INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2854
www.rsisinternational.org
AI-Powered Real-Time Recommendations in Livestream Tourism
Marketing: Effects on Customer Engagement and Booking Decisions
- A Systematic Review
Taylor Harris, F. S.
*
& La Geer-Jeremiah, A. O. S.
Central China Normal University, No. 152 Luoyu Road, Hongshan District, Wuhan, Hubei Province,
430079, PRC China
*
Corresponding author
DOI: https://dx.doi.org/10.47772/IJRISS.2025.914MG00220
Received: 02 November 2025; Accepted: 10 November 2025; Published: 22 November 2025
ABSTRACT
The integration of AI-powered real-time recommendations is transforming livestream tourism marketing, yet a
unified understanding of its impact on customer engagement and booking decisions is absent. Following the
PRISMA protocol, this systematic review synthesizes evidence from 15 empirical studies to address this gap.
The analysis is guided by the theoretical frameworks of the Stimulus-Organism-Response (S-O-R) model and
Uses and Gratifications (U&G) theory. The analysis reveals that AI’s power is not direct but psychologically
mediated, operating through a critical human-AI synergy. The streamer’s authenticity is foundational for
building trust, while the AI functions as a real-time decision-support tool, amplifying rather than replacing the
human element. However, a key paradox emerges: while personalization boosts engagement, it risks perceived
intrusiveness and suppresses the serendipitous discovery vital to tourism. We conclude that success depends on
designing for ‘value-centric transparency,’ where AI augments the human connection. Given the review’s
heavy reliance on data from the Chinese market, we critically highlight the urgent need for cross-cultural
validation and a decisive shift from studying behavioral intentions to tracking actual bookings.
Keywords: AI-powered recommendations, livestream commerce, customer engagement, booking decisions,
tourism marketing, real-time personalization
INTRODUCTION
The convergence of artificial intelligence (AI) and livestream marketing is redefining the digital commerce
landscape and unlocking new opportunities for personalized customer interaction. In the tourism industry,
which is inherently experience-based and information-rich (Kotler, 2022; Tussyadiah, 2020), this synergy
holds particular promise for transforming how destinations and services are marketed and sold (López Naranjo
et al., 2025; Zhang et al., 2025). Livestream commerce, defined by its real-time, video-centric, and interactive
format, has emerged as an effective tool for mitigating the intangibility of tourism products. This allows
potential customers to experience a hotel room, tour, or destination before booking (Liu et al., 2022;
Moghddam et al., 2025). The integration of AI-powered recommendation systems into these livestreams
represents the next evolutionary step, aiming to hyper-personalize the viewer’s journey by analyzing their real-
time behavior and preferences to suggest relevant products instantly (Li & Zheng, 2025).
To conceptualize the impact of this technological integration in examining how AI-powered recommendations
influence customer engagement and decision-making in livestream tourism marketing, this study is guided by
two foundational theoretical frameworks: the Stimulus-Organism-Response (S-O-R) model and the Uses and
Gratifications (U&G) theory. The S-O-R model, developed by Mehrabian and Russell (1974), as cited in
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2855
www.rsisinternational.org
Hochreiter et al. (2022), posits that external environmental stimuli (S), such as AI-powered recommendations,
influence a user’s internal cognitive and affective states (O), such as cognitive engagement or emotional
arousal, which in turn drive their behavioral responses (R), such as a booking decision or sharing the stream.
Meanwhile, U&G theory (Katz et al., 1973), as cited in (Sichach, 2023), provides a lens to understand why
audiences choose specific media. In this context, it helps explain the gratifications users seek from livestream s
(for example, information, entertainment, social interaction) and how effectively AI recommendations fulfill
these needs to foster deeper engagement. Together, these theories provide a robust framework for analyzing
how AI-driven stimuli within a livestream platform influence the viewer’s psychological processes and
subsequent behaviors.
Existing research has shown that traditional, non-real-time recommendation systems in e-commerce can
significantly influence click-through rates and sales conversions by reducing information overload and
simplifying choice (Lo et al., 2021). Concurrently, a growing body of literature has begun to document the
persuasive power of livestream commerce itself, highlighting its ability to generate trust through streamer
authenticity, foster a sense of social presence, and trigger impulse purchases through real-time interactivity like
coupons and limited-time offers (Chen et al., 2024; Moghddam et al., 2025). However, research on algorithmic
recommendations and livestream dynamics has largely developed separately. Studies on livestream marketing
often treat the streamer’s recommendations as a manual, human-driven element, while research on AI
recommendations predominantly focuses on static webpages or app interfaces. This has created a significant
conceptual gap: The synergistic effects of AI-powered, real-time recommendations within the dynamic,
immersive, and social context of livestream s remain poorly understood (Moghddam et al., 2025; Qu et al.,
2025).
This gap is particularly critical in the tourism industry, where products are high-involvement, complex, and
require significant consumer trust (Cohen et al., 2014). While an AI might successfully recommend a low-cost
consumer good in real-time, the mechanisms through which it influences a decision to book a costly vacation
package, a decision fraught with higher perceived risk, are likely to be far more complex. This complexity
requires AI systems capable of processing multi-attribute preferences (for example, travel dates, budget,
family-friendly amenities, accessibility needs) in real-time (Seabra et al., 2020). Additionally, the influence of
these recommendations is mediated by distinct factors, such as trust in the algorithm’s accuracy for high-stakes
decisions, the credibility of the streamer endorsing the AI’s suggestion, and the emotional atmosphere of the
live chat (Cheng & Jiang, 2022; Tussyadiah, 2020). Furthermore, the specific metrics for customer
engagement in this hybrid context remain undefined, oscillating between behavioral measures (clicking the
recommendation, using a real-time coupon), cognitive measures (prolonged viewing time, querying the AI),
and affective measures (expressed excitement in chat). A systematic synthesis is required to disentangle these
effects and mechanisms.
Therefore, this study addresses a crucial research void by systematically investigating the integration of AI-
powered real-time recommendations within livestream tourism marketing. The significance of this research
lies in its potential to provide a unified theoretical understanding of how this emerging technology influences
the customer decision journey in a high-stakes industry. By synthesizing existing but fragmented evidence, this
review aims to move beyond the isolated examination of either technology or medium and instead focus on
their powerful intersection. Consequently, this systematic review is guided by the following research question:
“How do AI-powered real-time product recommendations in livestream marketing impact customer
engagement and booking decisions in tourism?”
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2856
www.rsisinternational.org
METHODOLOGY
This study’s methodology aimed to systematically investigate the impact of AI-powered real-time
recommendations on customer engagement and booking decisions in livestream tourism marketing. To
synthesize the existing but fragmented evidence on this emerging topic, the study utilized a systematic
literature review (SLR) approach, which is known for its rigor, transparency, and replicability in identifying,
evaluating, and interpreting relevant research and adhering to the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA 2020) guidelines (Page et al., 2021). A systematic review involved
formulating a precise research question, developing predefined eligibility criteria, conducting a comprehensive
search across multiple databases, systematically extracting data, and critically appraising the quality of
included studies. This approach is well-suited for exploring technological interventions like AI-powered
recommendations and understanding their impact on user behavior and decision-making. Recognized as the
highest level of evidence in evidence-based practice (Brignardello-Petersen et al., 2024; Howie, 2019), this
approach was ideal for consolidating findings, identifying overarching themes and gaps, and offering a unified
theoretical understanding of the topic.
The Search Strategy
The search strategy involved a comprehensive and systematic search conducted in March 2025 across four
major academic databases: Scopus, Web of Science, EBSCOhost Business Source Complete, and ACM Digital
Library. Filters for subject areas such as Business, Management, Marketing, Tourism, Hospitality, and
Computer Science were applied where applicable to ensure results were precisely targeted to the
interdisciplinary nature of the topic. The search targeted peer-reviewed journal articles using English keywords
and phrases. These included core concepts such as “AI-powered recommendations,” “real-time
personalization,” “livestream commerce,” “customer engagement,” “booking decisions,” and “tourism
marketing.” Boolean operators (AND, OR) were used to combine these terms into complex query strings to
balance breadth with precision, for example, (“AI recommend*” OR “algorithmic recommend*”) AND
(“livestream” OR “livestream commerce”) AND (tourism OR “hospitality industry”). The search strategy was
limited to studies published between 2020 and 2025 to capture the evolution of the technology from its
emergent phase to its current state. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) guidelines were followed to document and visualize the screening, inclusion, and exclusion
processes.
Inclusion and Exclusion Criteria
Studies were included if they (1) were peer-reviewed journal articles published between 2020 and 2025 in
English; (2) employed empirical methods (quantitative, qualitative, or mixed methods); (3) focused on AI-
powered recommendation systems within livestream tourism marketing; (4) explicitly examined their effects
on customer engagement, booking decisions, or related behavioral and psychological outcomes; and (5)
addressed real-time personalization and interactive features specific to livestream commerce in a tourism or
hospitality context. While the search strategy retrieved some high-quality systematic reviews, these were used
solely for contextual analysis and were excluded from the final systematic and quality appraisal, which focused
on primary research.
Studies were excluded if they (1) focused on non-tourism sectors, non-livestream platforms, or AI applications
unrelated to recommendation systems; (2) were non-peer-reviewed (for example, conference abstracts, grey
literature, or commentaries); (3) reported insufficient sample sizes (n < 20), unclear methodology, or
unverifiable data; or (4) were published outside the 2020–2025 timeframe.
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2857
www.rsisinternational.org
Id
en
tifi
ca
tio
n
Records identified through database search
Scopus
Web of Science
EBSCOhost Business
Source Complete
ACM Digital Library
n = 100
n = 100
n = 80
n = 60
Total
n = 340
Reports assessed for eligibility
(n = 112)
Reports excluded after full text
reviews.
(n = 70)
Full-text assessed for
eligibility
(n = 42)
Reports excluded based on
inclusion and exclusion
criteria.
(n = 27)
In
cl
ud
ed
Full-texts included
(n = 15)
Fig. 1 Flow diagram of search results [in line with PRISMA Guidelines (PRISMA’s Four Phases Flow
Information Diagram of a Systematic Review)] (Page et al., 2021)
Sc
re
en
in
g
Records after duplicate removed
(n = 209)
Records screened
(n = 209)
Records deemed irrelevant after
reading the titles and abstract of
articles.
(n = 97)
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2858
www.rsisinternational.org
Ensuring Quality and Credibility
Following the identification and retrieval of 15 pertinent primary research articles using the PRISMA protocol,
a systematic quality appraisal was undertaken to ensure the review’s findings were grounded in robust and
credible evidence. The Standard Quality Assessment Criteria for Evaluating Primary Research Papers from a
variety of Fields (Han & Cumming, 2022; Kmet et al., 2004) was applied to all included studies. This tool is
adaptable for both qualitative and quantitative research. For mixed-methods studies, the qualitative and
quantitative components were assessed separately using the corresponding Kmet Checklists, and the average of
both scores was recorded as the final quality rating.
Each of the 15 studies was independently reviewed by the authors, who assessed whether the criteria were
fully met, partially met, or not met. Discrepancies in scoring were resolved through collaborative discussion
until consensus was reached. Scores for each study were calculated as a proportion of the total possible score,
converted into a fraction (For example, 9/10) for clarity in the summary table
The classification of quality followed the threshold adapted from (Han & Cumming, 2022; McGarty &
Melville, 2018): Strong: > 7.5/10, Moderate: 5.5–7.5/10, and Weak: < 5.5/10. Of the 15 included articles, 11
were rated strong (Scores from 8/10 to 9.5/10), 3 were rated moderate (scores between 6/10 and 7/10), and 1
was rated weak (score of 5/10). This quality appraisal served not as an exclusion criterion, but as a means to
contextualize the weight of evidence in the synthesis (Munn et al., 2020; Page et al., 2021). The high
proportion of strong studies supports the overall validity and reliability of the review’s conclusion.
The inclusion of one study rated as weak (Zhang & Wu, 2025) was justified by its relevance to the core
research question, providing insights into the adoption pathways for recommendations. To ensure robustness, a
sensitivity analysis was conducted by excluding this study. The four primary themes remained fully supported
by the remaining 14 studies, confirming the stability of our conclusions
Data Collection Process and Data Items
The authors independently reviewed all 15 studies included in this systematic review. Data extraction was
conducted systematically using a standardized protocol to ensure consistency and rigor across all studies.
Relevant data were extracted from each article’s abstract, methodology, results, and discussion sections, with
particular attention to disaggregated findings related to AI-powered recommendations and their effects. Each
study was coded according to key analytic categories relevant to the research questions, including: (a) study
characteristics (author, year, country, research design); (b) sample characteristics and context (participant
demographics, tourism sector, livestream platform); (c) AI recommendation system features (type of
algorithm, personalization approach, real-time capabilities); (d) measurement approaches for customer
engagement (behavioral, cognitive, affective dimensions); (e) measurement of booking decisions (intentions,
conversions, actual bookings); and (f) key findings regarding impacts and mediating factors. Discrepancies
between independent extractions were resolved through collaborative discussion and consensus among the
authors. To facilitate comprehensive analysis, the extracted data were synthesized into summary tables that
organize core insights across methodological approaches and contextual variables, enabling clearer comparison
of findings across the current landscape of AI-powered recommendation research in livestream tourism
marketing.
Research Results
The systematic review process, conducted in accordance with the PRISMA 2020 guidelines (Page et al., 2021),
culminated in the inclusion of 15 primary research articles for final synthesis and analysis. The methodological
distribution of these studies was as follows: 11 quantitative, 1 qualitative, and 3 mixed-methods. The
quantitative studies predominantly utilized cross-sectional surveys, structural equation modeling (SEM), and
experimental designs to investigate the causal relationships and mediating mechanisms between AI-powered
recommendations, customer engagement metrics, and booking intentions. The qualitative and mixed-methods
studies provided richer, contextual insights into industry implementation challenges, user perceptions, and the
nuanced interplay between AI and human elements in the livestream environment. This multi-method corpus
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2859
www.rsisinternational.org
of literature produced a comprehensive perspective on how AI-driven stimuli influence the viewer’s
psychological journey and behavioral outcomes in tourism livestream s.
Demographic Results
Table 1 A consolidated overview of the demographic and contextual characteristics of the participants and
studies included in this systematic review.
Descriptor
Characteristic
Total Number of Participants
~ 6,300+ (from self-reported sample sizes)
Quantitative Study Participants
~ 5,800+
Qualitative Study Participants
27 (industry experts)
Mixed-Methods Study Participants
~ 500+ (consumers & industry experts)
Geographical Context
China-based studies; also includes South Korea and
conceptual/modeling studies without a specific country context.
Primary Livestream Platforms
Dominantly Douyin (TikTok), Taobao, and JD.COM; also
includes Ctrip.com and other Online Travel agencies (OTA) platforms.
Participant Profile
Primarily, consumers and users of livestream e-commerce platforms
who have engaged with tourism or hospitality content. Also includes
industry experts (AI specialists, marketers) and live streamers.
Tourism Sector Context
Hotels, destinations, tour packages, and general hospitality services.
Implication
The heavy focus on the Chinese market indicates a mature research and
commercial environment for livestream tourism. This highlights a
significant gap in understanding its application and effects in other
cultural and technological contexts. The participant profile underscores
the
need
to
view
users
as
both
media
consumers
and
potential
customers.
Limitations of the Included Literature
This synthesis is constrained by several limitations inherent in the primary literature. The most significant is
pronounced geographic bias, with most studies conducted in China using platforms like Douyin and Taobao.
This limits generalizability to Western markets where consumer behavior, privacy expectations, and cultural
dimensions (for example, individualism-collectivism) may fundamentally alter the human-AI synergy and
trust-transfer mechanisms identified.
Additionally, heavy reliance on self-reported behavioral intentions creates an intention-behavior gap, as only
one study measured actual conversions. Methodologically, the predominance of cross-sectional surveys (11 of
15 studies) risks self-selection and social desirability biases, while limited experimental designs constrain
causal inference about AI recommendation effects specifically.
Consequently, findings should be interpreted as primarily reflective of Chinese contexts and behavioral
intentions until validated cross-culturally with actual booking data.
Findings from the Analysis
The synthesis of 15 studies revealed a cohesive set of thematic findings that directly address the central
research question: “How do AI-powered real-time product recommendations in livestream marketing impact
customer engagement and booking decisions in tourism?” The analysis indicates that the impact is not
monolithic but operates through a series of interconnected psychological, social, and technical
mechanisms, which collectively illustrate that AI's power is not in replacing human elements but in
augmenting them within a psychologically nuanced, trust-mediated environment. Four primary themes
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2860
www.rsisinternational.org
emerged: (1) The Primacy of the S-O-R Psychological Mechanism, (2) The Imperative of Human-AI Synergy,
(3) Trust as the Critical Mediator, and (4) The Dual-Edged Sword of Personalization. These themes, examined
sequentially below, provide a unified theoretical understanding of how AI recommendations influence the
customer journey in livestream tourism.
Table 2: Summary of 15 Reviewed Studies on AI-Powered Real-Time Recommendations in Livestream
Tourism Marketing
Author(s)
(Year)
Participants/C
ontext
Methodology/
Design
Key Findings on AI
Recommendations,
Customer Engagement &
Booking Outcomes
Identified
Challenges
Quality
Assessment
Zheng et
al. (2022)
260 Chinese
users who
watched a
destination
livestream in
the past three
months
Quantitative;
online survey;
Structural
Equation
Modeling
(SEM)
Establishes the S-O-R
model in livestreams,
demonstrating that
technological stimuli (e.g.,
interactivity, immediacy)
enhance trust and presence
to drive travel intention.
This provides the
foundational mechanism
through which AI-powered
recommendations, as a key
stimulus, are theorized to
influence engagement and
decisions
Varying effects
based on
destination type
and viewer
segment
Strong
(8.5/10)
Mei et al.
(2025)
1,084 consumers
from major
Chinese
livestream e-
commerce
platforms
(Douyin,
Taobao, JD.com)
Quantitative;
online survey;
principal
component
analysis and
mediation
regression
Identified key AI-driven
factors (product quality, AI
marketing strategy, anchor
ability) that enhance
consumer trust, mediating
purchase intentions with
trust levels ranging from
66.7% to 98.3%.
Inaccurate AI-
generated data
can undermine
trust and
influence
purchasing
decisions.
Strong
(9/10)
Zhang and
Wu (2025)
382 users of
Chinese
livestreaming
platforms
Quantitative;
cross-sectional
survey analyzed
using Hayes’
PROCESS
macro
This study
of streamer recommendatio
ns identifies perceived
value and credibility as
universal pathways for
recommendation
adoption, providing a
critical framework for how
AI-powered systems must
be designed to earn user
trust and demonstrate value
The conflation
of streamer-
driven and AI-
powered
recommendatio
ns complicates
adoption.
Weak (5/10)
Wu and
Yusof
(2024)
Analysis of
multi-platform
user data (n=93)
from e-
commerce,
streaming, and
social media
Quantitative,
simulation-
based offline
evaluation of a
Multi-behavior
Streaming
Recommender
System using
The proposed real-time,
multi-behavior AI
recommendation system
(MbSRS) demonstrates
strong efficacy,
outperforming a baseline
system with higher
accuracy (Precision:
0.82,
Adapting and
validating the
multi-behavior,
real-time
recommendatio
n framework
from standard
e-commerce to
Moderate
(7/10)
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2861
www.rsisinternational.org
key metrics
Recall: 0.68, NDCG: 0.74),
a 10.5% higher conversion
rate (4.20% vs. 3.80%),
and a low latency of 120ms
suitable for live interaction
effectively
support the
high-
involvement,
trust-sensitive
booking
decisions
characteristic of
tourism within a
livestream
environment
Ding et al.
(2025)
Douyin users in
China.
Quantitative;
simulation-
based offline
evaluation of
the Multi-
behavior
Streaming
Recommender
System
(MbSRS)
AI-driven proactive
recommendations enhance
engagement by aligning
product discovery with
entertainment, thus
increasing purchase intent
Intrusive
personalization
and privacy
concerns can
hinder
engagement;
influencer trust
can alleviate
skepticism
Strong
(9/10)
Xie et al.
(2022)
Tourism live
streamers on
Ctrip.com; 608
consumers
Mixed-
methods;
content analysis
and a large-
scale PLS-SEM
survey
Identifies the key value
attributes (e.g., product
value, emotional
connection, service
convenience) that drive
purchase intentions in
tourism livestream s. This
framework outlines
the critical value
propositions that AI-
powered recommendations
must fulfill to enhance
customer engagement and
drive bookings
Balancing
consistency and
authenticity in
value delivery
to avoid
skepticism
Strong
(8.5/10)
Zhou et al.
(2025)
Conceptual
analysis within
a duopoly
competition
model between
an online
retailer using
livestreaming
vs. a brick-and-
mortar (B&M)
retailer
Quantitative
game theory
analysis of
human vs. AI
livestreaming in
duopoly
competition.
AI livestreaming is
strategically advantageous
in situations with high
consumer hassle costs, as it
streamlines the process. It
offers a competitive edge
and cost-effectiveness in
moderate hassle scenarios,
with the choice between
human and AI hosts being
a strategic decision based
on costs and consumer
barriers
The need for
strategic
alignment
between
livestreaming
formats
(Human/AI)
and market
conditions to
avoid negative
profitability,
requiring
further
empirical
validation in
tourism
Moderate
(7/10)
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2862
www.rsisinternational.org
contexts
Liu et al.
(2024)
1,257 total
participants
across three
experimental
studies on
tourism E-
commerce
Livestreaming
(ECLS)
Quantitative
experimental
study using
three factorial
designs to test
causal effects of
livestreaming
features
(n=1,257)
AI must complement, not
replace, core human-
streamer elements. The
interaction between a
streamer's personal brand
equity and the telepresence
they create significantly
drives booking intention,
operating through an S-O-
R mechanism where
streamer credibility and
broadcast immersion are
crucial internal states.
Effectiveness is contingent
on this specific human-
centric configuration
Integrating AI
recommendatio
ns without
disrupting the
human-centric
streamer-viewer
relationship and
telepresence
that drive
engagement
Strong
(8.5/10)
Tolloso et
al. (2025)
Large-scale
fashion e-
commerce data
from H&M;
1.3M users,
105K articles,
31M
transactions,
using simulated
streaming
scenarios
Quantitative
methodology:
heterogeneous
graph neural
network with
contrastive
learning and
model
distillation for
real-time
recommendatio
n.
Performance: The full
model (HGNN + 1-week
personalization) achieved a
77.7% higher F1-score
than LightGCN and 38.9%
higher than PinSage on the
proprietary dataset.
Efficiency: Recommendati
ons generated in ~1.5ms
(CPU) with a 700KB
memory footprint per user.
Model adaptation takes
<100ms.
Ablation Study: Confirmed
that both HGNN pre-
training (for cold-start) and
continual personalization
are critical for optimal
performance. Visual
features (CNN) were found
to be more critical for
recall than global
interaction patterns alone
A short model
lifespan (1-2
weeks), data
sparsity
requiring multi-
behavior inputs,
and complex
dual-model
architecture.
Moderate
(7/10)
Chen and
Wei (2024)
Questionnaire
data from
travelers to
assess factors
influencing
travel decisions
and tourism
sales
Quantitative
(questionnaire,
ANOVA)
AI-based recommendations
and informativeness
significantly influence
travel decision-making and
tourism product sales
Ensuring AI-
driven
recommendatio
ns provide
sufficient
informativeness
and value, as
social media
alone showed
Strong
(8.5/10)
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2863
www.rsisinternational.org
no significant
impact on travel
decisions
Zhu et al.
(2023)
Surveyed 566
customers in
China with
experience
using AI
chatbots on
Online Travel
Agency (OTA)
platforms.
Quantitative
(survey,
structural
equation
modeling)
AI chatbot interactivity and
information quality
increase trust and purchase
intention in online travel
settings
The
effectiveness of
AI chatbots in
building trust is
significantly
lower for
customers who
are unfamiliar
with the tourism
products being
offered
Strong
(8.5/10)
Bulchand-
Gidumal et
al. (2024)
Surveyed 300
industry experts
and consumers
within the
hospitality and
tourism sectors
regarding the
impact of AI on
customer
experience and
engagement.
Mixed-methods
(qualitative
interviews with
industry
experts,
quantitative
surveys with
consumers)
AI-driven, real-time
personalized
recommendations enhance
customer engagement,
drive booking decisions,
and increase loyalty by
tailoring suggestions based
on individual preferences.
High costs of
data acquisition,
privacy
concerns, and
challenges in AI
integration
within existing
systems limit its
widespread
adoption in
tourism
marketing.
Strong
(9/10)
Kim (2023)
Surveyed 575
South Korean
consumers who
engaged with
travel
livestreaming
content in the
past three
months
Quantitative
(survey,
structural
equation
modeling)
The study establishes that
human streamer attributes
are the primary driver of
trust and
booking, highlighting the
crucial human-centric
context that AI-powered
recommendations must be
integrated into without
disrupting the streamer-
viewer dynamic
Integrating AI
tools without
diminishing the
central role of
the human
streamer’s
credibility and
relatability
Strong
(9/10)
Antczak
(2025)
Analyzed 20
industry reports,
academic
articles, and
case studies
related to AI
adoption in
tourism
marketing
Qualitative
(systematic
literature
review, case
study analysis)
AI in tourism marketing
enhances personalization,
customer engagement, and
operational efficiency,
offering tailored
experiences and optimizing
marketing strategies
through data analytics and
predictive tools
Ethical
concerns,
privacy issues,
and the need for
transparency in
AI usage; high
infrastructure
costs, and the
shortage of
skilled
Strong
(9/10)
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2864
www.rsisinternational.org
personnel for
implementing
AI systems
(Philp &
Nepomuce
no, 2024)
Participants/Co
ntext: 387 users
of major online
travel platforms
(for example,
Booking.com,
Expedia).
Quantitative;
online survey
analyzed using
Partial Least
Squares
Structural
Equation
Modeling (PLS-
SEM).
AI-powered
personalization
significantly increases
booking intentions by
boosting customer
engagement, but this effect
is contingent on effectively
mitigating users' privacy
concerns.
The
personalization-
privacy
paradox, where
the data
collection
needed for
effective AI
recommendatio
ns risks eroding
user trust and
negating
engagement
benefits.
Strong
(8.5/10)
The Primacy of the S-O-R Psychological Mechanism
The synthesis of the reviewed literature strongly affirmed the Stimulus-Organism-Response (S-O-R) model as
a robust theoretical anchor for deciphering the psychological impact of AI-powered recommendations in
livestream tourism. The findings delineate a clear causal pathway: AI-generated suggestions serve as potent
external stimuli (S), characterized by their real-time delivery and interactive nature (Wu & Yusof, 2024). These
stimuli directly influence the viewer’s internal states (O), triggering critical cognitive assessments (for
example, perceived relevance and value) and affective reactions (for example, emotional arousal and a sense of
telepresence) (Liu et al., 2024; Zheng et al., 2022). These heightened internal states, in turn, are the primary
drivers of behavioral responses (R), most notably increased engagement metrics and the intention to book (Mei
et al., 2025).
For instance, (Zheng et al., 2022) empirically established that technological stimuli in livestream s, a category
to which AI recommendations belong, enhance trust and social presence, which subsequently drive travel
intention. This provides a foundational mechanism for AI’s role. Further supporting this, (Mei et al., 2025)
demonstrated that the telepresence and credibility generated by a human streamer (an ‘Organism’ state) are
crucial mediators for booking intention, underscoring that AI recommendations operate within a broader
psychological environment.
The critical contribution of the S-O-R framework, as revealed by this synthesis, is its ability to unpack the
“black box” of customer decision-making. It moves beyond establishing a direct link between AI stimuli and
booking outcomes, instead illuminating the essential intermediary psychological processes, particularly trust,
emotional connection, and perceived authenticity, that must be successfully activated to convert a real-time
recommendation into a high-involvement purchase decision in tourism.
The Imperative of Human-AI Synergy
A central finding of this synthesis is that the efficacy of AI-powered recommendations is contingent upon
their complementary integration within the human-centric social environment of the livestream. The reviewed
literature consistently positions the human streamer, not the AI, as the foundational element for building trust
and para-social rapport, which are prerequisites for high-involvement booking decisions in tourism (Kim,
2023; Mei et al., 2025). The credibility, authenticity, and personal brand equity of the streamer create a
telepresence that immerses the viewer, establishing a psychological context in which AI suggestions can be
effectively received (Liu et al., 2024).
Within this context, AI and the human streamer assume distinct but interdependent roles. The AI system’s
value lies in its computational power to process multi-attribute user data in real-time, offering personalized
Page 2865
www.rsisinternational.org
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
suggestions that mitigate choice overload and enhance relevance (Wu & Yusof, 2024). Conversely,
the streamer’s value lies in their ability to authentically curate, endorse, and contextualize these AI-generated
options, lending them social credibility and emotional weight (Zhang & Wu, 2025). This division of labor was
exemplified in findings by Ding et al. (2025), where AI-driven proactive recommendations enhanced
engagement only when they aligned with the stream’s entertainment value and were perceived as being vetted
by a trusted influencer.
The imperative for synergy arises from the risk of disruption. When AI recommendations are perceived as
intrusive or overshadow the streamer-viewer relationship, they can undermine the very trust they are designed
to exploit (Liu et al., 2024). Therefore, the optimal model emerging from the evidence is not a balance of
power, but a hierarchical collaboration: the human streamer acts as the trusted curator and emotional anchor,
while the AI serves as a powerful, real-time decision-support tool that amplifies the streamer’s ability to cater
to individual viewer needs, ultimately leading to more persuasive and satisfying customer journeys.
Trust as the Critical Mediator
The synthesis unequivocally identifies trust as the critical and multifaceted mediator governing the pathway
from AI-powered recommendations to customer engagement and booking decisions. The findings reveal that
the relationship is not direct. Instead, AI stimuli influence customer behavior through a dual-object trust
dynamic, comprising trust in the AI’s competence and trust in the human streamer’s authenticity (Mei et al.,
2025). The high-perceived risk inherent in tourism purchases makes this mediating role particularly
pronounced, as customers are highly sensitive to signals of reliability and honesty.
The findings reveal that trust in the AI algorithm is contingent upon its perceived accuracy and transparency.
For instance, Wu and Yusof (2024) demonstrated that a high-performance AI system directly increased
conversion rates, an outcome predicated on users trusting its outputs. Conversely, Zhu et al. (2023) found that
inaccurate or intrusive recommendations erode this trust, inducing skepticism. Crucially, this synthesis shows
that trust in the often-opaque AI is frequently calibrated through the trusted human channel. The credibility of
the streamer who endorses, contextualizes, or even simply presents the AI’s suggestion serves as a powerful
heuristic for viewers, allowing trust to transfer from the streamer to the technology (Kim, 2023; Zhang & Wu,
2025).
The mediating power of trust is quantifiable. Mei et al. (2025) reported that trust mediated purchase intentions
with a substantial influence, with path coefficients indicating it explained a significant portion of the variance
in decision-making. Therefore, this review establishes that trust is not merely a facilitator but a non-negotiable
prerequisite. It acts as the psychological gatekeeper that modulates cognitive and affective responses,
determining whether an AI-powered stimulus will result in mere viewing behavior or will successfully
transition into a high-involvement booking decision in the livestream context.
The Dual-Edged Sword of Personalization
This synthesis revealed a paradox: AI-driven personalization is a powerful tool for engagement, but its
effectiveness has limits. Beyond a certain threshold, it can lead to negative psychological reactions. On one
hand, the reviewed evidence confirms that value-adding personalization, where recommendations are highly
aligned with a viewer’s demonstrated preferences (for example, travel dates, budget), significantly boosts
cognitive engagement and purchase intent (Ding et al., 2025). This efficacy stems from its ability to reduce
decision-making fatigue and create a sense of exclusive relevance, seamlessly integrating product discovery
into the hedonic flow of the livestream.
On the other hand, the literature identifies the risks of exceeding this threshold. When personalization feels
intrusive or mechanical, it can raise privacy concerns and create a sense of being monitored, which may erode
the trust necessary for making tourism purchases (Philp & Nepomuceno, 2024; Xie et al., 2022). Furthermore,
the review identifies a less obvious but critical risk: the suppression of serendipity. In an experience-driven
industry like tourism, the over-optimization of recommendations can create a “filter bubble,” limiting exposure
to inspiring, novel destinations or activities that fall outside a user’s immediate preference history but are
central to the allure of travel (Zhou et al., 2025).
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
Page 2866
www.rsisinternational.org
Therefore, the central challenge identified by this synthesis is not merely one of technical calibration, but
of strategic design. The optimal application of personalization in livestream tourism is by moving beyond
simple data usage to a model that focuses on understanding context and delivering value-driven customization.
This involves designing AI systems that transparently enhance the viewer’s journey without encroaching
on the viewer’s autonomy, and that strategically balance the presentation of perfectly matched options with the
curated introduction of serendipitous discoveries, thereby preserving the emotional excitement and exploratory
nature of travel planning.
DISCUSSION
This systematic review synthesized evidence from 15 empirical studies to address the research question: How
do AI-powered real-time product recommendations in livestream marketing impact customer engagement and
booking decisions in tourism? The findings reveal that this impact is not a direct causal relationship but a
complex, mediated process. The subsequent discussion interprets these synthesized findings through a
theoretical lens, elucidating the central psychological mechanisms, reconciling the imperative of human-AI
synergy, and examining the critical, dual-faced role of trust and personalization.
The findings of this systematic review collectively demonstrate that the influence of AI-powered
recommendations in livestream tourism is not a simple input-output phenomenon, but a complex psychological
journey, which the S-O-R model compellingly explains. The confirmation of this model’s robustness in this
context is a critical theoretical contribution (Liu et al., 2024; Mei et al., 2025; Zheng et al., 2022). It moves the
discourse beyond a technocentric view of AI efficacy to a psychographic one, illustrating that the real-time,
interactive nature of AI recommendations serves as a powerful environmental stimulus (S) that must first be
processed through the user’s internal world (O). This synthesis reveals that AI’s success is contingent on its
ability to trigger not just cognitive assessments of relevance, but also affective states of telepresence and
emotional arousal (Liu et al., 2024; Zheng et al., 2022), which are particularly vital for intangible, high-
involvement tourism products (Cohen et al., 2014). This aligns with and extends the Uses and Gratifications
theory, suggesting that effective AI recommendations fulfill deeper viewer needs for both personalized
information and immersive entertainment, thereby fostering the engagement that precedes a booking decision
(R). Ultimately, this review positions the S-O-R model as an indispensable framework for understanding how
technological features translate into psychological value and, consequently, commercial outcomes in the
dynamic context of livestream marketing.
One might assume that the most advanced AI would dominate the livestream, yet our synthesis challenges this
narrative of algorithmic dominance by revealing a non-negotiable synergy with the human streamer (Kim,
2023; Liu et al., 2024; Mei et al., 2025). This raises a critical question: if AI cannot build trust de novo, how
should its role be fundamentally designed? This finding can be interpreted through the lens of para-social
interaction theory, which explains the one-sided, intimate feeling a viewer develops for a media persona. The
reviewed evidence suggests that the human streamer is the primary vessel for this para-social relationship,
generating the authenticity and trust that AI, as a purely logical entity, cannot manufacture de novo (Kim,
2023; Liu et al., 2024). The AI, in this context, functions not as a replacement but as a force multiplier for the
streamer’s effectiveness (Ding et al., 2025; Liu et al., 2024). While this dynamic appears to align with
the “Computers Are Social Actors” (CASA) paradigm, our synthesis critically refines it. We must ask: Does
CASA hold when a human is present? Our findings suggest that in a livestream setting, the social credibility
of the human actor is not just a factor, but a prerequisite for the AI to be accepted as a competent “social” actor
(Kim, 2023; Zhang & Wu, 2025). The identified model of “hierarchical collaboration”, where the streamer
curates and the AI calculates, provides a strategic blueprint for designing these systems (Ding et al., 2025; Liu
et al., 2024; Wu & Yusof, 2024). It asserts that the most effective AI is one that operates transparently in
service of enhancing the human connection, rather than seeking to supplant it, thereby ensuring that
technological advancement does not erode the fundamental social drivers of livestream engagement and
purchase decisions. The non-negotiable synergy between human streamers and AI, a central tenet of this
review, was established within a specific cultural context. The para-social intensity and authority granted to
streamers in East Asian markets may be a key reason for their foundational role. It is a critical open question
whether this ‘hierarchical collaboration’ model holds with the same force in Western cultures, where audiences
Page 2867
www.rsisinternational.org
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
might place different values on authenticity versus algorithmic efficiency, and where influencer authority may
be less pronounced.
The identification of trust as the critical mediator and “psychological gatekeeper” elevates it from a peripheral
concern to the central construct governing the entire customer journey in AI-infused livestreams (Mei et al.,
2025; Zhu et al., 2023). This finding powerfully resonates with and extends Morgan and Hunt (1994)
commitment-trust theory of relationship marketing, which posits trust as the key mediator between a partner’s
actions and the relationship’s outcomes. In this context, the “partners” are the viewer, the streamer, and the AI
system. The synthesis reveals a sophisticated “trust transfer” mechanism, whereby the high-credibility, para-
social bond with the human streamer serves to bootstrap trust in the otherwise opaque AI algorithm (Kim,
2023; Zhang & Wu, 2025). This process is crucial for overcoming the “black box” problem often associated
with complex AI, where users cannot scrutinize the logic behind a recommendation. The dual-object nature of
trust, in the streamer's authenticity and the AI’s competence, means that a failure in either dimension can sever
the pathway to a booking decision (Liu et al., 2024; Wu & Yusof, 2024; Zhu et al., 2023). Consequently, this
review contends that the primary function of the livestream’s socio-technical environment is not merely to
inform or entertain, but to actively construct and maintain this fragile ecosystem of trust. This reframes the key
performance indicator for AI in this domain; it is not just algorithmic accuracy, but the system’s
demonstrable contribution to building a trustworthy experience that mitigates the high perceived risks of
tourism purchases. The identified ‘trust-transfer” mechanism, where the streamer’s credibility bootstraps trust
in the AI, must be critically examined through a cultural lens. In high-trust-ambiguity contexts (like a new
technology), individuals in collectivist cultures may rely more heavily on in-group cues (the streamer) to
reduce uncertainty. This mechanism may be less potent in individualistic cultures where personal assessment
of the technology’s utility and transparency is paramount. Thus, the universal applicability of trust as a
mediator calibrated primarily through the human channel is a key hypothesis for future research, not a fact.
Finally, the paradox of personalization, its capacity to both engage and alienate, reveals that its implementation
is less a technical challenge and more a strategic exercise in consumer psychology (Ding et al., 2025; Xie et
al., 2022). This finding directly engages with the “personalization-privacy paradox,” a well-established
concept where users desire relevant content but resist the data collection required to enable it (Philp &
Nepomuceno, 2024). Our synthesis deepens this understanding by identifying a second, equally critical
tension: the clash between efficiency and exploration. While AI-driven personalization excels at optimizing for
known preferences (exploitation), it inherently risks stifling the serendipitous discovery (exploration) that is a
fundamental gratification of both livestreaming and travel planning. This creates a “filter bubble” that can
undermine the perceived authenticity and excitement of the experience (Xie et al., 2022; Zhou et al., 2025).
Therefore, the optimal “sweet spot” is not found by merely tuning an algorithm, but by designing for what can
be termed “value-centric transparency” (Ding et al., 2025; Philp & Nepomuceno, 2024; Xie et al., 2022)” This
means systems must be architected to not only be accurate but also to transparently communicate their benefit
to the viewer, for instance, by explicitly curating a “surprising discovery” alongside hyper-personalized
options, thereby justifying the use of their data. This moves the paradigm from covert data extraction to a
value-driven partnership, transforming personalization from a potential source of skepticism into a cornerstone
of a trustworthy and captivating digital hospitality experience.
Practical Implications
For tourism marketers and livestream platforms, this review offers several evidence-based recommendations:
1. Human-Centric AI Integration: Position AI as a decision-support tool for streamers, not a replacement.
Train streamers to effectively contextualize and endorse AI recommendations to leverage trust transfer.
2. Transparency by Design: Implement explainable AI features that briefly justify recommendations (for
example, "Based on your interest in beach destinations"). Balance hyper-personalized options with
curated "surprising discoveries" to maintain exploratory appeal.
3. Trust-Calibration Mechanisms: Use streamer credibility to bootstrap trust in AI. Consider visual cues
indicating streamer endorsement of AI suggestions.
4. Privacy-Preserving Personalization: Adopt granular privacy controls and value-justified data usage,
clearly communicating how personalization benefits viewers.
Page 2868
www.rsisinternational.org
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
5. Cross-Platform Adaptation: Recognize that the human-AI synergy model may require adjustment for
Western audiences with different cultural expectations regarding influencer authority and privacy.
Implications for Future Research
This synthesis not only consolidates current knowledge but also charts a course for future inquiry. A sensitivity
analysis confirmed that the exclusion of the sole weak-quality study did not alter the fundamental thematic
structure or conclusions, reinforcing the robustness of the synthesized findings. Based on the remaining
identified gaps and complexities, we propose the following research directions:
1. Cross-Cultural Validation: The overwhelming China-centric bias of the extant literature represents the
most urgent gap. Future research must test the proposed theoretical model in Western (For example,
using Amazon Live, TikTok Shop) and other cultural contexts. These studies should explicitly
investigate how cultural dimensions, such as individualism-collectivism, uncertainty avoidance, and
power distance, moderate the core relationships identified here, particularly the strength of the human-
AI synergy and the trust-transfer mechanism. This will determine whether the ‘hierarchical
collaboration’ model is universal or culturally specific.
2. From Intentions to Actual Behavior: Future research must bridge the intention-behavior gap by
prioritizing longitudinal studies and field experiments that track actual booking conversions and
spending, not just intentions. Partnerships with tourism platforms to analyze real user data are critical
to validate the commercial impact of AI recommendations.
3. Algorithmic Transparency and Serendipity: Experimental studies are needed to design and test AI
interfaces that operationalize value-centric transparency and quantify the optimal balance between
personalized recommendations and serendipitous discovery. For instance, future research should test
propositions such as: AI interfaces that explicitly justify recommendations with value statements (for
example, “We suggest this based on your preference for family-friendly amenities”) will generate
higher trust and engagement than opaque recommendations, while also maintaining acceptance when
introducing serendipitous options.
4. Ethical and Data Governance Frameworks: Qualitative and policy-oriented research is required to
develop industry-specific ethical guidelines and data governance models that address privacy concerns
without stifling the benefits of personalization.
CONCLUSION
This systematic review set out to synthesize how AI-powered real-time recommendations impact customer
engagement and booking decisions within the livestream tourism marketing landscape. The synthesis of 15
empirical studies reveals that the influence is not a simple direct effect but a complex, psychologically
mediated process. The findings robustly confirm the Stimulus-Organism-Response (S-O-R) model as a
foundational framework, demonstrating that AI recommendations act as powerful external stimuli that must
first navigate the user’s internal cognitive and affective states, such as perceived trust, emotional arousal, and
telepresence, before culminating in behavioral responses like heightened engagement and booking intentions.
A central tenet emerging from this review is the non-negotiable synergy between human streamers and AI
(Kim, 2023; Liu et al., 2024). The evidence decisively refutes the narrative of algorithmic dominance,
establishing instead a model of hierarchical collaboration where the human streamer serves as the trusted
curator and emotional anchor, and the AI functions as a powerful, real-time decision-support tool. Within this
dynamic, trust operates as the critical psychological gatekeeper, a dual-faceted mediator contingent upon both
the streamer’s authenticity and the AI’s perceived competence. Furthermore, the review uncovers the dual-
edged nature of personalization, where its power to engage is bounded by thresholds of perceived intrusiveness
and the suppression of serendipity, necessitating a shift towards value-centric transparency in design.
The primary theoretical contribution of this review lies in its unified framework that explains the psychological
mechanisms through which AI recommendations generate engagement and booking intentions in livestreams.
Page 2869
www.rsisinternational.org
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
For practitioners, the findings provide a strategic blueprint: success hinges on designing AI systems that
augment, rather than disrupt, the human streamer’s credibility and the viewer’s trust, while carefully balancing
personalized relevance with exploratory discovery. While this review consolidates a fragmented field, the
heavy geographical focus on a single market and the reliance on behavioral intentions in the extant literature
underscore the critical need for the future research directions outlined. Ultimately, this review establishes that
in the high-involvement context of tourism, the most effective AI is not the most autonomous, but the one that
most seamlessly strengthens the human connection and builds a trustworthy, captivating digital experience.
Ethical Considerations
Ethical Approval
This study is a systematic review of previously published literature. As such, it did not involve the direct
collection of new data from human or animal subjects, and therefore, ethical approval was not required.
Conflict of Interest
The authors declare no potential conflicts of interest with respect to the research, authorship, or publication of
this article.
Data Availability
The data supporting the findings of this systematic review are derived from the 15 primary research articles
included in the analysis, which are all cited in the reference list. The full dataset extracted and analyzed during
this review is available in the summary tables (Tables 1 and 2) within this manuscript.
REFERENCES
1. Antczak, B. (2025). The Impact of Artificial Intelligence on Tourism Industry: A Marketing
Perspective.Journal of Modern Science, 62(2), 450-472.
https://doi.org/https://doi.org/10.13166/jms/208128
2. Brignardello-Petersen, R., Santesso, N., & Guyatt, G. H. (2024). Systematic reviews of the literature: an
introduction to current methods. American Journal of Epidemiology, kwae232.
https://doi.org/https://doi.org/10.1093/aje/kwae232
3. Bulchand-Gidumal, J., William Secin, E., O’Connor, P., & Buhalis, D. (2024). Artificial intelligence’s
impact on hospitality and tourism marketing: exploring key themes and addressing challenges. Current
Issues in Tourism, 27(14), 2345-2362. https://doi.org/https://doi.org/10.1080/13683500.2023.2229480
4. Chen, C., & Wei, Z. (2024). Role of Artificial Intelligence in travel decision making and tourism product
selling.
Asia
Pacific
Journal
of
Tourism
Research,
29(3),
239-253.
https://doi.org/https://doi.org/10.1080/10941665.2024.2317390
5. Chen, J., Luo, J., & Zhou, T. (2024). Research on determinants affecting users’ impulsive purchase
intention in live streaming from the perspective of perceived live streamers’ ability. Behavioral Sciences,
14(3), 190. https://doi.org/doi: 10.3390/bs14030190
6. Cheng, Y., & Jiang, H. (2022). 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. https://doi.org/https://doi.org/10.1108/JPBM-05-2020-2907
7. Cohen, S. A., Prayag, G., & Moital, M. (2014). Consumer behaviour in tourism: Concepts, influences
and opportunities. Current Issues in Tourism, 17(10),
872-909.
https://doi.org/https://doi.org/10.1080/13683500.2013.850064
8. Ding, L., Antonucci, G., & Venditti, M. (2025). Unveiling user responses to AI-powered personalised
recommendations: a qualitative study of consumer engagement dynamics on Douyin. Qualitative Market
Research: An International Journal, 28(2), 234-255. https://doi.org/https://doi.org/10.1108/QMR-11-
2023-0151
9. Han, C., & Cumming, T. M. (2022). Teachers’ beliefs about the provision of education for students with
autism spectrum disorder: A systematic review. Review Journal of Autism and Developmental
Disorders, 11(3), 545-563. https://doi.org/https://doi.org/10.1007/s40489-022-00350-6
Page 2870
www.rsisinternational.org
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
10. Hochreiter, V., Benedetto, C., & Loesch, M. (2022). The stimulus-organism-response (SOR) paradigm
as a guiding principle in environmental psychology: comparison of its usage in consumer behavior and
organizational culture and leadership theory. Global Business Conference 2022 Proceedings,
11. Howie, C. (2019). Conducting your first systematic review. PsyPag Quarterly.
https://doi.org/https://doi.org/10.53841/bpspag.2019.1.113.32
12. Kim, M. (2023). Parasocial interactions in digital tourism: Attributes of live streamers and viewer
engagement dynamics in South Korea. Behavioral Sciences, 13(11), 953.
https://doi.org/https://doi.org/10.3390/bs13110953
13. Kmet, L. M., Cook, L. S., & Lee, R. C. (2004). Standard quality assessment criteria for evaluating
primary research papers from a variety of fields. https://doi.org/https://doi.org/10.7939/R37M04F16
14. Kotler, P. (2022). Marketing for hospitality and tourism. UK.: Pearson Education Limited.
15. Li, C., & Zheng, W. (2025). Nipping trouble in the bud: A proactive tourism recommender system.
Information & Management, 62(1), 104062. https://doi.org/https://doi.org/10.1016/j.im.2024.104062
16. Liu, J., Hsu, F. C., Ma, Y., & Huang, Y. (2024). Personal Brand Equity and Telepresence's Role in
Tourism Electronic Commerce Live Streaming. International Journal of Tourism Research, 26(5),
e2767. https://doi.org/https://doi.org/10.1002/jtr.2767
17. Liu, X., Zhang, L., & Chen, Q. (2022). The effects of tourism e-commerce live streaming features on
consumer purchase intention: The mediating roles of flow experience and trust. Frontiers in psychology,
13, 995129. https://doi.org/https://doi.org/10.3389/fpsyg.2022.995129
18. Lo, C., Yu, H., Yin, X., Shetty, K., He, C., Hu, K., Platz, J. M., Ilardi, A., & Madhvanath, S. (2021).
Page-level optimization of e-commerce item recommendations. Proceedings of the 15th ACM
Conference on Recommender Systems,
19. López Naranjo, A. L., Puente Riofrío, M. I., Carrasco Salazar, V. A., Erazo Rodríguez, J. D., & Buñay-
Guisñan, P. A. (2025). Artificial intelligence in the tourism business: a systematic review. Frontiers in
Artificial Intelligence, 8, 1599391. https://doi.org/https://doi.org/10.3389/frai.2025.1599391
20. McGarty, A. M., & Melville, C. A. (2018). Parental perceptions of facilitators and barriers to physical
activity for children with intellectual disabilities: A mixed methods systematic review. Research in
developmental disabilities, 73, 40-57. https://doi.org/https://doi.org/10.1016/j.ridd.2017.12.007
21. Mei, L., Tang, N., Zeng, Z., & Shi, W. (2025). Artificial Intelligence Technology in Live Streaming E-
commerce: Analysis of Driving Factors of Consumer Purchase Decisions. International Journal of
Computers Communications & Control, 20(1).
22. https://doi.org/https://doi.org/10.15837/ijccc.2025.1.6871
23. Moghddam, H. A., Mortimer, G., Ahmadi, H., & Sharif-Nia, H. (2025). How livestream engagement
inspires tourist purchasing behaviour: A multi-study approach. International Journal of Information
Management, 83, 102903. https://doi.org/https://doi.org/10.1016/j.ijinfomgt.2025.102903
24. Morgan, R. M., & Hunt, S. D. (1994). The commitment-trust theory of relationship marketing. Journal
of marketing, 58(3), 20-38. https://doi.org/https://doi.org/10.1177/002224299405800
25. Munn, Z., Barker, T. H., Moola, S., Tufanaru, C., Stern, C., McArthur, A., Stephenson, M., &
Aromataris, E. (2020). Methodological quality of case series studies: an introduction to the JBI critical
appraisal tool. JBI evidence synthesis, 18(10), 2127-2133. https://doi.org/10.11124/JBISRIR-D-19-
00099
26. Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L.,
Tetzlaff, J. M., Akl, E. A., & Brennan, S. E. (2021). The PRISMA 2020 statement: an updated guideline
for reporting systematic reviews. bmj, 372. https://doi.org/https://doi.org/10.1136/bmj.n71
27. Philp, M., & Nepomuceno, M. V. (2024). How reviews influence product usage post-purchase: An
examination of video game playtime. Journal of Business Research, 172, 114456.
https://doi.org/https://doi.org/10.1016/j.jbusres.2023.114456
28. Qu, C., Dai, S., Guo, K., Zhao, L., Niu, Y., Zhang, X., & Xu, J. (2025). KuaiLive: A Real-time
Interactive Dataset for Live Streaming Recommendation. arXiv preprint arXiv:2508.05633.
https://doi.org/https://doi.org/10.48550/arXiv.2508.05633
29. Seabra, C., Reis, P., & Abrantes, J. L. (2020). The influence of terrorism in tourism arrivals: A
longitudinal approach in a Mediterranean country. Annals of tourism research, 80, 102811.
https://doi.org/doi:10.1016/j.annals.2019.102811
30. Sichach, M. (2023). Uses and gratifications theory-background, history and limitations. History and
Limitations (November 17, 2023). https://doi.org/http://dx.doi.org/10.2139/ssrn.4729248
Page 2871
www.rsisinternational.org
INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)
ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XIV October 2025 | Special Issue on Management
31. Tolloso, M., Bacciu, D., Mokarizadeh, S., & Varesi, M. (2025). Real-time and personalized product
recommendations for large e-commerce platforms. International Conference on Artificial Neural
Networks,
32. Tussyadiah, I. (2020). A review of research into automation in tourism: Launching the Annals of
Tourism Research Curated Collection on Artificial Intelligence and Robotics in Tourism. Annals of
tourism research, 81, 102883. https://doi.org/https://doi.org/10.1016/j.annals.2020.102883
33. Wu, Y., & Yusof, Y. (2024). Emerging Trends in Real-time Recommendation Systems: A Deep Dive into
Multi-behavior Streaming Processing and Recommendation for E-commerce Platforms. J. Internet Serv.
Inf. Secur, 14(4), 45-66. https://doi.org/https://orcid.org/0000-0002-2720-2441
34. Xie, C., Yu, J., Huang, S. S., & Zhang, J. (2022). Tourism e-commerce live streaming: Identifying and
testing a value-based marketing framework from the live streamer perspective. Tourism management,
91, 104513. https://doi.org/https://doi.org/10.1016/j.tourman.2022.104513
35. Zhang, L., & Wu, X. (2025). Exploring the underlying mechanisms of customers’ intention to adopt
product recommendations from live streamers: A moderated mediation approach. PloS one, 20(2),
e0314682. https://doi.org/https://doi.org/10.1371/journal.pone.0314682
36. Zhang, Z., Qin, M., Li, Z., & Zhu, H. (2025). The impact of travel live streaming on tourists’ travel
intention on the spot. PloS one, 20(10), e0333811. https://doi.org/doi:10.1371/journal.pone.0333811
37. Zheng, S., Wu, M., & Liao, J. (2022). The impact of destination live streaming on viewers’ travel
intention. Current Issues in Tourism, 26(2), 184-198.
38. https://doi.org/https://doi.org/10.1080/13683500.2022.2117594
39. Zhou, Y., Liu, S. S., Xu, X., & Cheng, T. (2025). Enhancing competitive advantage through AI-driven
live streaming sales. IEEE Transactions on Engineering Management.
40. https://doi.org/10.1109/TEM.2025.3550400
41. Zhu, Y., Zhang, R., Zou, Y., & Jin, D. (2023). Investigating customers’ responses to artificial intelligence
chatbots in online travel agencies: The moderating role of product familiarity. Journal of Hospitality and
Tourism Technology, 14(2), 208-224. https://doi.org/https://doi.org/10.1108/JHTT-02-2022-0041
