 2022). Unlike conventional e-commerce, which predominantly relies on visual interfaces accessed through websites or mobile applications, voice commerce engages consumers through spoken dialogues, which, in turn, create experiences that closely resemble everyday human conversations (Zierau et al., 2023).

A report by Forbes reveals that a staggering 72% of voice commerce users engage with these tools daily for various functional tasks and commercial endeavors (Crail and Bottorf, 2022). Consequently, global e-commerce transactions via voice assistants alone have catapulted from US$4.6bn in 2021 to US$19.4bn in 2023 (Chevalier, 2023), thus making voice commerce a promising modality for firms to deliver their products and services. However, a PwC study underscores a noteworthy distinction, suggesting that while users are increasingly comfortable deploying voice assistants for information searches, they exhibit reservations when it comes to actual purchasing (Hayes and Wagner, 2018). Despite this, the rise of voice assistants marks a transformative shift in human–computer interactions, pivoting toward a more organic and impactful mode of conversational communication (Flavián et al., 2023; Guzman, 2019; Hearst, 2011; Nass and Brave, 2005)—an evolution that carries profound implications for e-commerce. Consequently, there is a growing interest among scholars to dissect the details of this modality, especially focusing on the challenges and opportunities inherent to voice commerce (Malodia et al., 2024; Rzepka et al., 2020; Zaharia and Würfel, 2021).

In Maity and Dass’ (2014) study, they juxtaposed consumer purchase processes and interactions across e-commerce, mobile commerce (m-commerce), and traditional in-store purchases. However, voice commerce introduces a distinct set of service dynamics, primarily due to the involvement of voice assistants — intelligent software agents capable of engaging in spoken dialogues with users (Kraus, 2019; Kraus et al., 2019). Voice assistants represent a specific type of conversational agent, alongside chatbots, which instead facilitate text-based interactions (Bérubé et al., 2024). Importantly, voice assistants differ fundamentally from the general concept of voice assistance, which broadly refers to any form of support provided through voice communication without necessarily involving intelligent conversational capabilities (Simms, 2019). Contrasting traditional e-commerce, the service encounter in voice commerce comprises the trio of the voice-powered conversational agent integrated with the service delivery platform, the service provider, and the customer (Elshan et al., 2022). The interactions stemming from this newly established service paradigm present unique challenges, particularly those related to symmetrical information exchanges across both AI-AI and human–AI interfaces. Crucially, these challenges influence the fulfillment of customer expectations and, consequently, adoption intentions (Lee et al., 2020).

Recognizing this complexity, Robinson et al. (2020) proposed a service encounter framework that accentuates the multi-actor nature of service engagements facilitated by voice assistants. This framework clearly underscores the knowledge gaps in human–AI engagements—termed as “interspecific” within their framework—in online services. To enhance the utility and acceptance of such technology, understanding the implications of these interactions on customers’ evaluations and their subsequent intentions—one of the most accessible and reliable proxies for predicting actual behavior—is paramount. Thus, this investigation endeavors to pinpoint attributes affecting the integration and utilization of voice assistants in voice commerce, with a keen focus on attributes affiliated with the services and service providers in voice commerce.

Despite a recent surge in attention toward voice commerce research, existing literature predominantly focuses on the role of voice assistant attributes (e.g. anthropomorphism, interactivity, usability) in influencing consumer behavior (Berriche et al., 2022; Dellaert et al., 2020; McLean et al., 2021; Moriuchi, 2019). However, limited attention has been directed toward attributes specifically related to the services and their providers. This oversight limits theoretical understanding and leaves important elements influencing consumer behavior underexplored. Therefore, to develop novel theoretical insights and clearly isolate the impact of these overlooked attributes, clear analytical boundaries are deliberately set in this study, explicitly excluding voice assistant attributes from the empirical scope.

Service attributes refer to features directly associated with the service offering itself, such as service complexity (i.e. the degree of difficulty involved in executing and delivering the service) and service consumption phase (i.e. the specific stage—pre-purchase, during-purchase, or post-purchase—at which consumers interact with the service). Prior research indicates consumer behavior in e-commerce significantly varies based on product and service characteristics, including tangibility or intangibility and the complexity required in the consumer’s decision-making process (Lovelock, 1983; Phau and Meng Poon, 2000; Vijayasarathy, 2002). Service provider attributes, on the other hand, encompass attributes associated with the firm delivering the service, including brand familiarity (i.e. consumer’s prior recognition and knowledge of the brand) and multichannel integration (i.e. seamless integration of voice commerce with other service delivery channels).

Further evidence that scholarly understanding regarding how service and service provider attributes shape consumer purchase intentions remains notably limited is presented in Table 1. For example, Fernandes and Oliveira (2021) examined user perceptions of interaction quality with voice assistants, whereas Lu et al. (2020) investigated general technology acceptance attributes, yet neither sufficiently addressed the influence of specific service offerings or their providers’ attributes on consumer purchase decisions in voice commerce. Such oversight is notable because voice commerce inherently involves consumers, service providers, and voice assistants (i.e. the technological platforms through which service is rendered), each possessing distinct yet interrelated attributes influencing consumer behavior (Lim et al., 2022; Rabassa et al., 2022). This gap is especially critical in voice-based interactions, where the absence of visual and tactile elements significantly alters consumers’ perceptions of service complexity, risk evaluation, and trust toward service providers (Rabassa et al., 2022). While prior studies highlight voice assistant attributes such as anthropomorphism and interactivity as vital for initial user engagement, they neglect the critical relationship these attributes share with the provider’s brand image or multichannel integration of services. This omission is problematic because reduced visual and tangible cues in voice commerce heighten the importance of brand familiarity and seamless integration for consumer reassurance and satisfaction (Tassiello et al., 2021; van Doorn et al., 2017).

The present study addresses these identified theoretical gaps by offering three distinct yet interconnected theoretical contributions aligned with MacInnis' (2011) explication of conceptual contributions. First, this study identifies and delineates novel concepts in voice commerce—namely, service attributes (e.g. service complexity and service consumption phase) and service provider attributes (e.g. brand familiarity and multichannel integration)—which have been largely overlooked in prior literature. In explicitly defining and situating these concepts within voice commerce, this study enriches existing understanding of service delivery and consumer interaction with the digital platforms that support this delivery. Second, this study provides an integrative conceptual framework that holistically encapsulates these service and service provider attributes along with consumer attributes, in contrast to prior research, which often considered these attributes partially or separately (Malodia et al., 2023; Moriuchi, 2019). This integrative approach advances theoretical understanding by highlighting that voice commerce is inherently a multi-actor phenomenon involving interactions among multiple attribute categories, rather than focusing exclusively on consumer or voice assistant attributes, as seen in prior research (Dellaert et al., 2020; Lu et al., 2020; van Doorn et al., 2017). Third, through elucidating explicit relationships among these attribute categories, this study’s conceptual framework provides theoretical clarity and comprehensiveness in explaining how these attributes collectively influence consumer purchase intention in voice commerce. In this way, this study’s contribution moves beyond identification and classification of relevant constructs by specifying precise interrelationships, thereby enriching scholarly understanding and setting the stage for future empirical investigations. Understanding these attributes is not merely academic but also essential for firms to optimize their use of voice assistants to enhance consumer engagement and satisfaction, and by extension, improve their competitive positioning and drive sales in an increasingly digital marketplace. Without this knowledge, firms risk misaligning their voice-commerce strategies with consumer preferences and usage patterns, leading to suboptimal service delivery, missed opportunities for engaging consumers effectively, and shortfalls in fully harnessing the value generated by voice commerce.

This study begins by asking the research question “What key service attributes, service provider attributes, and consumer attributes influence purchase intentions in voice commerce?” (RQ₁) and then follows up with the research question “How does the interplay between these attributes influence purchase intentions in voice commerce?” (RQ₂) to deliver both breadth and depth in insights. To comprehensively address these research questions, it is essential to consider the evolving landscape of voice commerce. As voice commerce continues to evolve and dominate e-commerce (Lim et al., 2022), understanding the various attributes influencing consumer behavior becomes paramount (Lim et al., 2023). Despite extensive research focused on its technological attributes, a significant gap remains regarding the attributes of the service and the service provider. This gap is underscored by the Journal of Service Management’s call for a special issue on the “rise of voice conversation capabilities in smart service systems.” In recognition of this gap, we conducted in-depth interviews with both service providers and consumers, with the goal of uncovering a comprehensive perspective of the attributes that sway purchase intentions in voice commerce. This approach distinguishes the present study from prior ones, which have typically focused on individual attribute categories in isolation (Lim et al., 2022), resulting in less comprehensive insights.

In addition, while recent studies indicate that consistent utilization of conversational agents can enhance a brand’s image (Hernández-Ortega et al., 2022; Rzepka et al., 2022), the present study is distinctly centered on third-party providers delivering services via voice assistants. This focus is critical because these external providers must ensure that their offerings can seamlessly integrate within platforms they do not control, thereby posing distinct challenges that remain underexplored in the existing literature. Recognizing the scant research dedicated to the attributes of both services and service providers in voice commerce, the present study’s approach aligns with Braun and Clarke’s (2006) and Lim’s (2025) recommendations for an inductive research methodology. Through in-depth interviews with service providers and consumers, this study identified novel attributes relevant to three critical categories: the service, the service provider, and the consumer. These findings reveal the key attributes that shape purchase intentions in voice commerce and spotlight interaction dynamics between these attribute categories. Therefore, this study endeavors to enrich the nascent field of voice commerce by shedding light on these attributes, which mold consumer behavior toward purchasing via voice commerce.

The rest of this article is structured as follows: Subsequent to this introduction, an extensive review of the literature pertaining to voice commerce is pursued, which reveals precedents influencing diverse consumer behavior. Following this literature review, the research methodology is articulated, which encapsulates both data acquisition and its subsequent analysis. A detailed explanation of the findings is then presented. Subsequently, the conceptual framework that has emerged from this exploration is dissected, and its implications for both academia and industry are discussed. The article concludes with a reflection on the limitations inherent to the study while outlining potential avenues for future scholarly endeavors.

Voice commerce represents a two-sided market where voice assistants — such as Amazon’s Alexa, Apple’s Siri, and Google Assistant — mediate interactions between buyers and sellers (Rabassa et al., 2022). A voice assistant is a software-based dialogue system that can be accessed via smartphones or speaker devices (e.g. Amazon’s Echo) and is designed for voice-based interactions with its user (Guzman, 2019; Hoy, 2018; Kraus et al., 2019). The backend technology powering voice assistants relies on AI systems that use NLP in combination with other voice technologies like speech recognition, voice identification, text-to-speech, and speech synthesizers (Galanxhi-Janaqi and Nah, 2004; Kraus et al., 2019). Hence, these artificial agents can interpret commands, engage in appropriate actions, gather information (e.g. checking the weather), execute specific tasks (e.g. activating the air conditioner or coffee maker), and even facilitate product purchases (Rabassa et al., 2022).

Early research in human–computer interaction (HCI) and computer-mediated communication (CMC) provides a foundational understanding essential for the subsequent business applications of voice assistants. Specifically, HCI studies explored the effectiveness of speech- and text-based communication modes, revealing that while text-based interactions offered greater efficiency (Le Bigot et al., 2004, 2007), speech-based modes provided a more natural form of human communication (Novielli et al., 2010), surpassing even visual interfaces involving facial expressions and body language (Kock, 2004). As a result, voice-based interfaces are particularly valuable in contexts that require problem-solving or the exploration of solutions (Krämer et al., 2009), as they uniquely enable hands-free, real-time dialogue through spoken interactions. Unlike graphical or text-based user interfaces that require visual attention and manual interaction, voice-based interfaces allow users to perform tasks spontaneously, especially in situations where reading or typing is impractical (e.g. driving or multitasking). This advantage facilitates iterative, conversational exchanges, which significantly reduce cognitive load and accelerate decision-making during problem-solving tasks. Nonetheless, the inherent reliance on auditory cues exclusively constrains voice interfaces from effectively conveying detailed or comparative information, thus making them less suitable for highly complex or visually intensive tasks. Recognizing these specific strengths and shortcomings enables firms to accurately identify contexts where voice commerce solutions provide maximum value. Aligning voice-based technologies with scenarios in which they naturally excel, while acknowledging circumstances where visual or detailed information is essential, allows firms to enhance consumer engagement, optimize service delivery, and effectively leverage the unique benefits offered by voice commerce (Chattaraman et al., 2019; McLean et al., 2021; Moriuchi, 2019).

Early proponents of multichannel e-commerce have also suggested the potential of voice commerce to reduce call-center operating costs and improve customer service (Galanxhi-Janaqi and Nah, 2004). Much of the early research on voice commerce focused on the effects of text-to-speech technology on consumer satisfaction and trust in customer service (Chang and Jang, 2009; Qiu and Benbasat, 2005). However, the accuracy of speech recognition remained quite low until the recent development of sophisticated algorithms and AI, which have allowed users to engage in natural and complex dialogues with voice assistants, drastically expanding the scope of voice interfaces for commercial applications (Chung et al., 2017; McTear et al., 2016).

Interestingly, the potential of voice commerce is not being leveraged solely by providers of voice assistants (like Amazon or Google) but also by third-party firms that develop software functionalities—called Skills for Amazon Alexa and Actions for Google Assistant—that can be downloaded onto the voice assistants (Bahmani et al., 2022; Sabir et al., 2022). These functionalities enable customers to receive services offered by third-party firms through the voice interface of the voice assistant, making voice commerce a multi-actor context involving the voice assistant, the service provider, and the consumer. Voice functionalities are enabling service providers to partake in novel “service encounters,” where voice assistants augment or even replace the role of service employees (Kim et al., 2018; Larivière et al., 2017). Thus, voice commerce is now defined as a subset of conversational commerce, which is a subset of e-commerce, where voice technologies are used to facilitate, execute, and process business transactions (Kraus et al., 2019).

The subsequent sections offer an in-depth review of attributes tied to various facets related to voice assistant adoption, consumer behavior, as well as service and service providers in the context of voice commerce.

Research has primarily focused on the attributes influencing the use of voice assistants for e-commerce, often leaning on technology acceptance and use models (Davis et al., 1989; Venkatesh et al., 2003). Their adoption spans various domains, including content marketing (Reisenbichler et al., 2022), interactive advertising (Park et al., 2022), hospitality and tourism (Buhalis and Moldavska, 2022; Cao et al., 2022; Fan et al., 2022; Jiménez-Barreto et al., 2023), logistics and transportation (Hsiao and Chang, 2019), retail shopping (Aw et al., 2022; Kautish et al., 2023), and service engagements (Malodia et al., 2023). Noteworthily, the conversational interactions users have with voice assistants significantly enhance customer experience and wellbeing (de Oliveira et al., 2023; Kang and Shao, 2023) while current studies centered on online shopping have explored the impact of using voice assistants on facets of consumer behavior, such as brand engagement, patronage, purchase intention, satisfaction, and loyalty (Berriche et al., 2022; Chattaraman et al., 2019; King et al., 2022; McLean et al., 2021). These studies underscore that due to the conversational nature of interactions with these agents, users find them intuitive and valuable for brand engagement and online shopping tasks (Chattaraman et al., 2019; McLean et al., 2021; Moriuchi, 2019), wherein their perceived usefulness encourages customer engagement, subsequently amplifying customer loyalty (McLean et al., 2021; Moriuchi, 2019).

Relying solely on auditory and verbal interactions with voice assistants, however, may impose cognitive strain on users, demanding more effort to comprehend and interpret the agent’s feedback (Berriche et al., 2022). Yet, the recent technological evolution has ushered in voice assistants integrated with screens, allowing users to also utilize their touch and visual senses, thereby easing the overall cognitive load (Balakrishnan et al., 2024). Notable examples encompass devices like Amazon’s Echo Show and Google’s Nest Hub. Regardless of screen integrations, these agents typically adopt human names (e.g. Alexa, Siri) and communicate using human-like voices. Such anthropomorphic traits offer both functional and pleasure-driven benefits to users—key drivers for voice commerce adoption (Aw et al., 2022; Malodia et al., 2024; Rzepka et al., 2020). Beyond enhancing adoption and effectiveness, anthropomorphism fosters a sense of social presence in users, which makes them feel as if they are interacting with a human-like entity (Fernandes and Oliveira, 2021; Guzman, 2019; Lee, 2004; Thellman et al., 2016; van Doorn et al., 2017). This perception frames voice assistants as helpful, proficient, and intelligent, thus increasing their overall appeal (McLean et al., 2021; van Doorn et al., 2017). Moreover, services provided by such anthropomorphic agents are perceived as more responsive, appealing, and personalized to individual needs (van Doorn et al., 2017).

The quality of interaction with voice assistants is further heightened when they exhibit socially acceptable behaviors and emotions (Fernandes and Oliveira, 2021). A dialogue style that is more informal and relationship-driven fosters deeper trust and customer loyalty compared to a more rigid, task-focused communication style (Chattaraman et al., 2019). The agent’s interactivity is pivotal for voice commerce as it diminishes perceived risks and elevates the social presence, which, in turn, influences the customer’s intent to utilize these agents for online shopping (Lee et al., 2022).

It is worth noting that extant literature has extensively investigated the technological attributes of voice assistants, particularly anthropomorphism, interactivity, and usability, highlighting how these attributes influence consumer engagement and satisfaction (Berriche et al., 2022; McLean et al., 2021; Moriuchi, 2019). Although these technological attributes provide foundational insights into voice commerce, the present study intentionally shifts the analytical lens toward attributes less examined yet equally critical—the attributes of the services themselves and those of service providers. This strategic shift is justified by the unique attributes of voice commerce as a multi-actor service encounter, wherein the nature and complexity of services, the providers’ brand familiarity, and their capability for cross-channel integration critically determine consumer willingness to transact via voice. Through explicitly emphasizing these underexplored attributes, this study complements existing research by extending theoretical discourse beyond technological acceptance models and into the larger, multi-actor domain of voice-based service delivery.

Engaging with voice assistants, users might encounter psychological obstacles rooted in their values, needs, past experiences, and functional barriers relating to their perceptions about the effort and ease associated with using these agents (Berriche et al., 2022). Recent research, in light of these barriers (Beeler et al., 2022), has segmented conversational agent users into four distinct groups:

1. Active resistant users who consider the conversational agent as misaligned with their needs and, despite finding it easy to use, opt not to engage with it.

2. Passive resistant users who perceive the conversational agent as mismatched with their needs and difficult to operate.

3. Hedonistic adopters who recognize the conversational agent’s alignment with their needs but deem it complicated, hence they utilize it for elementary tasks.

4. Utilitarian adopters who recognize the conversational agent both aligned with their needs and user-friendly, extending their usage to regular purchasing.

When interacting with firms, consumers might either expect firms to genuinely prioritize their well-being or might lean toward a more transaction-based interaction (Clark and Mils, 1993). Those with a communal orientation would likely expect consideration and warmth from a voice assistant. This means that an agent’s social presence enhances its perceived warmth as well as the attractiveness, adaptability, and receptivity of the services provided (van Doorn et al., 2017). In contrast, exchange-oriented consumers would anticipate the agent to exhibit competence and expertise. This suggests that an agent’s social presence amplifies its perceived proficiency. In addition, consumers well-versed with technology view voice assistants as more adept and helpful, with services perceived as more enticing and personalized (Kendall et al., 2020; van Doorn et al., 2017).

Relational aspects, such as trust and rapport between the user and the agent, are deemed pivotal for the agent’s adoption (Fernandes and Oliveira, 2021; Hasan et al., 2021). Specifically, within voice commerce, consumer trust in the agent can influence choices, often leading them to opt for default product suggestions and shaping their purchase satisfaction (Mari and Algesheimer, 2021). Moreover, external societal influences, or the conviction that acquaintances and peers would commend their usage of voice assistants, can enhance user engagement, which, in turn, strengthens brand loyalty (Moriuchi, 2019).

Concerning risks in voice commerce, research has identified potential pitfalls like bias, privacy, and security (Rzepka et al., 2020), alongside the possibility of overreliance or compulsive behavior stemming from agent use (Ramadan, 2021). There is a perception that voice assistants could hinder product comparison by favoring their parent brand or affiliate brands, resulting in a potentially skewed product recommendation (Rabassa et al., 2022). Such biases could lead consumers to sense unfairness, especially if their preferences diverge significantly from the agent’s suggestions (Dash et al., 2022). The underlying technology, which often involves transmitting voice commands to cloud servers, raises privacy and security concerns (Bolton et al., 2021; Renz et al., 2022). Attributes like trust and privacy concerns have been observed to influence users’ overall voice commerce experiences, often mediating the effects of their inherent traits (Bawack et al., 2021; Hasan et al., 2021).

When utilizing voice assistants, consumers are increasingly prone to being swayed by sellers and their offerings. This vulnerability arises due to the rich personal data and purchase behavior logged by sellers when transactions are routed through these voice assistants (Dellaert et al., 2020). Consequently, it is crucial to scrutinize the characteristics of sellers and their offerings in relation to consumer behavior in voice commerce. Notably, this dimension remains under-researched with a striking absence of focus on the attributes of the seller or service provider and service offerings in the context of voice commerce.

Consumer behavior in e-commerce varies significantly based on product and service characteristics, particularly the tangibility or intangibility of the offering and the level of complexity or involvement required in the consumer’s decision-making process (Lovelock, 1983; Phau and Meng Poon, 2000; Vijayasarathy, 2002). Tangible products, defined by physical attributes, often necessitate visual inspection or tactile assessment—making voice-based interfaces less effective when consumers must closely evaluate features (e.g. color, style, texture), as is typically the case with clothing or electronics. In contrast, intangible or routine services (e.g. subscriptions, repeat orders) are better suited to voice commerce because they demand fewer visual confirmations and lower cognitive effort. While existing research has begun exploring how these differences shape consumer behavior (Sharma et al., 2024), substantial gaps remain. Early evidence suggests that consumers gravitate toward voice assistants for low-involvement, habitual purchases—such as household consumables—due to the convenience and immediacy of spoken commands (Tassiello et al., 2021), whereas non-transactional tasks (e.g. basic information searches) tend to be driven by contextual needs and social norms (Malodia et al., 2023; Moriuchi, 2019). High-involvement or visually complex products (e.g. electronics, luxury items), in contrast, frequently require detailed comparisons and visual inspections, making purely auditory interactions insufficient (Dellaert et al., 2020). Similarly, complex or high-stakes services—such as financial consulting or personalized healthcare—often discourage consumers from completing transactions solely through voice, instead preferring channels offering more comprehensive information and personalization (Malodia et al., 2023).

In terms of outcomes resulting from service delivery through voice assistants, previous studies demonstrate that platforms—whether apps, stores, or websites—featuring these agents are perceived as more engaging, dynamic, and user-friendly, which strengthens customer loyalty (Moriuchi, 2019; van Doorn et al., 2010). Although service offerings facilitated by these providers may appear more appealing and customizable, relatively few investigations focus on whether these favorable perceptions translate into purchases (Berriche et al., 2022; Malodia et al., 2023; Tassiello et al., 2021). Moreover, while a handful of studies explore how voice assistant attributes and consumer attributes interact (Malodia et al., 2023; Moriuchi, 2019), the multi-actor nature of these service encounters, encompassing consumers, service providers, and voice assistants, remains underexplored. Some researchers have called for deeper inquiries into these interaction effects (Grewal et al., 2022; Tassiello et al., 2021; van Doorn et al., 2017), which suggests a gap in understanding how service-side attributes (e.g. service complexity, service consumption phase, brand familiarity, multichannel integration) intersect with consumer attributes to drive purchases in voice commerce.

Our literature review highlights that research frequently proposes certain offering- and seller-related characteristics as strategic recommendations, yet seldom clarifies why and how these attributes influence decision-making in a voice-driven context. Our literature review highlights a twofold gap: (1) insufficient examination of how service-related attributes interact with consumer attributes to shape behavior in voice commerce, and (2) limited attention to purchase intention as a proxy of actual purchase as a distinct outcome in these settings. Table 1 reveals a disproportionate focus on consumer traits and voice assistant (technological) attributes, whereas service- and service provider-specific attributes—likely pivotal in determining perceived value and viability—have remained comparatively neglected. Although one might assume that a lack of research indicates low importance, we argue that the opposite is true here: the unique interplay of service providers’ brand familiarity, multichannel integration, service complexity, and service consumption phase in auditory-only transactions demands closer scholarly scrutiny, especially given the potential for misalignment when such attributes are overlooked.

Addressing the aforementioned gap holds significant theoretical and managerial relevance. Theoretically, concentrating solely on consumer or voice assistant attributes obscures the multi-actor ecosystem inherent in voice commerce, hindering a full understanding of purchase drivers. Through examination of service and service provider attributes, existing frameworks can be refined to encompass the understanding of how these attributes converge to influence purchase intentions among consumers. Managerially, ignoring these underexplored attributes can undermine strategic decision-making, particularly for firms offering complex or intangible services. Considering that voice commerce often functions as a dual-sided market, in which service providers list their offerings for potential buyers through AI-driven channels (Rabassa et al., 2022), a deeper grasp of service and service provider attributes can guide critical design choices—ranging from brand positioning to platform integration. Hence, this study explicitly addresses these shortcomings by focusing on the interplay of service and service provider attributes in shaping purchase decisions (Figure 1), thereby extending prior research beyond consumer or voice assistant attributes alone.

Extant research on voice commerce has predominantly employed conceptual, empirical, or experimental approaches to inform the development of voice-based systems (e.g. Dellaert et al., 2020; Galanxhi-Janaqi and Nah, 2004; Le Bigot et al., 2004; Novielli et al., 2010; van Doorn et al., 2017), many of which sought to establish or validate theoretical models (e.g. Baier et al., 2018; Berriche et al., 2022; Hernández-Ortega et al., 2022; McLean et al., 2021). Given that our study aims to explore the interplay among service attributes, service provider attributes, and consumer attributes within voice commerce—a phenomenon that remains underexplored—we chose to adopt an inductive approach using grounded theory (Charmaz, 2006; Gilbert and Stoneman, 2015; Glaser and Strauss, 2017; Urquhart et al., 2010). This approach enables a deep exploration of the contextual and socially constructed dimensions inherent in interactions between consumers and service providers who leverage voice assistants in e-commerce. Such an approach is especially suited to emerging fields such as voice commerce, where traditional quantitative (e.g. confirmatory, experimental) methods might overlook the dynamics of exchanges and interactions (Lim, 2025). Although technology acceptance has a well-established history in e-commerce (Chatterjee et al., 2002; Molesworth and Suortti, 2002) and information systems (Davis et al., 1989; Venkatesh et al., 2003) research, the field of voice commerce remains nascent and rapidly evolving (Lim et al., 2022), wherein its profound disruption of purchase decision-making and the limited number of exploratory studies necessitate an approach that can capture these complexities, taking into account voices from the ground. Our goal is, therefore, to construct a conceptual model that derives insights from a comprehensive depiction of discerned attributes and their interrelations, consistent with grounded theory (Wiesche et al., 2017) ^[1].

A targeted analytical scope was adopted, focusing exclusively on attributes of services, service providers, and consumers, while deliberately omitting voice assistant attributes from the analysis. This decision is justified by the extensive existing literature already providing rich insights into voice assistant attributes (Table 1) and by the critical need to illuminate the comparatively underexplored attributes associated with other actors involved in voice commerce. Although previous studies have examined consumer attributes, revisiting them here is necessary because their significance may shift or intensify in interactions uniquely mediated by service and service provider attributes within voice commerce. Concentrating on these neglected yet essential dimensions thus enables the identification of novel insights regarding the dynamics that distinctly shape consumer perceptions and purchase intentions within voice commerce. This targeted approach not only offers a valuable theoretical extension to prior research but also provides practical guidance for optimizing service design and platform integration strategies in this emerging field.

Our research is anchored in the Mexican market, a setting where leveraging computers and smartphones for e-commerce transactions has become both prevalent and normalized. In this ecosystem, prominent online retail giants like Amazon are evident, complemented by a myriad of smaller e-commerce entities. Although the marketplace is competitive, owing to its vast potential, vying solely on the basis of price has become untenable, especially in the face of the large players’ streamlined operations. Consequently, the only viable strategy left for new entrants is to distinguish their offerings, or in others, to engage in product differentiation.

Voice commerce emerges as a potential solution to offer a purchase experience that mirrors verbally requesting a family member at home to procure an item from a nearby store. Yet, the process of making a purchase in voice commerce diverges considerably from conventional e-commerce, primarily due to the lack of visual feedback. This requires substantive re-education, moving away from the now-established e-commerce transaction blueprint. Past studies have repeatedly evidenced that younger demographics exhibit a heightened tendency to embrace novel technologies (Lim et al., 2023), particularly when considering low-stakes product categories (Martínez-Domínguez and Mora-Rivera, 2020). Consequently, our study centers predominantly on this younger cohort of online shoppers, individuals who seamlessly integrate smart devices into their daily routines.

Our study concentrates exclusively on voice assistants that have achieved widespread smartphone integration. Our focus thus converges on agents such as Amazon’s Alexa, Apple’s Siri, and Google Assistant. While the breadth and depth of voice commerce functionalities varied among these agents, each presents basic voice commerce capabilities for their user base, spanning connected, smart devices.

Our study is situated in Mexico, a developing digital economy with a rapidly growing internet and e-commerce market. The choice of the Mexican market is underpinned by attributes such as market growth, technological adoption, and cultural aspects that make it an ideal setting for this study. This market has witnessed significant growth and transformation in e-commerce, driven by both private and public endeavors, mirroring the development of digital economies observed in other early adopters such as the USA (Palacios, 2003). In addition, cultural attributes play a crucial role in e-commerce adoption, and Mexico offers a representative example of a developing economy with a diverse population and a blend of traditional and modern values. This diversity allows us to explore the interplay between cultural attributes and new technology adoption. The Mexican market’s familiarity with e-commerce (Datta, 2011; Larios-Hernandez, 2024) sets a precedent for addressing voice commerce-specific challenges by leveraging existing digital infrastructure and consumer behaviors, making it a suitable setting for such studies.

For data collection, we employed a multifaceted methodology to encapsulate the perspectives of various stakeholders within voice commerce. Initially, we engaged in in-depth interviews with 15 business developers affiliated with e-commerce platforms in Mexico. This approach sought to explore their motivations steering the design and establishment of their voice commerce ventures. Unlike platform designers or IT managers, e-commerce platform business developers provide a holistic view, encompassing both strategic and operational aspects across diverse channels of their organizations. We explored their expectations concerning customer usage of their platforms and their strategy pertaining to the conception, development, launch, and expansion of this innovative business venture. Our preference for business developers emanated from their proximate interactions with customers and a comprehensive grasp of challenges customers encounter in routine e-commerce activities. While none of our participants’ firms perceived voice commerce as a direct e-commerce substitute, insights gleaned from e-commerce challenges proved invaluable for voice commerce design and development.

Next, we conducted in-depth interviews with 18 voice commerce users in Mexico, encompassing a spectrum of online shopping experiences from elementary reservation inquiries to tangible product acquisitions. These dialogues enriched our understanding of user motivations prompting transactions in specific scenarios and impediments in others. Given the innately personal and pioneering nature of chat and voice-driven commerce, innovative data gathering techniques were mandated, enabling users to candidly discuss experiences with their peers in a naturalistic setting.

To deepen our understanding of the adoption trajectories and barriers encountered by regular voice commerce users, we conducted six detailed triads by re-engaging previously interviewed users. In this context, triads refer explicitly to small, three-person focus groups—an approach well-established in qualitative disciplines such as sociolinguistics (Schiffrin, 1984) and nursing (Kendall et al., 2009). Triads promote dynamic yet intimate discussions, which allow participants to openly articulate, validate, or challenge one another’s experiences and viewpoints. This peer interaction facilitates richer, more authentic insights regarding shared attitudes and collective behaviors compared to individual interviews or larger focus groups. In particular, triads proved especially advantageous for examining voice commerce, as their intimate setting encouraged participants to comfortably share personal experiences, express reservations, and mutually validate perspectives about interactions with voice assistants—nuances that might otherwise remain undisclosed in alternative methodological formats.

The interviews were guided by an interview protocol that ensured consistency while allowing flexibility to explore emerging themes. The main topics covered in the interviews included participants’ experience with voice assistants and voice commerce, perceived benefits and challenges, attributes influencing adoption and use, strategies for designing and implementing effective solutions (for business developers), and future expectations and concerns.  Appendix Table A1 presents background information of the interview participants, including their roles, industries, experience levels, and interview durations. The interviews with business developers lasted between 60 and 120 min (average: 90 min) while user interviews ranged from 45 to 80 min (average: 70 min). The data collection process spanned four months. Tables 2 and 3 below provide a more granular overview of our sample and data collection strategy.

The collected written materials and interview transcripts underwent rigorous evaluation. Following transcription, qualitative data were analyzed using a grounded theory-based coding procedure consistent with established methodological practices in qualitative research (Corbin and Strauss, 2008; Gioia et al., 2013; Lim, 2025; Saldaña, 2015). We employed open coding to systematically analyze each data segment for its inherent meaning and assigning initial conceptual labels reflecting participants’ narratives (Charmaz, 2006; Corbin and Strauss, 2008; Lim, 2025; Saldaña, 2015). Throughout this analytical phase, we consistently compared individually coded segments across different transcripts and triads—a practice known as constant comparative analysis—to refine and verify emerging concepts (Lim, 2025). This iterative coding process was facilitated using the NVivo v.12 software. Subsequently, initial codes that shared conceptual similarities were grouped under higher-order conceptual categories—referred to as thematic coding—to identify broader, meaningful patterns (Gioia et al., 2013; Miles and Huberman, 1994). Any discrepancies in code interpretation among the research team were addressed through collaborative discussions, thus ensuring alignment and conceptual clarity. Our comprehensive inductive coding strategy aligns with the established guidelines by Corbin and Strauss (2008). A detailed breakdown of our coding classifications is provided in  Appendix Table A2.

Our analysis of the interview data revealed several key attributes and themes that shape the adoption and success of voice commerce. While attributes, such as perceived ease of use and perceived usefulness, have been previously identified in the literature on conversational agents (e.g. Rzepka et al., 2020; Zierau et al., 2023), our study provides new insights into their specific manifestations and implications within the context of voice commerce. In addition, we identify novel attributes and themes that emerge from the unique perspectives of business developers and users, highlighting the importance of considering multiple stakeholder views in understanding voice commerce. Our findings suggest that the adoption of voice commerce for procuring services is shaped by three fundamental elements: service attributes, service provider attributes, and consumer attributes. We dive into each in detail below.

Service complexity refers to the difficulty and variability associated with delivering a service, which can significantly influence a consumer’s choice of interaction channel and intention to purchase. This complexity encompasses not only the type of service but also how the service is delivered. For example, depending on the tangibility of the service, a consumer’s preference for one channel over another might shift. Both business developers and users acknowledged that certain types of products and services, such as those that are highly visual or require extensive comparison, may be less suited to voice commerce. As one business developer explained, “We have to be strategic about which products we push through conversational channels. Some things [services] just don’t translate well to voice or text-based interactions” (BD9, Transportation industry). Users also expressed hesitation about making high-stakes purchases, such as expensive electronics or financial products, through voice assistants, with one noting, “For big-ticket items, I still prefer to do my own research and talk to a human before making a decision” (U11, Regular user). The decision-making process to request, execute, and obtain services is tied to the service type. An interviewee illustrated this by saying, “For a restaurant reservation and you know what you’re going to order [I might just say] make my usual reservation and order my usual plate” (U2, Novice user). Yet, for high-involvement product categories, the decision is not as straightforward. Another interviewee mentioned, “If it’s a plane ticket, a lot of things can go wrong” (U6, Regular user).

We observed that for services impacting an individual’s personal appearance or well-being, like haircuts, travel, dining or medical attention, users would typically order via voice commerce if they had past experience with the service provider. Without this familiarity, users leaned more toward just researching with a voice assistant. However, for generic services not tied to an individual’s personal image or body, the feedback was more favorable toward voice commerce. One participant remarked, “For something like Netflix, there’s little room for error. Whether it’s three screens or five, there’s not much else to it” (U11, Regular user), while another stated “I’d use it for Coursera, but for a physical university, I’d research online, consider my options, then maybe make a call or visit the campus in person” (U9, Experienced user).

Service complexity, therefore, plays a crucial role in determining the suitability of voice commerce for different types of services. Firms, therefore, need to carefully consider the nature of the service when designing and implementing voice commerce solutions to ensure they meet consumer needs and preferences effectively.

Each phase of service consumption—from pre-purchase decisions, the actual service encounter, to post-encounter activities—can unfold across varied channels. During the pre-purchase phase, while there is an assumption that users might favor initiating service searches using a voice assistant, our findings indicate hesitation in using voice search for unfamiliar or complex services, as previously revealed. This reluctance stems from voice recognition hurdles such as tone, pace, accents, and so forth. Even for pre-purchase actions like setting up an account, users pointed out potential complexities in the process. Nevertheless, they felt at ease ordering frequently used services (e.g. hailing a cab) via a voice assistant. One participant articulated, “For routine payments or straightforward transactions, I’m fine with just using my voice. I prefer to avoid complications or scenarios where I’d need visual confirmation” (U11, Regular user). These findings align with Lemon and Verhoef’s (2016) customer journey framework, which conceptualizes the customer journey as a series of interconnected phases—pre-purchase, purchase, and post-purchase—each involving distinct consumer behaviors and touchpoints. The pre-purchase phase is particularly influenced by information search and evaluation, where voice assistants may serve as facilitators, yet are limited by their inability to convey detailed visual information, impacting trust and consumer confidence. This theoretical perspective underscores why consumers may prefer traditional channels for complicated searches but find voice assistants adequate for routine inquiries.

However, for the same phase, a business developer noted: “Voice assistants can play a crucial role in guiding customers through their initial research and decision-making process. While they might not make the final purchase through voice, these interactions can significantly influence their choices” (BD7, Retail industry). This is consistent with Howard and Sheth’s (1969) theory of buyer behavior, which outlines that consumer decision-making processes are shaped by external stimuli, including marketing information and social influences. Voice assistants, despite being novel, act as a source of such stimuli, guiding consumer perceptions and preferences during the early stages of the decision-making process.

Barring a few exceptions, such as music subscriptions, the actual consumption of most services does not typically occur through a voice assistant. For the consumption phase, many users migrate to channels they deem more fitting. That said, they are often open to gathering more information or service reviews from the agent. This sentiment is encapsulated in one respondent’s remark about “leveraging voice only where it feels appropriate” (U9, Experienced user) during the service experience. Lemon and Verhoef’s (2016) customer journey framework supports this view, suggesting that the effectiveness of service interactions depends on aligning the right channels with the consumer’s preferred touchpoints during the consumption phase.

In the post-service engagement phase, users expressed mixed feelings about using voice assistants. While a few users saw potential for quick feedback and issue resolution, others preferred traditional channels for more complicated follow-ups. One user noted, “For simple things like confirming my order arrived, I’d be happy to use a voice assistant. But if I had a complaint or needed to return something, I’d probably want to speak to a real person or use the app” (U7, Experienced user). A business developer echoed this sentiment, stating, “We see voice as a complementary channel for post–purchase support. It’s great for quick checks and simple queries, but we still need to offer other options for more complex customer service issues” (BD3, Retail industry). These insights suggest that while voice commerce has potential in the post-service phase, its application may be limited to simpler interactions.

Clearly, the suitability of voice assistants varies across the different phases of service consumption. While users show hesitation in using voice for complex pre-purchase decisions, they are comfortable with routine tasks and post-service engagements. Although pre-purchase and post–purchase phases may not involve direct transactions, they are crucial components of the voice commerce ecosystem. These phases significantly influence consumer decision-making and overall satisfaction, impacting, in turn, future purchase behavior. For instance, pre-purchase interactions can shape consumer expectations and preferences while post–purchase engagement can affect brand loyalty and repeat purchases. This highlights the need for firms to strategically adopt and integrate voice commerce where it offers the most convenience and least friction for users, thereby ensuring a seamless and satisfying experience across all phases of service consumption.

Different voice assistants rely on varying backend service providers to facilitate commerce. For instance, while Amazon’s Alexa primarily uses its proprietary distribution platform, Google Assistant, Apple’s Siri, and others often rely on third-party service providers to handle tasks like hotel reservations (e.g. via Booking.com) or taxi bookings (e.g. through Uber). Interestingly, the distinction between in-house or third-party service provisioning was not a major attribute for consumers. Rather, their inclination to utilize a voice commerce platform was predominantly influenced by their familiarity with and trust in the service provider. Within this context, we spotlighted two critical attributes impacting their purchase intentions.

Brand familiarity encompasses the prior understanding and recognition a customer holds about a particular brand, product, or service. Hence, it is not exclusively about knowledge. Other elements contribute to this sense of familiarity, including the brand’s visibility or recognizability. Noteworthily, brand familiarity can be distilled down to three core elements: awareness, past experiences, and influence.

Our data suggested a compelling link between users’ familiarity with a brand and their propensity to engage with it via voice commerce. Users were markedly more inclined to transact with brands they had previously encountered on different channels. Their past experiences, thus, played a pivotal role in their decision-making process. Illustrating this, one participant shared, “If it isn’t an established platform, I’d be hesitant. I would only be at ease accessing services online from a recognized brand or platform” (U6, Regular user) while another remarked “I’d refrain from making a purchase or placing an order on any website, app, or platform unfamiliar to me” (U2, Novice user).

Furthermore, even without being a direct consumer of a particular brand, mere recognition could tilt the scales in its favor. This was particularly evident with luxury brands, where users inherently expected consistency in quality and experience across various sales platforms. As an interviewee articulated, “Brand reputation is crucial. Trust is pivotal. If you’re acquainted with and have confidence in a brand, you’re more inclined towards it, even if you haven’t directly transacted with it” (U14, Regular user).

From the business perspective, the essence of brand familiarity was not lost. Professionals were acutely aware of the hurdles voice commerce presented, especially when it came to forging a strong brand image. This was due to the inherent constraints of a non-visual medium. They expressed, “In our rollout phase, we’ve introduced only a limited selection of services on the voice commerce platform, in contrast to our expansive online offerings. The absence of a visual medium poses challenges for branding. Having a distinct brand imprint in potential customers’ minds can significantly lower the entry barriers for them” (BD4, Hotel industry).

These findings suggest that brand familiarity significantly influences consumers’ purchase intention in voice commerce. Users are more likely to transact with brands they recognize and trust, thereby highlighting the importance of strong brand presence across various channels. For firms overcoming the non-visual nature of voice commerce to establish a recognizable and trusted brand image is crucial for lowering entry barriers and enhancing user adoption.

Given that a service can be availed through an array of channels, such as voice assistants, mobile applications, or websites, the choice often hinges on a channel’s perceived benefits over its alternatives. Users typically weigh the merits of each channel before settling on one. One business developer elucidated the expectations from voice commerce, “Users anticipate that service providers will streamline and enhance the purchasing experience [via voice assistants]. It should not only be efficient and expedient but also have an edge over other mediums, positioning it as the primary choice. The seamless voice-based search and findability of services are [also] vital. This necessitates companies to optimize their web assets to ensure effortless recognition and access [by the voice assistant]” (BD5, E-commerce industry).

Many users articulated that voice assistants lacked the intuitiveness seen in apps or websites. As a result, they would prefer using voice assistants for certain stages of the purchase journey while deferring to conventional channels like websites for others. For service providers, this underscores the importance of adopting a cohesive multichannel strategy, especially in the interim. This sentiment resonates with the feedback from our interviewees:

The necessity for seamless integration across multiple channels is paramount to providing a unified customer experience. Users expect consistency and continuity when transitioning between different platforms, such as starting a transaction on a voice assistant and completing it on a website. Service providers must ensure that their systems are interconnected and that user data flows smoothly across channels to facilitate this seamless experience.

Perceived relative advantage is defined as the extent to which an innovation is viewed as better than the product it supersedes, and it plays a pivotal role in the adoption rate of a new product (Rogers, 2003). This perspective hinges on consumers’ perceptions, which might not always align with the actual attributes of the product or service in question. This lens allows firms to gauge if consumers would truly prefer their offering over a competitor’s. In the context of the current study, perceived relative advantage refers to the extent to which users view voice assistants as superior to traditional e-commerce channels regarding its ease of use, convenience, and efficiency. Our findings underscore a positive sentiment toward voice assistants. The predominant advantages spotlighted by users were the ease and speed of accomplishing specific tasks requiring minimal cognitive effort. Users further emphasized convenience and potential utility, contingent on the service context. Whereas, the business developers we interviewed illuminated the unique convenience offered by voice commerce, asserting: “This platform isn’t poised to supplant e-commerce. In certain contexts, like routine FMCG acquisitions, it might progressively appear more advantageous compared to traditional channels … a sentiment reminiscent of what Amazon Dash tried to achieve” (BD1, E-commerce industry).

Although perceived relative advantage as a general construct is well established, our findings provide novel depth by highlighting specific dimensions uniquely relevant to voice-based interactions. Users emphasized that voice assistants significantly reduce cognitive and physical effort compared to traditional text-based searches or visual navigation, primarily because voice interactions eliminate manual input, visual attention, and complex screen interactions. These distinct advantages manifest particularly strongly in scenarios demanding multitasking or hands-free capabilities, such as driving or cooking, or among user groups with physical limitations—such as older adults or individuals with disabilities. For example, one user explicitly mentioned, “Given the challenges I face with typing, voice-based interactions remove barriers, [which allow me to experience] effortless online transactions” (U11, Regular user). These context-specific insights demonstrate that perceived relative advantage in voice commerce encompasses reduced cognitive load, enhanced physical convenience, and improved accessibility— dimensions largely overlooked in prior e-commerce literature that predominantly focuses on visual or tactile interfaces.

When taken collectively, these findings suggest that perceived relative advantage significantly influences consumers’ intention to make a purchase through voice commerce. Therefore, understanding and leveraging these perceived advantages can enable firms to better position their voice commerce offerings to meet consumer needs and preferences, thus enhancing the likelihood of widespread adoption.

Perceived privacy risk encapsulates consumers’ apprehensions about potential unpredictability and unfavorable outcomes tied to a purchase decision. This risk, while perceived, may not always be grounded in reality but is indeed present in the consumer’s mind (Pathak and Pathak, 2017). Financial, performance, and service risks emerged as predominant concerns users grapple with, not only in traditional e-commerce but also in voice commerce. Touching upon financial risks, apprehensions stemmed from potential monetary losses due to transaction verification lapses and open access to voice assistants within homes. One individual voiced, “I hesitate to fully trust my conversational agent with transaction finalization. What if, for instance, my son unintentionally interrupts when I’m confirming a payment?” (U5, Experienced user). Another expressed concerns about privacy and security, “The thought of private conversations being exposed and potentially leading to blackmail or misuse is downright unsettling” (U11, Regular user).

Both business developers and users emphasized the need for voice assistants to provide accurate, reliable, and transparent information to build trust and encourage users to engage in commercial transactions. As one business developer noted, “Trust is the foundation of any successful voice commerce strategy. If users don’t trust the agent, they won’t trust the recommendations or the transactions” (BD1, E-commerce industry). Users echoed this sentiment, with one stating, “I need to know that the agent has my best interests in mind and isn’t just trying to push products or services on me” (U3, Novice user).

Our analysis further reveals a notable tension concerning privacy issues between consumers and service providers in voice commerce. On one hand, consumers express significant concern regarding the potential misuse or unintended capture of sensitive information due to continuous, unrestricted access of voice assistants within their homes. One user explicitly described these apprehensions, “My biggest worry is accidentally sharing private details—conversations in the background, sensitive financial data, or personal health issues. With voice assistants constantly listening, I fear unintended consequences and misuse of my data” (U8, Regular user). This perspective underscores a prevalent fear among users related to inadvertent exposure of highly sensitive personal information due to the constant auditory monitoring embedded in voice assistant technologies.

In contrast, business developers emphasize the commercial value of voice-based data collection, framing it as essential for providing personalized services, enhancing customer experiences, and improving convenience and efficiency. A business developer clearly articulated this perspective, “To remain competitive, we must rely on detailed consumer insights captured by voice assistants. Personalized services significantly boost engagement and conversions, and data analytics are essential to achieve this level of personalization” (BD2, Retail industry). This viewpoint emphasizes how privacy trade-offs directly link to business performance outcomes.

Juxtaposing these divergent views illuminates a critical tension inherent to voice commerce, that is, while convenience and personalization may drive consumer adoption, these same factors amplify consumer concerns regarding privacy risks. Consumers simultaneously value tailored interactions and express reservations about sharing extensive personal information through voice platforms, thus highlighting a critical managerial challenge: firms must balance personalization with privacy protections to foster consumer trust and sustained engagement. As succinctly expressed by another user, “I like the idea of personalized recommendations, but I’m not sure I want the agent to know everything about me” (U1, Novice user).

These findings show that perceived privacy risk significantly impacts user trust and engagement in voice commerce, thus affecting their intention to make purchases through voice assistants. Addressing these risks through robust privacy measures and transparent communication about data usage can help mitigate consumer concerns. Balancing personalization with privacy will be crucial for fostering the widespread adoption of voice commerce.

User familiarity refers to the degree of ease and comfort users experience when interacting with technology, shaped by their prior experiences and the intuitive quality of the interface. This familiarity is particularly critical in voice commerce, as users must depend primarily on auditory interactions rather than visual or textual cues. Greater familiarity, achieved through ease of learning, intuitive design, and habitual use, significantly reduces cognitive effort because users can rely on previously established mental models to interact more efficiently with unfamiliar interfaces (Agarwal and Karahanna, 2000). In this regard, familiarity enables users to experience more confident and seamless interactions that reap the benefits of voice-based environments. As an experienced user noted, “Once familiarized, it certainly streamlines the process, saving time and proving its efficiency” (U9, Experienced user). Similarly, another user observed, “After successfully navigating it a few times, my trust would be stronger. It’s like cultivating trust with an algorithm” (U11, Regular user).

Business developers similarly recognize that user familiarity is crucial for widespread adoption of voice commerce. According to one developer, “Drawing on our e-commerce acumen, we know that for customers to even try a new platform, it must promise an experience as smooth as other established retail avenues, so we try to keep our design simple and fluid” (BD9, Transportation industry). This highlights the importance of intuitive design and user-friendly interfaces that facilitate quick learning and habitual engagement.

Our findings underscore how the distinctive cognitive demands of voice-based interactions place user familiarity at the forefront of adoption. Unlike traditional e-commerce, which relies on visual cues and clear navigational markers, voice commerce requires users to process and interpret auditory information in real time. This shift heightens the need for familiarity—defined as repeated, comfortable interaction patterns—because it lowers the cognitive burden associated with voice interactions. For instance, one regular user explained, “After successfully navigating it a few times, my trust becomes stronger—much like building trust with a reliable friend. If I can confidently predict how the assistant will respond, it significantly reduces uncertainty and makes me willing to use voice commerce more frequently” (U11, Regular user).

Similarly, business developers recognized that replicating familiar interaction cues reduces initial user hesitance. As one developer stated, “We consciously replicate familiar interaction patterns from our other digital channels, [allowing] a frictionless transition to voice. The absence of visual navigation demands intuitive, familiar auditory interactions, such as confirmation chimes, to prevent user frustration and cognitive fatigue” (BD9, Transportation industry).

Consequently, user familiarity stands out as a critical factor influencing purchase intentions in voice commerce. Unlike visually oriented platforms, voice commerce calls for carefully designed, cognitively effortless auditory interactions that capitalize on users’ established mental models. Strategically leveraging these familiarity dimensions can, therefore, meaningfully enhance adoption and sustained usage.

Social influence and normative pressures refer to the impact that societal expectations and the behaviors of significant others have on an individual’s decision to adopt or reject a technology. This hinges on how perceptions of social acceptability and the actions of family, friends, and influencers shape user behavior (LaMorte, 2019).

Individuals are more inclined to engage with a product or service when they observe its widespread adoption within their social circle. This social validation reinforces the behavior, making users gravitate toward brands or services that resonate with their peers. As voice commerce becomes increasingly prevalent, it is poised to transition from a niche offering to an integral part of daily routines.

An interviewed business developer elucidated this trajectory, stating: “At the onset, we anticipate a mix of skepticism and indifference from potential users. While novelty might spark curiosity, it’s insufficient for genuine engagement. The true challenge lies in refining the platform until it garners widespread acceptance, propelling it into the mainstream. However, the timeline for this evolution remains speculative at best” (BD5, E-commerce industry).

This insight highlights the critical role of social influence in consumers’ adoption and, in turn, purchase intention through voice commerce. Observing family, friends, and influencers using voice commerce platforms can significantly impact an individual’s decision to adopt and use these technologies. Peer recommendations and endorsements from trusted sources often serve as powerful motivators, reducing perceived risks and enhancing the perceived benefits of adoption. Moreover, normative pressures, or the perceived societal expectations to conform to certain behaviors, also play a significant role. Individuals may feel compelled to use voice commerce if they believe it is becoming a societal norm or if there is a perception that not using such technology would be seen as outdated or non-conformist. This pressure can drive the usage of voice assistants to make purchases even among those who might initially be hesitant or indifferent.

The findings indicate that attributes exert not only independent influences but also combine to form interactions among service, service provider, and consumer attributes, thereby revealing unique aspects of voice commerce. Constraints inherent to voice-based interactions—particularly the absence of visual feedback—intensify the relationship between service complexity (a service attribute) and brand familiarity (a service provider attribute). One user explained, “For straightforward inquiries, such as checking product availability, I trust any assistant. However, when the decision involves considerable uncertainty, like booking travel, my confidence heavily depends on whether the voice assistant uses a familiar and reputable brand” (U12, Regular user). Another user added, “When Alexa connects me to well-known services like Domino’s Pizza, I trust the process entirely. When directed to smaller restaurants, the absence of visuals makes me skeptical … I always verify through the app or website afterward” (U4, Experienced user).

While the previously outlined attributes strongly influence consumers’ intent to use voice assistants for purchases, the intensity of such intentions varies according to the service scenario, the type of provider, and individual consumer characteristics. Several users specifically illustrated the interaction between the service consumption phase (a service attribute) and brand familiarity (a service provider attribute):

Additional tensions emerge at the intersection of consumer attributes, notably perceived privacy risks, and service provider attributes, such as multichannel integration. Business developers reported that seamless integration across channels often strengthens consumer trust, since visual confirmations in mobile apps or websites address gaps in auditory interactions. A developer in the retail industry commented, “Consumers might initiate a voice transaction, but for critical steps like payment confirmation or detailed customization, they often prefer the reassurance of visual cues provided on our app. Thus, integrating channels enhances trust and addresses privacy concerns” (BD8, Retail industry). However, users expressed growing anxiety about privacy, since voice assistants often listen continuously, gathering information even when not actively engaged. One novice user observed, “I appreciate the convenience of seamless integration, but I worry about how much personal information is collected across these connected platforms. The convenience becomes unsettling when you realize your voice data moves freely between platforms without explicit confirmation” (U3, Novice user).

Interactions involving service complexity, brand familiarity, and perceived relative advantage further highlight the interdependent nature of these attributes. Participants reported a higher perceived advantage for voice interactions when services were routine and the provider was already familiar. A regular user explained, “I would comfortably reorder groceries from familiar brands through voice, but for an unfamiliar clothing brand, I would not trust a voice assistant without visually inspecting the product first” (U10, Regular user). Another participant noted, “When I’m pressed for time heading to work, I often hail an Uber using Assistant. No need to fumble with my phone, just a voice command does the trick” (U9, Experienced user). Familiarity with the provider can, therefore, mitigate constraints imposed by the auditory-only format, which, in turn, reinforce purchase intentions even for more complex or tangible offerings.

These examples illustrate tensions and complementarities that characterize voice commerce interactions. The heightened consumer sensitivity to privacy, the compensatory role of brand familiarity, and the occasionally conflicting nature of multichannel integration show how these attributes collectively shape consumer behavior in voice commerce. This critical, integrated perspective enables firms to better navigate consumer demands for seamless, personalized experiences while addressing significant concerns around privacy and trust, thereby strengthening adoption and sustained engagement, including purchases, in voice commerce.

This study’s inductive analysis moves beyond voice assistant characteristics and identifies three critical attribute categories—i.e. service attributes, service provider attributes, and consumer attributes—that exert noteworthy influences on consumers’ purchase intentions in voice commerce. The findings derived from this analysis enrich understanding by unpacking the ways in which these previously overlooked attributes collectively shape consumer behavior in voice commerce. While prior studies predominantly focused on isolated attributes, particularly the technological attributes of voice assistants (Fernandes and Oliveira, 2021; Moriuchi, 2019), the present study shifts theoretical attention toward understanding voice commerce as a context involving multifaceted interactions. For example, the findings demonstrate how brand familiarity compensates for service complexity and how multichannel integration can effectively mitigate consumer privacy concerns. These interconnected relationships highlight attribute-level dependencies critical for explaining consumer decision-making within e-commerce environments involving voice assistants. Figure 2 visually represents this integrated conceptual framework derived from our analysis.

Focusing on service and service provider attributes, our findings introduce novel insights by extending the established concept of fit from platform technology to service-specific contexts. Specifically, this study advances prior discussions on how product characteristics influence e-commerce adoption (Phau and Meng Poon, 2000; Vijayasarathy, 2002) by emphasizing the unique challenges and opportunities inherent in voice-based sales channels. Rather than restating established fit constructs, we highlight their increased relevance and distinct aspects arising explicitly from auditory-only interactions. For example, the analysis demonstrates that user familiarity significantly reduces cognitive barriers stemming from the absence of visual cues, which, in turn, enhances consumer comfort and trust when engaging through voice-driven platforms. Consequently, firms should strategically develop voice interfaces specifically tailored to products or services inherently compatible with auditory interactions, such as routine grocery reordering or streaming subscriptions, while maintaining traditional e-commerce channels for offerings perceived as higher-risk, visually dependent, or involving high-stakes decisions, such as expensive travel bookings or complex health-related appointments.

Further, this study illuminates tensions arising from differing priorities of business developers and users in voice commerce. Business developers commonly prioritize personalization, perceiving detailed user data as essential for enhancing consumer engagement and securing competitive advantage through tailored experiences that drive sales and loyalty. Conversely, users express pronounced reservations regarding extensive personalization due to inherent privacy risks in voice interactions, such as continuous data capture and unintended sharing of private conversations within the home. Prior studies frequently portray attributes such as anthropomorphism and usability as unequivocally beneficial for enhancing consumer engagement (Fernandes and Oliveira, 2021; Moriuchi, 2019). However, the present findings reveal that these attributes simultaneously intensify consumer privacy concerns—for instance, users expressed unease about voice assistants inadvertently capturing sensitive background discussions, financial information, or personal health data—which creates a strategic dilemma for service providers. Adopting a balanced approach becomes essential because providers must carefully navigate between offering personalized interactions (e.g. tailored product suggestions or transaction convenience) and safeguarding user privacy (e.g. clear opt-in mechanisms, explicit consent for data sharing, transparent communication regarding data usage). Implementing such integrative strategies is critical to driving sustained engagement and establishing user trust in voice-based services.

Finally, this study underscores the importance of adopting a holistic perspective that conceptualizes voice commerce as a dynamic, multisided platform ecosystem rather than viewing it simply as another channel of e-commerce. Such an ecosystem view explicitly accounts for the continuous evolution of consumer behaviors, technological advancements, and strategic interactions among various actors involved in voice commerce. For instance, service attributes like complexity and consumption phases influence consumer behaviors uniquely within auditory-only interactions, where the absence of visual confirmation heightens consumer reliance on trust signals such as brand familiarity and multichannel integration. Similarly, consumer concerns related to privacy escalate within voice-based contexts due to always-on auditory monitoring, thus intensifying tensions between the personalization sought by providers and data security demanded by users. These dynamics highlight the value of integrating insights from various perspectives to better understand how rapidly evolving voice technologies and shifting user expectations collectively shape service delivery strategies. An integrated lens, therefore, helps clarify the interdependencies among consumers, service providers, and services within voice commerce while pointing to strategic challenges that firms must address in catering to stakeholder demands.

This study provides several theoretical contributions that enrich our understanding of voice commerce and set the stage for future research in this emerging area of e-commerce. Two key overarching insights emerge from the findings: first, a holistic and multifaceted approach is essential to fully capture how different attributes collectively influence purchase intentions within voice commerce (RQ₁), and second, recognizing the interactions among these attributes is crucial, as their collective impact shapes the adoption and usage patterns of voice-based platforms in e-commerce (RQ₂). These insights advance existing literature in several noteworthy ways.

To begin, this study accentuates the importance of service and service provider attributes, alongside the established consumer and voice assistant attributes, in shaping voice commerce adoption. Voice commerce introduces unique service complexities due to its reliance on auditory communication, which removes the visual and tactile cues readily available in traditional e-commerce. This absence of visual confirmation significantly reshapes the trust dynamics and decision-making processes of consumers, making the strategic design of service attributes—such as clarity during the pre-purchase phase or reliability in post-purchase support—essential for shaping consumer purchase intentions. Consequently, assumptions prevalent in traditional e-commerce settings, which typically depend on visual navigation elements (Kumar et al., 2021), may inadequately explain user behavior in voice-based interactions, thereby reinforcing the need to expand theoretical frameworks to explicitly incorporate service and service provider attributes alongside consumer and technological (voice assistant) attributes.

Building on this expanded theoretical perspective, this study draws on Lovelock’s (1983) service classification theory, which distinguishes services according to their target (individuals or physical goods) and their tangible or intangible nature. Recent work applying this classification has positioned service robots, including voice assistants, primarily as facilitators of intangible services (Wirtz et al., 2018). The present findings reveal that consumers exhibit considerable caution in using voice commerce for services directly linked to personal appearance, well-being, or those perceived as high-stakes (e.g. medical appointments, haircuts, expensive travel bookings), unless they have established familiarity or trust with the provider. Conversely, consumers displayed greater willingness to use voice assistants for intangible services characterized by lower perceived risk, such as online learning and streaming subscriptions, and tangible services unrelated to personal well-being, such as routine household purchases. This preference distinction suggests that current technological capabilities of voice assistants are not equally suited across all service types or phases of consumption, thus highlighting the necessity to carefully reconsider service dimensions such as complexity, tangibility, and personal involvement when designing voice-based service interactions.

Connecting the implications of service characteristics to consumer attributes, this study enhances theoretical understanding of voice commerce by offering context-specific insights into how consumer attributes influence their purchase intentions via voice-based interactions. While perceived risks and subjective norms have previously been identified as influential in shaping decisions to purchase services through voice assistants (Bawack et al., 2021; Mariani et al., 2023; Moriuchi, 2019; Rabassa et al., 2022), the present study expands this understanding by emphasizing perceived relative advantage and user familiarity as essential determinants of purchase intention. Specifically, this study reveals that perceived relative advantage in voice commerce involves distinct dimensions such as reduced cognitive load, hands-free convenience, immediate interactions, and enhanced comfort, diverging notably from traditional conceptions of general “usefulness.” Interestingly, ease of use did not emerge as a prominent factor within perceived relative advantage, likely because some users encountered challenges with voice recognition (Palma et al., 2020). Instead, user familiarity emerged as a key driver of consumer behavior, distinctly significant due to the reliance on auditory rather than visual or tactile cues, thus extending beyond the generalized technological experience commonly highlighted in conventional adoption theories.

Developing this integrated consumer perspective further, this study enriches understanding by examining how these consumer attributes collectively interact with service and service provider attributes. This combined perspective emphasizes the necessity of viewing voice commerce encounters as multifaceted interactions (Rabassa et al., 2022; Robinson et al., 2020). Rather than concentrating narrowly on voice assistant adoption alone, the analysis redirects scholarly attention to the interconnected relationship between consumers, services, and service providers. Such an integrated focus addresses a gap in extant literature, which has tended to overlook how these distinct attribute categories influence each other. Through highlighting these relational dynamics, this study provides a more comprehensive understanding of the factors uniquely shaping consumer intentions and behaviors within voice commerce.

Expanding upon this comprehensive attribute analysis, this study contributes to the service innovation literature by positioning voice commerce as an emerging form of service delivery that challenges traditional assumptions around service design and management (Elshan et al., 2022; Kraus, 2019; Kraus et al., 2019). Central to this contribution is the heightened role of user familiarity and social norms, both of which are especially prominent when users interact with AI-based agents instead of conventional digital channels (Lim et al., 2022). Specifically, consumers’ comfort level with voice interfaces and societal acceptance of interactions with non-human entities become critical determinants of their adoption decisions. This finding diverges significantly from conventional e-commerce research, where familiarity and social acceptance are typically overlooked (Kumar et al., 2021). Consequently, the present results urge researchers to closely examine how prior user experiences with voice technologies and societal perceptions of AI distinctly shape purchasing behavior in voice contexts—factors that often remain underexplored in e-commerce (Luo et al., 2024).

Further extending this argument, this study highlights the dynamic nature of voice commerce as a continually evolving sociotechnical phenomenon (Elshan et al., 2022; Rabassa et al., 2022). Unlike conventional e-commerce channels, voice commerce experiences rapid technological improvements in voice assistant capabilities, ongoing shifts in consumer expectations and behaviors, and continuous adaptations by service providers. These developments create unique tensions and synergies among consumer attributes (such as perceived privacy risk and relative advantage), service provider attributes (notably brand familiarity and multichannel integration), and service attributes (including complexity). Consequently, voice commerce requires service providers to constantly adjust their strategies to accommodate evolving technological advancements and shifting consumer demands, highlighting a uniquely dynamic context for consumer interactions and strategic decision-making. This explicit recognition of the dynamic interplay among attributes distinguishes voice commerce from static, traditional e-commerce models (Kumar et al., 2021), thereby offering scholars and practitioners deeper insights into navigating rapidly changing consumer behaviors and platform strategies.

The findings offer several practical implications for strategically managing voice commerce through a clear understanding of consumer attributes, service provider attributes, and service attributes.

Firstly, the relationship between brand familiarity (service provider attribute) and service complexity (service attribute) emerges as particularly significant. Given the absence of visual and tactile cues in voice commerce, consumer uncertainty increases, especially when transactions involve complex or high-risk services. Firms can effectively reduce this uncertainty through strategies that initially introduce familiar brands for lower-risk services. Instead of transferring strategies directly from visual e-commerce, firms should design voice-specific experiences, gradually building consumer trust and habitual usage. For example, firms might initially offer voice-based interactions to reorder familiar, low-risk products before gradually extending to more complex services.

Secondly, the tension between perceived privacy risk (consumer attribute) and multichannel integration (service provider attribute) requires careful managerial consideration. Voice commerce inherently raises consumer privacy concerns due to continuous auditory monitoring and data collection. Firms can address these concerns by implementing integrated multichannel strategies, providing consumers with immediate visual confirmations through supplementary channels such as mobile applications or websites. For instance, immediately following a voice transaction, a visual summary or confirmation sent via an associated mobile app can significantly reduce perceived privacy risks and improve consumer trust.

Thirdly, perceived relative advantage and user familiarity emerge as critical consumer attributes relevant to voice commerce. Firms can enhance perceived relative advantage by clearly communicating the distinctive benefits of voice assistants, such as convenience, hands-free multitasking capabilities, immediacy, and reduced cognitive effort. For instance, clearly positioning voice commerce in contexts where consumers typically face constraints—such as commuting, exercising, or managing household tasks—can accentuate these advantages. Moreover, user familiarity becomes especially critical in voice commerce due to its reliance on auditory cues without visual support. Firms should, therefore, structure initial user interactions around familiar patterns from established e-commerce channels, thereby enabling consumers to quickly acclimate and build habitual engagement.

Finally, firms must carefully evaluate which products and services are best suited to voice commerce. While powerful, voice commerce alone may not optimally support all service categories. Firms should strategically consider factors such as service complexity, perceived risk, and consumer involvement when determining voice commerce suitability. Routine transactions—such as ordering groceries, managing appointments, or simple account inquiries—lend themselves particularly well to voice interfaces. However, complex or high-risk transactions typically require support from traditional, visually oriented channels. Adopting an integrated multichannel strategy, in which voice commerce seamlessly interacts with conventional channels, ensures consistent, compelling, and reassuring customer experiences.

This article explores and examines the key attributes driving consumers’ purchase intentions in voice commerce. The findings demonstrate that consumer, service, and service provider attributes collectively shape consumer decisions, challenging perspectives that treat these elements as independent, since considering these attributes in isolation may lead to incomplete insights and ineffective management practices. As voice technology becomes increasingly widespread, firms face growing risks of unmet consumer expectations and inefficient resource allocation when these integrated attributes are not managed collectively. Consequently, this article advocates an integrative approach, emphasizing attribute interdependencies to better capture the complexity and evolving consumer expectations associated with voice commerce.

Although this study provides valuable insights into voice commerce, several limitations must be acknowledged. The inductive nature of the study approach, for instance, necessitates caution since the findings, while novel, are preliminary and require additional validation through subsequent research. Future studies can refine and empirically validate these foundational insights using complementary methodologies such as survey (correlation) and experimental (causal) research designs. In addition, the geographical specificity of this study, which focused exclusively on business developers and consumers in Mexico, adds depth to the analysis but also limits generalizability. To address this limitation, future research should replicate and extend this investigation across varied global contexts, which, in turn, can contribute to the generalizability or extension of insights herein this study.

The authors convey their deepest and most sincere appreciation to Handling Editors-in-Chief Jay Kandampully (previous) and Linda Alkire (present), Co-Guest Editor Bieke Henkens, and the anonymous reviewers for their constructive and developmental feedback that helped improve the quality of the article.