Copyright © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024 This

version of the article has been accepted for

publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance
improvements, or any corrections. The Version of Record is available online at: [Link] Springer Nature or its licensor (e.g. a society or other
partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this
article is solely governed by the terms of such publishing agreement and applicable law.(see: [Link]

AI-powered voice assistants: developing a framework for building consumer trust

and fostering brand loyalty

Vai Rawool
Brunel University, London, UK
E-mail: 2238263@[Link]

Pantea Foroudi
Brunel University, London, UK
E-mail: [Link]@[Link]; [Link]@[Link]

Maria Palazzo
Universitas Mercatorum
Email: [Link]@[Link]
Corresponding author
AI-powered voice assistants: developing a framework for building consumer trust
and fostering brand loyalty

ABSTRACT

This paper explores the role of artificial intelligence (AI)-powered voice assistants (VAs)

in the context of online shopping, with a specific focus on Indian consumers.

Through a quantitative research approach, data were collected with an online survey of

150 Indian participants based on constructs and measurement tools, which are

meticulously defined, ensuring the reliability and validity of the results.

The study (1) explores previous research to understand the use of AI-powered VAs in

online shopping and their varied antecedents/dimensions; (2) analyses the influence of

consumer trust on intention to use, satisfaction, and emotional attachment among

individuals who use AI-powered VAs in online shopping; (3) develops a framework that

reflects the antecedents of AI-powered VAs in online shopping and the outcome as brand

loyalty, while taking consumer trust as the mediator for users’ intention to use AI-

powered VAs, customer satisfaction, and emotional attachment in their journey towards

brand loyalty; (4) studies how Alexa, an AI-powered VA, influences the consumer

journey when shopping online in India and ultimately affects the brand loyalty of Indian

consumers.

Keywords: artificial intelligence-powered voice assistants, AI, Alexa, brand loyalty,

consumer trust, consumer satisfaction, brand relationship, e-commerce.

Introduction

Artificial intelligence (AI)-powered voice assistants (VAs) can be defined as software

programs that can run either on input-driven speakers or even smart phones (Acikgoz et

al., 2023; Kautish et al., 2023). AI refers to a host of technologies ranging from natural

language processing (NPL), machine learning, deep learning, computer vision, and so on

(Jones et al., 2004; Park, 2010; Haleem et al., 2022). As VA users rapidly accept and

adopt this voice technology, attention to the understanding and exploration of this brand-

consumer relationship (where VA is considered a brand) has grown considerably (Cheng

and Jiang, 2021).

To boost the brand-consumer relationship, AI-powered chat bots have been successful in

developing personal relationships with consumers. This is demonstrated by the quick

adoption due to the industry’s fast-growing nature (31.6% annual growth) as claimed by

Cheng and Jiang (2021) in their Business Insider Report. From this research’s

perspective, a reference to the chatbot industry offers a substantial backdrop to introduce

and further discuss AI-powered VAs and the relationship fostered between them and their

consumers (Ki et al., 2020; Kautish et al., 2023). Moreover, as AI is considered an

emerging technology, Cheng and Jiang (2021) believed it to have been instrumental in

creating a conducive environment for brands to innovate while achieving a competitive

advantage.

Indeed, it is accepted that this technology has been instrumental in creating tectonic

changes in how brands at large deliver value to consumers by managing their needs

through the delivery of offerings (Kumar et al., 2019; Badghish et al., 2024). Previous

research by Kumar et al. (2019) shows the changes in consumer preferences owing to

their interactions with brands that had a poorly curated online ecosystem in terms of the
price, place, and product recommendations. This is where the advent of AI, according to

Maroufkhani et al. (2022), can help brands get a step closer to delivering offerings in a

manner that leads to lasting consumer relationships.

Earlier, companies like Google, Microsoft, Amazon, and Facebook were more focused

on text-based communication to build stronger connections with consumers. However,

with the arrival of omnichannel purchasing technologies like AI-powered VAs,

consumers have found speech-based communication to be far more effective

(Maroufkhani et al., 2022). For this reason, about 27% of the global population is using

voice search specifically with a 100% predicted growth rate of at-home VAs. With AI-

powered VAs rapidly overtaking virtual assistants (Maroufkhani et al., 2022), the

prospect of growing brand love among consumers can no longer be ignored.

In such a time of digital revolution and disruptive innovations, it becomes important for

academics and managers to study not only how consumers perceive such VAs which

make their life easy (Cheng and Jiang, 2021; Aw et al., 2022; Maroufkhani et al., 2022;

Ashrafi and Easmin, 2023; Im et al., 2023) but also to explore the dimensions of

perceived risks and drivers that can affect consumer trust in AI (Hu et al., 2022; Pitardi

and Marriott, 2021).

In this field of study, AI as an emerging technology in the branding and marketing

verticals has been a well-followed phenomenon (Badghish et al., 2024; Xiao and Tan,

2024). Besides, it was analyzed that AI has paved the way for omnichannel purchasing

by way of VAs, according to Maroufkhani et al. (2022), when it comes to consumers

shopping online. In addition, substantial research attention has been paid to transforming

customer experiences (Aw et al., 2022), customer acceptance of VAs (Fernandes and

Oliveira, 2021), perceived risk (Gillath et al., 2020), as well as the psychological
exploration of consumer perceptions (Okoli, 2018; Whang and Im, 2020; Pitardi and

Marriott, 2021) and consumer motivations for AI adoption (Choi and Drumwright, 2021).

Nonetheless, while there has been a study of fashion shopping and the role of VAs

(Kautish et al., 2023), there is no specific study that explores the role of VAs in online

shopping from the mediating influence of consumer trust with the outcome of brand

loyalty.

Furthermore, when it comes to online shopping as an industry, certain gaps have been

identified in relation to India, as there have been theoretical papers published about the

influence of Alexa on consumers (Dogra and Kaushal, 2021; Malodia et al., 2023);

however, a survey from the brand loyalty perspective with consumer trust as a mediating

factor during the purchase journey has yet to be validated.

That said, the study attempts to explore this literature gap and is aimed at achieving the

following objectives: (i) explore the previous literature to understand the use of AI-

powered VAs in online shopping and their varied antecedents/dimensions; (ii) understand

the influence of consumer trust on intention to use, satisfaction, and emotional attachment

among individuals who use AI-powered VAs in online shopping; (iii) develop a

conceptual framework that reflects the antecedents of AI-powered VAs in online

shopping and the outcome as brand loyalty (combining brand relationship and attitude)

while taking consumer trust as the mediator for users’ intention to use AI-powered VAs,

customer satisfaction, and emotional attachment in their journey towards brand loyalty;

and finally, (iv) study how Alexa, an AI-powered VA for online shopping in India

influences the consumer journey and, ultimately, brand loyalty of Indian consumers.

In line with the objectives, the following research questions were formulated: (i) What

are the main dimensions of AI-powered VAs in online shopping and how are they related

to brand loyalty? (ii) Can consumer trust influence an individual’s intent to purchase,
satisfaction, and emotional attachment when using an AI-powered VA in online

shopping? and (iii) How does an AI-powered VA user’s online shopping journey relate

to their brand relationship and brand attitude?

LITERATURE REVIEW

Emergence of AI-powered voice assistants

Consumer interactions with corporate brand organizations have evolved over time,

especially with the advent of technology (Acikgoz et al., 2023; Monod et al., 2023;

Roslan and Ahmad, 2023). This is a direct result of the increasing adoption of AI as well

as of automated technologies like service robots, chatbots, and VAs (Fernandes and

Oliveira, 2021). Organizations adopt such technologies to maximize competitive

advantage while contributing more value to consumers through heightened

personalization and efficiency. Estimates indicate that about 95% of the customer

interactions will soon be supported by this technology. For the paper, close attention is

given to the existing literature about voice conversational agents like Alexa, which

belongs to the first generation of AI-powered digital VAs (Fernandes and Oliveira, 2021).

When it comes to the adoption of technology, acceptance is one of the important factors

that paves the way for success; hence, it was important to study Alexa with regard to its

root nature, i.e., an evolved version of a service robot, as pointed out by Fernandes and

Oliveira (2021). For this, a good starting point is the exploration of the Service Robot

Acceptance Model (sRAM) proposed by Wirtz et al. (2018) to examine the Technology

Acceptance Model (TAM). The TAM, which was first conceptualized by Fred Davis

(1985), draws parallels between user motivations, perceived usefulness, ease of use, and

brand attitude (Marangunić and Granić, 2014). Various studies by academics like Davis
(1989), Venkatesh et al. (2012) have further updated the TAM while identifying the

factors that work in tandem with the usage of AI-powered devices like VAs.

In keeping with the sRAM, Wirtz et al., (2018) defined service robots as autonomous

decision-making interfaces that are capable of frontline service-led interactions that affect

customers. The authors further discussed the social interaction angle through local input

channels like microphones, cameras, and sensors, while pointing out the robot’s ability

to access the internet for data and provide highly personalized and customized output to

the customer. It also explores the customer-robot relationship and the consequent micro

impact not only from a functional perspective but also from the emotional, social,

relational, and cognitive angles, thus highlighting positive implications.

These findings helped to understand where human intervention will be required and

where it will turn to human-robot collaboration. One of the important findings was that

privacy and security issues underlying trust can hinder the quick adoption of these robots

or VAs.

The adoption of VAs was further explored by Fernandes and Oliveira (2021) in an attempt

to empirically validate the sRAM and the previously created conceptual framework

involving the TAM (Wirtz et al., 2018). Fernandes and Oliveira (2021) collected data

from young consumers in order to identify and understand the correlation between the

drivers of adoption. They identified that the functional benefits, such as perceived

usefulness, were the biggest motivation factors, which, in turn, influenced ease of use and

social norms for frequent users. The findings were also suggestive of ease of use as an

antecedent of the TAM; however, it was also reported that users looking for human

interaction ranked it lower as a reason to adopt voice technology. Fernandes and Oliveira

(2021) were successful in reiterating customer trust, rapport, and social presence as
important drivers, thereby substantially validating the Wirtz et al. (2018) framework of

the sRAM.

The selected previous research offered a directional stance for businesses and brand

managers to understand what motivates consumers to adopt voice technology through

utilitarian benefits and how brand loyalty can be influenced by building reliable,

trustworthy delivery mechanisms (Wirtz et al., 2018; Fernandes and Oliveira, 2021;

Acikgoz et al., 2023). Having established a strong starting point, it is imperative to now

explore the customer-AI powered VA relationship (brand) in further detail, specifically

from the standpoint of online shopping.

AI-powered voice assistants in Online Shopping. Theoretical Perspectives

Aw et al. (2022) defined AI-powered VAs like Alexa as data-based programs that are

embedded in Internet of Things (IoT) devices or applications. Their paper focused on

exploring the TAM and sRAM by adding a parasocial angle to the pre-existing utilitarian

one. AI-powered VAs include Apple’s Siri, Amazon’s Alexa, Microsoft’s Cortana, and

Google Assistant (Aw et al., 2022) and thus the importance of studying the relationship

between AI-powered VAs (their human-like and technological attributes) and consumers

from a holistic perspective is highlighted. Consumers use VAs to do far more than just

perform daily tasks (Guha et al., 2023; Huh et al., 2023). The usage has extended to the

realm of retail and ecommerce, as studied by Aw et al. (2022), who predicted that VAs

in terms of online shopping will soon replace laptops or computers. They also explored

how the idea of humanness in VAs will prove to be a significant game changer, especially

owing to the para-social phenomenon (consumers interact with VAs as they would with

another human), which is explained in a similar vein to how individuals would perceive

their interactions with celebrities in the past through traditional media outlets. This fact
is further reflected in the observation by Huh et al. (2023) of a deeper brand loyalty that

customers now show towards brands, which has facilitated engagement through digital

VAs.

In the same line, Aw et al. (2022) reported important findings for the consumer shopping

experience. The promotion of the ‘smart shopper’ phenomenon has been accelerated by

the popularity of VAs; this has been based on the dependency factor due to the

intelligence and competence displayed by VAs as well as their perceived high

trustworthiness, and reliability, all of which positively influence a customer’s online

shopping experience thereby cementing their attitude and loyalty towards their favourite

brands. According to a study by Okoli (2018), digital VAs like Alexa have proved

instrumental in enabling consumers to overcome the issue of decision paralysis. To

further explore this issue, it is important to study the anthropomorphism angle; this is

especially significant because of the ‘voice’ – a powerful anthropomorphic cue that

humanizes the AI-powered VA that was explored in research by Whang and Im (2020).

Besides, it is also necessary to further explore the role played by concepts such as

perceived usefulness, perceived ease of use, perceived enjoyment, social cognition, and

privacy concerns of AI-powered VAS. Therefore, taking into account RQ1, the paper

suggests the following hypotheses:

H1. There is a positive relation between the perceived usefulness of AI-powered VAs in

online shopping and brand loyalty.

H2. There is a positive relation between the perceived ease of use of AI-powered VAs in

online shopping and brand loyalty.

H3. There is a positive relation between the perceived enjoyment of AI-powered VAs in

online shopping and brand loyalty.

H4. There is a positive relation between the social cognition of AI-powered VAs in

online shopping and brand loyalty.

H5. There is a positive relation between the privacy concerns of AI-powered VAs in

online shopping and brand loyalty.

The Role of Consumer Trust and Privacy Risk in Using Voice Assistants for Brand

Loyalty

Gillath et al. (2020) found that, as is the case with any new technology adoption, AI, too,

comes with a factor of perceived risk. This is mainly because while the prevalence of AI-

powered tools like Alexa has increased, people are yet to completely trust them. Surveys

conducted and published by Dujmovic (2017) and Partners for Automated Vehicle

Education (2020) found that only 44% of participants would willingly be a passenger in

a self-driving Uber car, while another study reported that 48% of participants would not

get into a self-driving car. The collated results of this paper aim to explore the reasons

for generating trust in AI, not from a cognitive perspective, but more from an affective

perspective.

According to Gillath et al. (2020), individuals do not perceive AI usage using human to

non-human parameters, and there are many underlying factors at play, in particular, an

individual’s attachment style (high attachment anxiety, high attachment avoidance,

securely attached). The same was confirmed through a survey conducted as part of their

research, where another finding was that familiarity and consistent intention to use proved

instrumental in fostering consumer trust. They also mentioned scope for future research

from psychological perspectives. Research by Hasan et al. (2020) about the VA Siri was

successful in identifying parameters such as consumer trust, interaction, perceived risk,

and novelty value as influences on brand loyalty. Their report touches upon the
importance of perceived risk and privacy concerns relevant to sharing sensitive personal

information with the VA, which affects how loyal consumers feel towards the brand. It

is also important to consider the dual humanness explored by Hu et al. (2021), as it

investigates how an AI-powered VA’s attributes of speaking and listening can influence

understanding as the deciding factor on the consumer’s part and can affect how competent

and intelligent they think their VA is.

Findings by Marangunić and Granić (2014), Fernandes and Oliveira (2021), Sharma et

al. (2020), Hasan et al. (2020), and Pitardi and Marriott (2021) have been built upon the

perspective that when the VA behaves in a more human-like way, for instance, asking

the user not to shut it down, the user has responded more positively to that request. This

indicates that giving a personality to a VA could influence consumer behavior and affect

their trust and brand loyalty.

Therefore, considering the need to further explore whether consumer trust influences an

individual’s intent to purchase, satisfaction, and emotional attachment when using an

AI-powered VA in online shopping (RQ2), this paper proposes the following:

H6. Customer trust is a moderator between AI-powered VAs in online shopping and

intention to use.

H7. Customer trust is a moderator between AI-powered VAs in online shopping and

customer satisfaction.

H8. Customer trust is a moderator between AI-powered VAs in online shopping and

emotional attachment.

The Consumer and the AI-Powered Voice Assistant: Focus on Brand Relationship

and Attitudes.
Choi and Drumwright (2021), who studied how VA users use wake-up words/phrases

like ‘Ok Google’, or ‘Alexa!’ to initiate their interactions found that using voice

technology has numerous exciting benefits to reshape consumers’ lives, be it for

entertainment purposes, information searches, or online purchases. According to

Poushneh (2021) and Im et al. (2023), the notion that individuals treat VAs as they would

treat another human warrants the exploration of the consumer-VA relationship by

examining the motivations to use, post-consumption perceptions, and behavioral

intentions.

According to Choi and Drumwright (2021), exploration of the consumer-VA relationship

based on the CASA (Computers as Social Actors) research paradigm and the study of

gratification theory allows a substantial foray into the socio-psychological aspects of

consumer behavior. Their paper outlines five motivation factors, namely, social

interaction, personal identity, conformity, life efficiency, and information, which are all

confirmed by the idea of consumer satisfaction and further influence the intent to use.

The communication aspect of an AI-powered VA, especially at a personal level, is what

makes socially attractive for a consumer to use and form somewhat of an attachment with

it, as concluded by Cheng and Jiang (2021), Ashrafi and Easmin (2023), Aw et al. (2022),

and the report by EY India (2023).

It can be noted that a VA can take on the role of a helper when it comes to consumers

navigating through the online purchase journey. Hence, two questions arise: What works?

and What doesn’t work? Carmel et al. (2020) showed that personalized preferences,

popularity, exploration, tolerance, and indirect relation can be seen as parameters of

relevancy when it comes to suggestions given by a VA to a customer while they look for

products to buy online. This finding can be used to draw parallels between consumer

satisfaction, intention to use, and emotional attachment, stemming from the idea of trust.
However, the purpose of this paper is to offer more insights into this field of study, as the

analysis investigates how an AI-powered VA user’s online shopping journey relates to

brand relationship and brand attitude (RQ3). Thus, the following hypotheses are

proposed:

H9. There is a positive relation between intention to use with brand relationship and

attitude.

H10. There is a positive relation between customer satisfaction with brand relationship

and attitude.

H11. There is a positive relation between emotional attachment with brand relationship

and attitude.

Based on the observations and understanding of the previous studies, this paper explores

the antecedents and outcomes shown in Figure 1. Moreover, this research considers

customer trust as a mediating factor and suggests that it has an impact on intention to use,

consumer satisfaction, and emotional attachment of AI-powered VAs in online shopping.

Thus, the paper presents the following conceptual framework (Fig. 1).

Figure 1: Proposed Conceptual Framework

METHODOLOGY

The paper focusses on primary research through a quantitative study with a positivist

approach to deductively test the hypotheses and answer the questions.

According to Rahman (2020), quantitative research makes use of deductive logic to

divide empirical constructs and represent them in terms of frequencies and to further

explore their association through statistical analysis. The methodology, as opposed to a

method, was chosen based on the argument by Walter and Andersen (2016) that it would

allow this study to present a clear picture of the relationship between social constructs

with statistical analysis at the base. The paper followed the technique suggested by

Rahman (2020) to collect numerical data and conduct a statistical analysis. Additionally,

the quick administration and evaluation of results and the measurement of the extent of

agreement or disagreement between the data sets collected by Choy (2014) justified the

preference for quantitative research.

According to Newman (2000), a quantitative study following a positivist approach works

on the concept of theory testing, where a general theory helps form a particular

hypothesis, which is then tested/confirmed, and conclusions are drawn through data

collection, analysis, and the report of the findings.

The paper took into account the industry reports of Mint (2021) and EY India (2023)

when conducting the broad literature review, which was done after identifying the

research gap or the research problem. This gap was used to carry out a study by

formulating hypotheses about the usage of AI-powered VAs and the mediating role

consumer trust plays in online shopping in India to build brand loyalty with a specific

nod to Alexa as the chosen brand.

Following the recommendation by Lakens (2022), the paper justifies the selected sample

design (sample size) through heuristics. The sample size to conduct this study was 150

participants, who were asked to answer the online close-ended survey questionnaire in

English (see Appendix I). Due to the online nature of the survey, the river sampling

method was chosen to recruit respondents as part of the non-probability approach

(Lehdonvirta et al., 2020). River sampling is also referred to as intercept sampling or

real-time sampling wherein participants are invited through an online survey link shared

on a website or via email-led platforms; it is considered more accessible than the older,

more traditional sampling methods propagated by Olivier (2011). To conduct this survey,

participants were asked to answer the online survey through email and social media

communication channels such as WhatsApp, LinkedIn, and Instagram. Data were

collected from January to October 2022.

DATA COLLECTION, ANALYSIS AND RESEARCH ETHICS

The use of a survey as a tool for data collection for this study helped to determine

causation (Nardi, 2018) between the dimensions/antecedents and the observed outcomes

which would later be interpreted in keeping with the hypothesis and conceptual

framework. This measurement tool proves advantageous (Nardi, 2018), as it offers easier

comparisons with studies of similar topics and anonymity for participants, and most

importantly, is a safe method to help participants answer questions that may carry

sensitive or personal connotations. The data were analyzed with the help of IBM SPSS

29 (Statistical Package for Social Sciences) software, wherein correlation was attempted

through a multi-item (5) Likert scale which ranged from ‘Strongly Disagree’ to ‘Strongly

Agree’. The multi-item analysis has been proven to be more valid, accurate, and reliable
compared to single-item analysis. Each construct as part of the survey design had multiple

questions to ensure that all perspectives would be considered when measuring each

construct.

Finally, it is important to highlight that this research paid close attention to ethical

considerations. Research ethics boards such as research ethics committees (RECs) and

the institutional review boards (IRBs) are created with the sole purpose of ensuring that

research projects follow ethical practices diligently, that they guarantee participant

protection, and that the advantages of the research outweigh any conceived harm

(Bouhouita-Guermech et al., 2023). This research was conducted in adherence to the

REC’s requirements, whereby data were collected only after an approval letter had been

received. For the online survey, participant protection was guaranteed online through the

‘Agree to Consent Form’ button of the eConsent Form, which works to uphold the respect

of research participants and the research principles (Yusof et al., 2022) as well as by

sharing a ‘Participant Information Sheet’.

The definition of constructs, the main survey questionnaire, and the initial pool of items

are discussed in detail in the following sections.

ITEM MEASUREMENTS OF CONSTRUCTS

According to Wilson (2023), when a researcher attempts to study human attributes like

their behavior and attitudes, it becomes imperative to create a measurement model, which

is typically used to collect and store information about each property. The measurement

model, presented in Table 1, focusses on the initial pool of items or constructs and their

measurement scales.
Table 1: Table of Item Measurements

Main concept: Perceived Usefulness

Item Measurements References

Alexa makes the information search for online shopping (Pitardi and Marriott,

easier. 2021)

Alexa provides me with good quality information. (Sohn, 2017)

Alexa provides me with up-to-date information. (Fernandes and Oliveira,

I think Alexa is useful to me. 2021)

Alexa helps me with time-effectiveness by gathering relevant (Choi and Drumwright,

information. 2021)

I find Alexa extremely useful when I am searching for

information while shopping online.

Alexa makes shopping online easier.

Using Alexa makes shopping online extremely productive.

Because Alexa is useful for multitasking.

Main concept: Perceived Ease of Use

Item Measurements References

Learning how to use Alexa is easy for me. (Pitardi and Marriott,

I think I can use Alexa without any help. 2021)

My interactions with Alexa are clear and understandable. (Sohn, 2017)

I find Alexa easy to use. (Fernandes and Oliveira,

Alexa helps me navigate online shopping easily. 2021)

I find it easy to become skilled at using Alexa while shopping

online.
I find Alexa to be always up and running when it comes to

online shopping

Main concept: Perceived Enjoyment

Item Measurements References

I find Alexa pleasant to interact with, (Pitardi and Marriott,

I find using Alexa enjoyable while shopping online. 2021)

I find using Alexa pleasant while shopping online. (Fernandes and Oliveira,

I have fun when using Alexa to shop online. 2021)

Because Alexa makes me look cool. (Choi and Drumwright,

Voice ordering is enjoyable/unenjoyable. 2021)

Voice ordering is thrilling.

Voice ordering is exciting.

Voice ordering is delightful.

Main concept: Social Cognition

Item Measurements References

When it comes to online shopping – (Pitardi and Marriott,

I feel that Alexa understands me. 2021)

I think Alexa is helpful. (Fernandes and Oliveira,

I think Alexa is warm. 2021)

I think Alexa has good intentions. (Choi and Drumwright,

I think Alexa is effective. 2021)

I think Alexa is intelligent.

I think Alexa is competent.

Because most people around me get excited about using

Alexa.
Main concept: Privacy Concerns

Item Measurements References

I have my doubts over the confidentiality of my interactions (Pitardi and Marriott,

with Alexa during online shopping. 2021)

I am concerned to perform financial transactions through (Sohn, 2017)

Alexa during online shopping.

I am concerned that my personal details stored on Alexa

during online shopping could be stolen,

I am concerned that Alexa collects too much information

about me while I am shopping online.

Alexa is secure.

Alexa does not share my personal information with any other

websites.

Alexa protects information about my online shopping

behavior.

Main concept: Consumer Trust

Item Measurements References

I feel that Alexa makes truthful claims and recommendations (Pitardi and Marriott,

during online shopping. 2021)

I feel that Alexa is trustworthy during online shopping. (Fernandes and Oliveira,

I believe what Alexa tells me/suggests during online 2021)

shopping. (Hu et al., 2021)

I feel that Alexa is honest with me during online shopping.

I feel that I can rely on Alexa to do what it is supposed to do.

I believe that Alexa provides accurate information.

Alexa performs all of its roles very well.

Alexa is proficient.

I believe that Alexa would act in my best interests.

Alexa is sincere and genuine.

If I ask for help, Alexa will do its best to assist me.

Main concept: Intention to use

Item Measurements References

It is likely that I will use Alexa in the future during online (Pitardi and Marriott,

shopping. 2021)

I intend to use Alexa frequently during online shopping. (Fernandes and Oliveira,

I expect to continue using Alexa during online shopping in the 2021)

future. (Choi and Drumwright,

I intend to use Alexa in the future. 2021)

Because Alexa helps me get information on things that

interest me.

Main concept: Consumer Satisfaction

Item Measurements References

I am satisfied with my decision to shop online through Alexa. (Kumar, Sachan and

If I had to shop online again, I may feel differently about Dutta, 2020)

using Alexa. (Poushneh, 2021)

My choice to use Alexa during online shopping was wise.

I think I did the right thing by using Alexa during online

shopping.

Overall, I am satisfied with Alexa.

Being a user of Alexa has been a satisfying experience when

it comes to online shopping.

Having experienced Alexa during online shopping, it was

pleasurable

Main concept: Emotional Attachment

Item Measurements References

Alexa can recognize my emotions during online shopping. (Pantano and Scarpi,

I like Alexa because I can have a casual conversation with it. 2022)

Because Alexa enhances my self-image. (Choi and Drumwright,

Because Alexa helps me express myself as an early adopter. 2021)

Alexa can adapt recommendations according to my emotions

during online shopping.

Alexa is a good listener.

Alexa uses the right tone of conversation during online

shopping

Main concept: Brand Relationship

Item Measurements References

I feel that Alexa tries to support a long-term commitment for (Sharma et al., 2020)

me during online shopping. (Ki, Cho and Lee, 2020)

Alexa and I both benefit from this relationship during online

shopping.

Most people like me are happy in their interactions with Alexa

during online shopping.

I am pleased with the relationship Alexa has established with

me.
 My Alexa is friendly.

When using my Alexa, I am completely absorbed in it.

When using my Alexa, I am usually deeply involved with my

interactions with it.

My Alexa excites my curiosity.

Main concept: Brand Attitude

Item Measurements References

Overall, I feel using Alexa during online shopping is a good (Pitardi and Marriott,

idea. 2021)

I generally have positive feelings towards Alexa during online (Ki, Cho, and Lee, 2020)

shopping.

The thought of using Alexa during online shopping is

appealing to me.

If my Alexa were a real person, I would disclose a great many

things to them.

If my Alexa were a real person, I would reveal myself

completely and honestly to them.

If my Alexa were a real person, I would be able to count on

them in my times of need

VALIDITY AND RELIABILITY OF CONSTRUCTS

The constructs of this study were defined in line with the previous literature and included

in the quantitative survey through item measurements (see Table 1). Besides, it was
important to test the validity and reliability of the constructs before moving on to

hypothesis testing. Appendix II shows the analysis of validity and reliability through

Cronbach’s alpha value analysis. It demonstrates a high level of internal consistency

between the constructs. The overall Cronbach’s alpha for all the items combined is 0.988,

indicating strong reliability. Indeed, all the constructs exhibit strong reliability with

Cronbach’s alpha values well above the accepted threshold of 0.7. Notably, constructs

like “Perceived Usefulness,” “Perceived Ease of Use,” “Perceived Enjoyment,” “Social

Cognition,” “Consumer Trust,” “Intention to Use,” “Consumer Satisfaction,” “Emotional

Attachment,” “Brand Relationship,” and “Brand Attitude” all surpass this threshold, with

alpha values ranging from 0.987 to 0.989. This further indicates that the items within

each construct are strongly correlated, reinforcing the reliability and internal consistency

of this study’s measurement tool, i.e., the online survey. The reliability analysis

demonstrated that the survey’s constructs were both valid and reliable, making it a strong

tool to capture respondents’ perceptions and attitudes regarding the use of Alexa in online

shopping contexts. This helped to progress testing the hypotheses and finding the answers

to the research questions.

FINDINGS

As discussed before, quantitative data were collected through river sampling from an

online survey that was shared with 150 participants in India. In terms of the demographic

information, the study identified the respondents, and no specific criteria were stipulated.

The only specific criterion in terms of demographic information was whether a

respondent was a user or a non-user of Alexa. The following tables represent the

statistical analysis done through cross tabulation based on the respondent’s age, gender,
and Alexa usage to draw up a demographic profile through statistical analysis for the

survey sample. This offered a demographic profile of the sample size. From above, it can

be observed that most Alexa user respondents belong to the age groups of 31-40 and 41-

50 years, with a notable concentration of males in the latter group. Meanwhile, non-users

of Alexa are most prevalent in the age group of 21-30 years, and there is a relatively

balanced distribution of genders among non-users.

The chi-square tests provide statistical significance for the relationships between Alexa

user status and age/gender. For Alexa users, both the Pearson Chi-Square and

Likelihood Ratio tests yield p-values above the conventional significance level of 0.05,

suggesting that there is no significant association between Alexa usage and age/gender

in this sample.

Table 2: Crosstabulation and Chi-square Test of Respondent’s Age, Gender, and

Alexa Usage

Age, Gender, Alexa_User Crosstabulation

Count
Gender
Alexa_User Female Male Total
Alexa_User Age 21 – 30 years 4 4 8
31 – 40 years 13 7 20
41 – 50 years 10 17 27
51 – 60 years 7 3 10
Total 34 31 65
Non-User Age 18 – 20 years 1 0 1
21 – 30 years 33 24 57
31 – 40 years 7 7 14
41 – 50 years 6 5 11
51 – 60 years 0 2 2
Total 47 38 85
Total Age 18 – 20 years 1 0 1
21 – 30 years 37 28 65
31 – 40 years 20 14 34
41 – 50 years 16 22 38
51 – 60 years 7 5 12
 Total 81 69 150
Chi-Square Tests
Asymptotic
Significance (2-
Alexa_User Value Df sided)
Alexa_User Pearson Chi-Square 5.087b 3 .166
Likelihood Ratio 5.171 3 .160
No. of Valid Cases 65
Non-User Pearson Chi-Square 3.599c 4 .463
Likelihood Ratio 4.722 4 .317
No. of Valid Cases 85
Total Pearson Chi-Square 3.649a 4 .456
Likelihood Ratio 4.026 4 .402
No. of Valid Cases 150

Similarly, for non-users, the tests indicate no significant relationship. While no

hypothesis was proposed to explore the influences of age and gender as part of this study,

the results of the demographic profile imply that Alexa adoption is not strongly

influenced by age or gender among this survey’s respondents (see Table 2).

RESULTS OF TESTING HYPOTHESES

Correlation analysis was conducted for each of the following hypotheses, following a

Pearson 2-tailed correlation on SPSS, and the results were as follows.

H1 - There is a positive relation between the perceived usefulness of AI-powered

VAs in online shopping and brand loyalty.

The survey respondents’ answers revealed strong, positive links between the perceived

usefulness of AI-powered virtual assistants in online shopping and brand loyalty.

Increased perceived usefulness across several dimensions, like information search,

quality, and overall utility, was tied to higher brand loyalty. Emotional connections with

VAs and perceiving them as friendly and capable of meeting needs also correlated

positively with brand loyalty for the participants. The result of a Pearson 2-tailed
correlation analysis in SPSS therefore supported the hypothesis (H1) of a positive

relationship between VAs’ usefulness in online shopping and brand loyalty. This

resonates with the findings of previous studies, such as Huh et al. (2023), that to boost

brand loyalty, businesses should prioritize enhancing VAs’ perceived usefulness and

fostering emotional bonds in the online shopping experience.

H2 - There is a positive relation between the perceived ease of use of AI-powered

VAs in online shopping and brand loyalty.

The results from the SPSS analysis revealed significant correlations between the

perceived ease of use of AI-powered virtual assistants in online shopping and brand

loyalty. Specifically, the data indicated a strong positive relationship between perceived

ease of use and various dimensions of brand loyalty, such as brand relationship, brand

attitude, and overall brand perception for the respondents. These findings supported the

second hypothesis (H2), which suggested a positive relationship between the perceived

ease of use of AI-powered VAs and brand loyalty. This resonates with the findings of

previous studies, such as Akdim and Casaló (2023). Besides, the statistically significant

correlations at the 0.01 level (2-tailed) across multiple dimensions suggested that when

consumers find AI-powered VAs easy to use in online shopping, they are more likely to

form strong and positive brand associations, ultimately fostering brand loyalty.

H3 - There is a positive relation between the perceived enjoyment of AI-powered

VAs in online shopping and brand loyalty.

The correlation analysis in SPSS 29 revealed strong and highly significant positive

relationships between the perceived enjoyment of AI-powered Virtual Assistants in

online shopping and brand loyalty based on the data collected from respondents. The

correlations among various aspects of perceived enjoyment, such as interaction

pleasantness, enjoyment while shopping, having fun, and even the perception that VAs
make users look ‘cool’, all exhibit strong positive associations with brand loyalty.

Additionally, the excitement, thrill, and delight associated with voice ordering through

VAs also showed remarkably strong positive correlations with brand loyalty. These

findings strongly support H3, as highlighted also by Huh et al. (2023), suggesting that

enhancing the perceived enjoyment of VAs in online shopping can foster greater brand

loyalty. Brands should therefore prioritize making the shopping experience with VAs

more enjoyable to build stronger brand loyalty among customers.

H4 - There is a positive relation between the social cognition of AI-powered VAs in

online shopping and brand loyalty.

The correlation analysis results indicate a strong and statistically significant positive

relationship between the social cognition of AI-powered VAs in online shopping and

brand loyalty. The correlations between various aspects of social cognition, such as the

understanding, helpfulness, warmth, good intentions, effectiveness, intelligence, and

competence of Alexa during online shopping, were all highly positive and significant (all

p < 0.01). Additionally, these aspects of social cognition were positively correlated with

different dimensions of brand loyalty, including the commitment to a long-term

relationship, mutual benefit, customer satisfaction, and emotional involvement during

interactions with Alexa while shopping. These findings suggest that as users perceive

Alexa as more socially competent, their brand loyalty and positive attitudes toward Alexa

in the context of online shopping also increase. Overall, the results support H4, indicating

a strong positive relation between social cognition and brand loyalty in the context of AI-

powered VAs in online shopping. This seems to be in line with the findings of other

studies (Maroufkhani et al., 2022; Huh et al., 2023).

H5 There is a positive relation between the privacy concerns of AI-powered VAs in

online shopping and brand loyalty.

The Pearson correlation analysis reveals a robust and positive connection between the

privacy concerns surrounding AI-powered VAs in online shopping and brand loyalty.

Notably, concerns about interaction confidentiality, financial transactions, and personal

data security on Alexa during online shopping exhibit strong, positive correlations with

diverse facets of brand loyalty, including commitment, mutual benefit, and satisfaction.

Additionally, anxieties regarding excessive data collection correlate positively with

overall brand attitude. These results affirm H5, emphasizing the pivotal role of addressing

privacy concerns in bolstering brand loyalty for AI-powered VAs in the context of online

shopping, as was previously analyzed by Kamoonpuri and Sengar (2023).

Considering that Consumer Trust played the role of a moderating factor, moderation

analysis through linear regression analysis was chosen to test the corresponding

hypotheses where intention to use, customer satisfaction and emotional attachment took

on the position of dependent variables each, and AI-powered VAs in online shopping

were the independent variables. The following are the results found for each hypothesis.

H6 Customer trust is a moderator between AI-powered VAs in online shopping and

intention to use.

Regarding the moderation analysis for Hypothesis 6 (Customer trust moderates the

relationship between AI-powered VAs and intention to use), the results from Table 3

show that the interaction term (Interaction_AIxTrust_Intention) has a significant effect

on intention to use (F=85.238, p<.001). The model with the interaction explains more

variance in intention to use (R^2=0.537) compared to the model without it (R^2=0.432).

The standardized coefficient for the interaction term (Beta=0.741) indicates that as

customer trust increases, the positive relationship between AI-powered VAs and intention

to use is strengthened. This suggests that customer trust indeed moderates the

relationship, supporting H6.

Table 3: Moderation Analysis for Hypothesis 6

ANOVAa
Model Sum of Squares df Mean Square F Sig.
1 Regression 91.194 1 91.194 112.336 <.001b
Residual 120.146 148 .812
Total 211.340 149
2 Regression 113.484 2 56.742 85.238 <.001c
Residual 97.856 147 .666
Total 211.340 149
Excluded Variablesa
Collinearity
Partial Statistics
Model Beta In T Sig. Correlation Tolerance
b
1 Interaction_AIxTrust_In .741 5.787 <.001 .431 .192
tention
a. Dependent Variable: Intention to Use [I intend to use Alexa in the future.]
b. Predictors in the Model: (Constant), AI-powered Vas

H7 Customer trust is a moderator between AI-powered VAs in online shopping and

customer satisfaction.

Regarding the moderation analysis for H7 (Customer trust as a moderator between AI-

powered VAs in online shopping and customer satisfaction), the results shown in Table

4 indicate that the interaction term (Interaction_AIxTrust_Satisfaction) significantly

influences customer satisfaction (F=71.387, p<.001). The model with the interaction term

explains more of the variance in customer satisfaction (R^2=0.493) compared to the

model without it (R^2=0.402).

The standardized coefficient for the interaction term (Beta=0.688) suggests that as

customer trust increases, the positive relationship between AI-powered VAs and

customer satisfaction becomes stronger. These findings support H7, confirming that

customer trust acts as a moderator in the relationship between AI-powered VAs and

customer satisfaction.

Table 4: Moderation Analysis for Hypothesis 7

ANOVAa
Model Sum of Squares Df Mean Square F Sig.
1 Regression 66.320 1 66.320 99.373 <.001b
Residual 98.773 148 .667
Total 165.093 149
2 Regression 81.343 2 40.672 71.387 <.001c
Residual 83.750 147 .570
Total 165.093 149
Excluded Variablesa
Collinearity
Partial Statistics
Model Beta In T Sig. Correlation Tolerance
b
1 Interaction_AIxTrust_S .688 5.135 <.001 .390 .192
atisfaction
a. Dependent Variable: Consumer Satisfaction [Overall, I am satisfied with Alexa.]
b. Predictors in the Model: (Constant), AI-powered Vas

H8 Customer trust is a moderator between AI-powered VAs in online shopping and

emotional attachment.

Table 5: Moderation Analysis for Hypothesis 8

ANOVAa
Model Sum of Squares Df Mean Square F Sig.
1 Regression 56.838 1 56.838 54.996 <.001b
Residual 152.956 148 1.033
Total 209.793 149
2 Regression 68.479 2 34.239 35.617 <.001c
Residual 141.315 147 .961
Total 209.793 149
Excluded Variablesa
Collinearity
Partial Statistics
Model Beta In T Sig. Correlation Tolerance
b
1 Interaction_AIxTrust_E .538 3.480 <.001 .276 .192
motional Attachment
a. Dependent Variable: Emotional Attachment [Alexa can adapt recommendations according to
my emotions during online shopping.]
b. Predictors in the Model: (Constant), AI-powered Vas

In the moderation analysis for H8 (see Table 5), Customer trust is a moderator between

AI-powered VAs in online shopping and emotional attachment; the results reveal that the

interaction term Interaction_AIxTrust_Emotional Attachment has a significant impact on

emotional attachment (F=35.617, p<.001). The model with the interaction term explains

a higher proportion of variance in emotional attachment (R^2=0.326) compared to the

model without it (R^2=0.271). The standardized coefficient for the interaction term

(Beta=0.538) indicates that as customer trust increases, the positive relationship between

AI-powered VAs and emotional attachment becomes more pronounced. These findings

support H8, confirming that customer trust serves as a moderator in the relationship

between AI-powered VAs and emotional attachment during online shopping, as

suggested also by Bilal et al. (2024).

The final set of hypotheses were tested through regression analysis where the construct

of brand relationship became the dependent variable, and the constructs of intention to

use, consumer satisfaction, emotional attachment, and brand attitude became independent

variables. The results were as follows.

H9 There is a positive relation between intention to use and brand relationship and

attitude.

The regression analysis results shown in Table 6 reveal important insights regarding H9,

which posits a positive relationship between intention to use, brand relationship, and

attitude. The models constructed for this analysis indicate that there is indeed a

relationship between these variables. In Model 1, intention to use demonstrates a

significant positive association with brand relationship (Beta = 0.677, p < 0.001),

explaining 45.9% of the variance in brand relationship. When Brand Attitude is

introduced in Model 2, the explanatory power increases substantially (R-squared =

0.690). Both intention to use (Beta = 0.096, p = 0.187) and brand attitude (Beta = 0.755,

p < 0.001) emerge as significant predictors of brand relationship.

Table 6: Regression Analysis Table for Hypothesis 9

ANOVAa
Model Sum of Squares Df Mean Square F Sig.
1 Regression 70.242 1 70.242 125.355 <.001b
Residual 82.931 148 .560
 Total 153.173 149
2 Regression 105.742 2 52.871 163.856 <.001c
Residual 47.432 147 .323
Total 153.173 149
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
b. Predictors: (Constant), Intention to Use [I intend to use Alexa in the future.]
c. Predictors: (Constant), Intention to Use [I intend to use Alexa in the future.], Brand Attitude
[Overall, I feel using Alexa during online shopping is a good idea.]
Coefficientsa
Unstandardized Standardized
Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 1.366 .196 6.965 <.001
Intention To Use [I .577 .051 .677 11.196 <.001
intend to use Alexa in
the future.]
2 (Constant) .708 .162 4.383 <.001
Intention To Use [I .081 .061 .096 1.327 .187
intend to use Alexa in
the future.]
Brand Attitude [Overall, .715 .068 .755 10.489 <.001
I feel using Alexa
during online shopping
is a good idea.]
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
Excluded Variablesa
Collinearity
Partial Statistics
Model Beta In T Sig. Correlation Tolerance
1 Brand Attitude [Overall, .755b 10.489 <.001 .654 .407
I feel using Alexa
during online shopping
is a good idea.]
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
b. Predictors in the Model: (Constant), Intention to Use [I intend to use Alexa in the future.]

This result is a novelty in this field of study, as the role played by intention to use and

brand attitude for strengthening brand relations has not been deeply explored in past

research. Importantly, brand attitude appears to exert a stronger influence than intention

to use. Thus, the results support H9, indicating a positive relationship between these
variables, particularly when considering the combined impact of intention to use and

brand attitude.

H10 There is a positive relation between customer satisfaction and brand

relationship and attitude.

The results of the regression analysis for Hypothesis 10 (see Table 7), which explores

the relationship between customer satisfaction, brand relationship, and attitude, provide

valuable insights. Model 1 demonstrates a significant positive relationship between

customer satisfaction and brand relationship (Beta = 0.803, p < 0.001), explaining

64.5% of the variance in brand relationship. When Brand Attitude is added in Model 2,

the explanatory power increases (R-squared = 0.709), signifying that both customer

satisfaction (Beta = 0.323, p < 0.001) and brand attitude (Beta = 0.543, p < 0.001) are

significant predictors of brand relationship. Notably, brand attitude exerts a stronger

influence than customer satisfaction in this context.

Table 7: Regression Analysis Table for Hypothesis 10

ANOVAa
Model Sum of Squares Df Mean Square F Sig.
1 Regression 98.815 1 98.815 269.044 <.001b
Residual 54.358 148 .367
Total 153.173 149
2 Regression 108.655 2 54.327 179.388 <.001c
Residual 44.519 147 .303
Total 153.173 149
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
b. Predictors: (Constant), Consumer Satisfaction [I am satisfied with my decision to shop online
through Alexa.]
c. Predictors: (Constant), Consumer Satisfaction [I am satisfied with my decision to shop online
through Alexa.], Brand Attitude [Overall, I feel using Alexa during online shopping is a good
idea.]
Coefficientsa
Unstandardized Standardized
Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) .814 .168 4.839 <.001
 Consumer Satisfaction .784 .048 .803 16.403 <.001
[I am satisfied with my
decision to shop online
through Alexa.]
2 (Constant) .630 .156 4.029 <.001
Consumer Satisfaction .315 .093 .323 3.390 <.001
[I am satisfied with my
decision to shop online
through Alexa.]
Brand Attitude [Overall, .514 .090 .543 5.700 <.001
I feel using Alexa
during online shopping
is a good idea.]
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
Excluded Variablesa
Collinearity
Partial Statistics
Model Beta In T Sig. Correlation Tolerance
1 Brand Attitude [Overall, .543b 5.700 <.001 .425 .218
I feel using Alexa
during online shopping
is a good idea.]
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
b. Predictors in the Model: (Constant), Consumer Satisfaction [I am satisfied with my decision to
shop online through Alexa.]

The ANOVA results indicate that both Model 1 and Model 2 are highly significant (p <

0.001), suggesting that these models provide a better explanation of the variance in brand

relationship than a null model with no predictors. In conclusion, the findings support

Hypothesis 10, confirming a positive relationship between customer satisfaction, brand

relationship, and attitude. Particularly, when considering the combined effects of

customer satisfaction and brand attitude, they significantly contribute to explaining the

variance in brand relationship. This underscores the importance of both customer

satisfaction and brand attitude in fostering a strong and enduring brand relationship in the

context of online shopping through Alexa.

H11 There is a positive relation between emotional attachment and brand

relationship and attitude.

The results of the regression analysis for Hypothesis 11 as part of Table 8, which

examines the positive relationship between emotional attachment, brand relationship, and

attitude, provide valuable insights into the interplay of these variables. In Model 1,

emotional attachment shows a significant positive relationship with brand relationship

(Beta = 0.677, p < 0.001), explaining 45.9% of the variance in brand relationship. When

Brand Attitude is introduced in Model 2, the explanatory power increases substantially

(R-squared = 0.707). Both emotional attachment (Beta = 0.198, p = 0.002) and brand

attitude (Beta = 0.691, p < 0.001) are significant predictors of brand relationship, with

brand attitude demonstrating a stronger influence.

Table 8: Regression Analysis Table for Hypothesis 11

ANOVAa
Model Sum of Squares Df Mean Square F Sig.
1 Regression 70.279 1 70.279 125.477 <.001b
Residual 82.894 148 .560
Total 153.173 149
2 Regression 108.280 2 54.140 177.280 <.001c
Residual 44.893 147 .305
Total 153.173 149
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
b. Predictors: (Constant), Emotional Attachment [Alexa can recognize my emotions during
online shopping.]
c. Predictors: (Constant), Emotional Attachment [Alexa can recognize my emotions during
online shopping.], Brand Attitude [Overall, I feel using Alexa during online shopping is a good
idea.]
Coefficientsa
Unstandardized Standardized
Coefficients Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 1.562 .180 8.702 <.001
Emotional Attachment .623 .056 .677 11.202 <.001
[Alexa can recognize
my emotions during
online shopping.]
2 (Constant) .657 .155 4.225 <.001
 Emotional Attachment .182 .057 .198 3.190 .002
[Alexa can recognize
my emotions during
online shopping.]
Brand Attitude [Overall, .655 .059 .691 11.155 <.001
I feel using Alexa
during online shopping
is a good idea.]
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
Excluded Variablesa
Collinearity
Partial Statistics
Model Beta In t Sig. Correlation Tolerance
1 Brand Attitude [Overall, .691b 11.155 <.001 .677 .519
I feel using Alexa
during online shopping
is a good idea.]
a. Dependent Variable: Brand Relationship [Alexa tries to support a long-term commitment for
me during online shopping.]
b. Predictors in the Model: (Constant), Emotional Attachment [Alexa can recognize my emotions
during online shopping.]

The ANOVA results indicate that both Model 1 and Model 2 are highly significant (p <

0.001), suggesting that these models provide a better explanation of the variance in brand

relationship than a null model with no predictors.

In conclusion, the findings support H11, confirming a positive relationship between

emotional attachment, brand relationship, and attitude. Particularly, regarding the

combined effects of emotional attachment and brand attitude, they significantly

contribute to explaining the variance in brand relationship, as was analyzed by

Ghorbanzadeh and Rahehagh (2021). This underscores the significance of emotional

attachment, in conjunction with brand attitude, in nurturing a strong and enduring brand

relationship in the context of online shopping through Alexa.

CONCLUSIONS AND IMPLICATIONS

This study has traversed the landscape of AI-powered VAs in online shopping,

uncovering insights that contribute significantly to the existing body of knowledge. The

theoretical implications of the findings offer a deeper understanding of the dynamics at

play in the context of customer behavior, brand loyalty, and the role of trust in the era of

AI-driven commerce.

The contributions of this study are threefold. First, the paper contributes to the literature

on AI-powered VAs, which still needs to be strengthened (Lu et al., 2020). Specifically,

the research reinforces the notion that perceived usefulness, ease of use, enjoyment, and

social cognition are pivotal factors that positively influence brand loyalty in the context

of AI-powered VAs. These findings align with the previous literature on technology

acceptance and customer satisfaction, emphasizing their continued relevance in the realm

of AI-driven shopping (Pitardi and Marriott, 2021; Sohn, 2017; Choi and Drumwright,

2021).

Secondly, the study contributes to the trust literature concerning interactions with AI‐

based technology. The exploration of trust as a moderator highlights its significance in

shaping consumer behavior (Fernandes and Oliveira, 2021; Hu et al., 2021). The paper

demonstrates that trust acts as an amplifier, strengthening the relationships between AI-

powered VAs and intention to use, customer satisfaction, and emotional attachment. This

novel insight adds depth to the understanding of trust dynamics in the context of emerging

AI-led voice technologies and online shopping.

Thirdly, this study enhances our understanding of brand loyalty in the context of

interactions with AI-powered VAs. The research also sheds light on the fact that brand

loyalty concerns in AI interactions can extend beyond the immediate relationship. While

the existing literature extensively documents the adverse impact of privacy issues on trust
(Pitardi and Marriott, 2021; Fernandes and Oliveira, 2021; Hu et al., 2021), this study

reveals an alternative pathway through which trust and privacy are interconnected in

virtual assistant interactions, focusing not only on antecedents but paying attention also

to consequences/outcomes that can be reached while strengthening intention to use,

customer satisfaction, and emotional attachment.

Additionally, the research conducted as part of the study holds practical implications that

can guide businesses, brand managers, and industries in harnessing the potential of AI-

powered VAs for enhancing customer engagement and brand loyalty.

Firstly, businesses and managers should prioritize improving the perceived usefulness,

ease of use, and overall enjoyment of AI-powered VAs in online shopping. User

experiences that emphasize utility, simplicity, and emotional attachment are more likely

to foster brand loyalty.

Additionally, understanding the role of trust is crucial. Companies should focus on

building trust with customers by ensuring data privacy, security, and transparency in AI

powered VA driven interactions. Trust-building initiatives can include, but need not be

limited to, clear data policies, secure transaction processes, and transparent data handling

practices.

Furthermore, as advancements in virtual assistant technology continue to evolve,

becoming increasingly human-like (or less machine-like), this study offers valuable

insights for designers and managers aiming to build, and sustain over a long time, user

trust in virtual assistants.

This research highlights the crucial role of specific elements in shaping trust during

interactions with virtual assistants. Therefore, it is recommended that producers and

developers concentrate on enhancing consumer trust by refining the conversational flow

of virtual assistants through machine learning and natural language processing. These
technologies can adapt to user preferences, creating a more personalized and socially

engaging experience, ultimately succeeding in establishing the base for brand loyalty.

Although current findings indicate that consumers are considering several facets while

interacting with virtual assistants, strengthening trust in virtual assistants could

potentially stimulate online shopping.

In terms of privacy, the positive perception of virtual assistants is contingent upon

delivering a more personalized user experience that can also be a key asset in establishing

brand loyalty. This relationship allows consumers to receive tailored content while

enabling predictive analytics to optimize marketing efforts, and to anticipate consumer

behavior and spending patterns during online shopping. However, practitioners must

tread carefully, as there exists a delicate balance between acceptable and intrusive levels

of virtual assistants’ usage, requiring ongoing awareness and consideration.

In conclusion, this study has ventured into the realm of AI-powered VAs in online

shopping to shed light on the factors influencing brand loyalty and the pivotal role of

trust. The paper highlights the significance of perceived usefulness, ease of use,

enjoyment, and social cognition, while practical implications guide businesses in

leveraging these insights to enhance customer engagement and brand loyalty.

Acknowledging the research’s limitations and exploring avenues for future research

ensures the ongoing evolution of the understanding of AI-powered VAs in online

shopping and their impact on consumer behavior and brand loyalty. As we navigate the

dynamic landscape of technology and consumer preferences, the pursuit of knowledge

remains principal, offering the promise of even more profound insights in the future.

LIMITATIONS AND FUTURE RESEARCH

While the study provides valuable insights, it is essential to acknowledge its limitations.

First, the research focused primarily on the Indian market, and the findings may not be

fully generalizable to other regions with distinct cultural and economic characteristics.

Second, the study relied on self-reported data from online surveys, which can be subject

to response bias on the part of participants and may not fully capture actual behavior in

real time. Third, the research did not explore the potential influence of other external

variables, such as individual differences and market conditions, which could affect

brand loyalty in the context of AI-powered VAs in online shopping.

The limitations of this study open doors for future research endeavors. This paper aims

to become a starting point to understand how trust affects brand loyalty in the AI-

powered voice commerce ecosystem. Future research could expand the geographical

scope to encompass diverse cultural contexts, enabling a more comprehensive

understanding of AI adoption and brand loyalty worldwide. Exploring the role of

individual differences, such as personality traits and technology readiness, in the

adoption of AI-powered VAs in online shopping and brand loyalty could provide

deeper insights into the intricacies of customer behavior.

Additionally, investigating the impact of market conditions, economic-political setups,

competitive landscapes, and industry-specific factors on brand loyalty in AI-powered

VA-led shopping contexts would enrich the understanding of the multifaceted interplay

between technology and consumer behavior. Further research could also delve into the

ethical and societal implications of AI-powered VAs in online shopping, including

issues related to privacy, data security, and the ethical use of AI technologies in line

with sustainable development goals.

Lists of abbreviations
Artificial intelligence (AI)

Computers as Social Actors (CASA)

Institutional review boards (IRBs)

Internet of Things (IoT)

Natural language processing (NPL)

Research ethics committees (RECs)

Service Robot Acceptance Model (sRAM)

Statistical Package for Social Sciences (SPSS)

Technology Acceptance Model (TAM)

Voice assistant (VA)

REFERENCES

Acikgoz, F., Perez‐Vega, R., Okumus, F., & Stylos, N. (2023). Consumer engagement

with AI‐powered voice assistants: A behavioral reasoning perspective. Psychology &

Marketing, 40(11), 2226-2243.

Akdim, K., & Casaló, L. V. (2023). Perceived value of AI-based recommendations

service: the case of voice assistants. Service Business, 17(1), 81-112.

Ashrafi, D. and Easmin, R. (2023) Okay Google, Good to talk to you… Examining the

determinants affecting users’ behavioural intention for adopting voice assistants: Does

technology self-efficacy matter?’, International Journal of Innovation and Technology

Management, 20(2), 2350004. doi: 10.1142/s0219877023500049.

Aw, E., Tan, G., Cham, T., Raman, R. and Ooi, K. (2022). Alexa, what’s on my shopping

list? Transforming customer experience with digital voice assistants’, Technological

Forecasting and Social Change, 180(5). doi: 10.1016/[Link].2022.121711.

Badghish, S., Shaik, A. S., Sahore, N., Srivastava, S., & Masood, A. (2024). Can

transactional use of AI-controlled voice assistants for service delivery pickup pace in the

near future? A social learning theory (SLT) perspective. Technological Forecasting and

Social Change, 198, 122972.

Bilal, M., Zhang, Y., Cai, S., Akram, U., & Halibas, A. (2024). Artificial intelligence is

the magic wand making customer-centric a reality! An investigation into the relationship

between consumer purchase intention and consumer engagement through affective

attachment. Journal of Retailing and Consumer Services, 77, 103674.

Bouhouita-Guermech, S., Gogognon, P., & Bélisle-Pipon, J. C. (2023). Specific

challenges posed by artificial intelligence in research ethics. Frontiers in Artificial

Intelligence, 6.

Carmel, D., Haramaty, E., Lazerson, A., Lewin-Eytan, L. and Maarek, Y. (2020) Why do

people buy seemingly irrelevant items in voice product search? Proceedings of the 13th

International Conference on Web Search and Data Mining (pp. 79-87).

Cheng, Y. and Jiang, H. (2021). Customer–brand relationship in the era of artificial

intelligence: Understanding the role of chatbot marketing efforts, Journal of Product &

Brand Management, 31(2), 252-264. doi:10.1108/jpbm-05-2020-2907.

Choi, T. and Drumwright, M. (2021). ‘OK, Google, Why do I use you?’ Motivations,

post-consumption evaluations, and perceptions of voice AI Assistants’, Telematics and

Informatics, 62, 101628. doi: 10.1016/[Link].2021.101628.

Choy, L. T. (2014). The strengths and weaknesses of research methodology: Comparison

and complimentary between qualitative and quantitative approaches, IOSR Journal of

Humanities and Social Science, 19(4), 99-104.

Davis, F. (1989) Perceived usefulness, perceived ease of use, and user acceptance of

information technology, MIS Quarterly, [online] 13(3), 319–340. doi:10.2307/249008.

Dogra, P., & Kaushal, A. (2021). An investigation of Indian generation Z adoption of the

voice-based assistants (VBA). Journal of Promotion Management, 27(5), 673-696.

Dujmovic, J. (2017). What’s holding back artificial intelligence? Americans don’t trust

it. [online] MarketWatch. Available at: [Link]

holding-back-artificial-intelligence-americans-dont-trust-it-2017-03-30.

EY India (2023). India@100: reaping the demographic dividend. [online] [Link].

available at: [Link]

Fernandes, T. and Oliveira, E. (2021). Understanding consumers’ acceptance of

automated technologies in service encounters: Drivers of digital voice assistants

adoption, Journal of Business Research, 122, 180–191. doi:

10.1016/[Link].2020.08.058.

Ghorbanzadeh, D., & Rahehagh, A. (2021). Emotional brand attachment and brand love:

the emotional bridges in the process of transition from satisfaction to loyalty. Rajagiri

Management Journal, 15(1), 16-38.

Gillath, O., Ai, T., Branicky, M., Keshmiri, S., Davison, R. and Spaulding, R. (2020).

Attachment and trust in artificial intelligence, Computers in Human Behavior, 115, 1-10.

doi: 10.1016/[Link].2020.106607.

Guha, A., Bressgott, T., Grewal, D., Mahr, D., Wetzels, M., & Schweiger, E. (2023).

How artificiality and intelligence affect voice assistant evaluations. Journal of the

Academy of Marketing Science, 51(4), 843-866.

Haleem, A., Javaid, M., Qadri, M., Singh, R. and Suman, R. (2022). Artificial intelligence

(AI) applications for marketing: a literature-based study, International Journal of

Intelligent Networks, 3(3), 119–132. doi: 10.1016/[Link].2022.08.005.

Hasan, R., Shams, R. and Rahman, M. (2020). Consumer trust and perceived risk for

voice-controlled artificial intelligence: The case of Siri, Journal of Business Research,

131(2). doi: 10.1016/[Link].2020.12.012.

Hu, P., Lu, Y. and Gong, Y. (Yale) (2021). Dual humanness and trust in conversational

AI: A person-centered approach, Computers in Human Behavior, 119(C). doi:

10.1016/[Link].2021.106727.

Hu, P., Lu, Y., Gong, Y. and Ding, A. (2022). Speaking vs. listening? Balance

conversation attributes of voice assistants for better voice marketing, International

Journal of Research in Marketing, 40(1). doi: 10.1016/[Link].2022.04.006.

Huh, J., Kim, H. Y., & Lee, G. (2023). “Oh, happy day!” Examining the role of AI-

powered voice assistants as a positive technology in the formation of brand

loyalty. Journal of Research in Interactive Marketing, (ahead-of-print).

Im, H., Sung, B., Lee, G. and Kok, K. (2023). Let voice assistants sound like a machine:

voice and task type effects on perceived fluency, competence, and consumer attitude,

Computers in Human Behavior, 145(C). doi: 10.1016/[Link].2023.107791.

Jones, E., Watson, B., Gardner, J., & Gallois, C. (2004). Organizational communication:

challenges for the new century. Journal of Communication, 54(4), 722-750.

Kamoonpuri, S. Z., & Sengar, A. (2023). Hi, May AI help you? An analysis of the barriers

impeding the implementation and use of artificial intelligence-enabled virtual assistants

in retail. Journal of Retailing and Consumer Services, 72, 103258.

Kautish, P., Purohit, S., Filieri, R. and Dwivedi, Y. (2023). Examining the role of

consumer motivations to use voice assistants for fashion shopping: The mediating role of
awe experience and eWOM, Technological Forecasting and Social Change, 1-14. doi:

10.1016/[Link].2023.122407.

Ki, C., Cho, E. and Lee, J. (2020). Can an intelligent personal assistant (IPA) be your

friend? Para-friendship development mechanism between IPAs and their users,

Computers in Human Behavior. doi: 10.1016/[Link].2020.106412.

Kumar, R., Sachan, A. and Dutta, T. (2020). Examining the impact of e-retailing

convenience dimensions on behavioral intention: The mediating role of satisfaction,

Journal of Internet Commerce, 19(4), 466–494. doi:10.1080/15332861.2020.1788367.

Kumar, V., Rajan, B., Venkatesan, R. and Lecinski, J. (2019). Understanding the role of

artificial intelligence in personalized engagement marketing, California Management

Review, [online], 61(4), 135–155. Available at:

[Link]

Lakens, D. (2022). Sample size justification, Collabra: Psychology, [online] 8(1), 1–28.

doi:10.1525/collabra.33267.

Lee, S., Oh, J. and Moon, W. (2022). Adopting voice assistants in online shopping:

Examining the role of social presence, performance risk, and machine heuristic,

International Journal of Human–Computer Interaction, 39(14), 1–15. doi:

10.1080/10447318.2022.2089813.

Lehdonvirta, V., Oksanen, A., Räsänen, P. and Blank, G. (2020). Social media, web, and

panel surveys: Using non‐probability samples in social and policy research, Policy &

Internet, 13(1), 134–155. doi:10.1002/poi3.238.

Lu, V. N., Wirtz, J., Kunz, W. H., Paluch, S., Gruber, T., Martins, A., & Patterson, P. G.

(2020). Service robots, customers and service employees: what can we learn from the

academic literature and where are the gaps? Journal of Service Theory and

Practice, 30(3), 361-391.

Malodia, S., Ferraris, A., Sakashita, M., Dhir, A., & Gavurova, B. (2023). Can Alexa

serve customers better? AI-driven voice assistant service interactions. Journal of Services

Marketing, 37(1), 25-39.

Marangunić, N. and Granić, A. (2014). Technology acceptance model: a literature review

from 1986 to 2013, Universal Access in the Information Society, 14(1), 81–95.

doi:10.1007/s10209-014-0348-1.

Maroufkhani, P., Asadi, S., Ghobakhloo, M., Jannesari, M. and Ismail, W. (2022). How

do interactive voice assistants build brands’ loyalty?, Technological Forecasting and

Social Change, 183, 121870.

Monod, E., Lissillour, R., Köster, A., & Jiayin, Q. (2023). Does AI control or support?

Power shifts after AI system implementation in customer relationship

management. Journal of Decision Systems, 32(3), 542-565.

Nardi, P. M. (2018). Doing survey research: A guide to quantitative methods. Routledge.

Newman, I. (2000). A conceptualization of mixed methods: A need for

inductive/deductive approach to conducting research. ERIC. Available at:

[Link]

Okoli, A. (2018). Humanizing digital 2020 - What consumers expect from you today

[Report]. Zoovu Blog. Available at: [Link]

report

Olivier, L. (2011). LEX OLIVIER: River Sampling Nonprobability sampling in an online

environment. LEX OLIVIER. Available at: [Link]

[Link]

Pantano, E. and Scarpi, D. (2022). I, Robot, You, Consumer: Measuring artificial

intelligence types and their effect on consumers emotions in service, Journal of Service

Research, 25(4), 583–600. doi:10.1177/10946705221103538.

Park, N. (2010). Adoption and use of computer-based voice over Internet protocol phone

service: Toward an integrated model. Journal of Communication, 60(1), 40-72.

Partners for Automated Vehicle Education (2020). PAVE Poll: Americans wary of AVs

but say education and experience with technology can build trust | PAVE Campaign.

PAVE [Link]. Available at: [Link]

wary-of-avs-but-say-education-and-experience-with-technology-can-build-trust/.

Pitardi, V. and Marriott, H. (2021). Alexa, she’s not human but… Unveiling the drivers

of consumers’ trust in voice‐based artificial intelligence, Psychology & Marketing, 38(4),

626–642. doi:10.1002/mar.21457.

Poushneh, A. (2021). Humanizing voice assistant: The impact of voice assistant

personality on consumers’ attitudes and behaviors, Journal of Retailing and Consumer

Services, 1-10. doi: 10.1016/[Link].2020.102283.

Rahman, M. S. (2020). The advantages and disadvantages of using qualitative and

quantitative approaches and methods in language "Testing and Assessment" research: A

literature review. Journal of Education and Learning, 6(1), 102-112.

Roslan, F. A. B. M., & Ahmad, N. B. (2023). The rise of AI-powered voice assistants:

Analyzing their transformative impact on modern customer service paradigms and

consumer expectations. Quarterly Journal of Emerging Technologies and

Innovations, 8(3), 33-64.

Sharma, S., Singh, S., Kujur, F. and Das, G. (2020). Social media activities and its

influence on customer-brand relationship: An empirical study of apparel retailers’

activity in India’, Journal of Theoretical and Applied Electronic Commerce Research,

16(4), 602–617. doi:10.3390/jtaer16040036.

Sohn, S. (2017). A contextual perspective on consumers’ perceived usefulness: The case

of mobile online shopping, Journal of Retailing and Consumer Services, [online] 38, 22–

33. doi: 10.1016/[Link].2017.05.002.

Statista (2022). India - e-commerce market size 2014-2027. [online] Statista. Available

at: [Link]

Venkatesh, V., Thong, J. and Xu, X. (2012). Consumer acceptance and use of information

technology: Extending the unified theory of acceptance and use of technology, MIS

Quarterly, 36(1), pp.157–178.

Walter, M., & Andersen, C. (2016). Indigenous statistics: A quantitative research

methodology. Routledge.

Whang, C. and Im, H. (2020). ‘I Like Your Suggestion!’ the role of humanlikeness and

parasocial relationship on the website versus voice shopper’s perception of

recommendations, Psychology & Marketing, 38(4), 581–595. doi:10.1002/mar.21437.

Wilson, M. (2023) Constructing measures [online] Taylor & Francis.

Wirtz, J., Patterson, P., Kunz, W., Gruber, T., Lu, V., Paluch, S. and Martins, A. (2018).

Brave new world: service robots in the frontline, Journal of Service Management, 29(5),

907–931. doi:10.1108/josm-04-2018-0119.

Xiao, Y., & Tan, H. Y. (2024). Why can’t you be all talk to get things done? Consumer

acceptance of voice-assisted products. Journal of Product & Brand Management, 33(1),

125-137.

Yusof, M. Y. P. M., Teo, C. H., & Ng, C. J. (2022). Electronic informed consent criteria

for research ethics review: a scoping review. BMC Medical Ethics, 23(1), 1-11.

THERE IS NO CONFLICT OF INTEREST.

DATA AVAILABILITY STATEMENT: The data that support the findings of this study

are available from the corresponding author upon request.