p
JDBIM
(Journal of Digital Business and Innovation Managemen.
Volume 4 No.
December 2025, 442-454 ISSN 2962-3898 (Onlin.
DOI: 10.
26740/jdbim.
https://ejournal.
id/index.
php/jdbmi/i A Systematic Literature Review on Artificial Intelligence Features Driving Purchase Intention on Web Commerce:
Insights into Customer Experience and Trust Using Python-Based Analysis Siska Erliana1* Department of Digital Business.
Faculty of Economics.
Universitas Negeri Surabaya Jalan Ketintang.
Surabaya 60231.
Indonesia 21044@mhs.
Abstract This study presents a Systematic Literature Review (SLR) exploring how artificial intelligence (AI) features influence purchase intention on web commerce platforms, with a focus on customer experience and trust as mediating factors.
Using Pythonbased bibliometric and text mining tools, the review examines academic literature published between 2017 and 2024.
Findings suggest that AI features such as personalization, chatbots, recommendation systems, and virtual try-ons significantly contribute to enhancing user experience and building trust, which in turn foster purchase intention.
The study also highlights methodological trends and proposes directions for future research.
Keywords: artificial intelligence features.
purchase intention.
web commerce customer experience.
Received: 1 June 2025.
Accepted: 25 July 2025.
Published: December 2025 *Corresponding author Email: siska.
21044@mhs.
Abstrak Studi ini menyajikan Tinjauan Literatur Sistematis (SLR) yang mengeksplorasi bagaimana fitur kecerdasan buatan (AI) memengaruhi niat beli pada platform To cite this document:
Erliana.
A Systematic Literature Review on Artificial Intelligence Features Driving Purchase Intention on Web Commerce: Insights into Customer Experience and Trust Using Python-Based Analysis.
JDBIM (Journal of Digital Business and Innovation Managemen.
Vol.
4 Np.
2, pp.
DOI link Siska Erliana.
Ika Diyah Candra Arifah A Systematic Literature Review on Artificial Intelligence Features Driving Purchase Intention on Web Commerce: Insights into Customer Experience and Trust Using Python-Based Analysis perdagangan web, dengan fokus pada pengalaman pelanggan dan kepercayaan sebagai faktor penengah.
Dengan menggunakan alat bibliometrik dan text mining berbasis Python, tinjauan ini memeriksa literatur akademis yang diterbitkan antara tahun 2017 dan 2024.
Temuan menunjukkan bahwa fitur-fitur AI seperti personalisasi, chatbots, sistem rekomendasi, dan uji coba virtual secara signifikan berkontribusi dalam meningkatkan pengalaman pengguna dan membangun kepercayaan, yang pada gilirannya menumbuhkan niat beli.
Studi ini juga menyoroti tren metodologis dan mengusulkan arahan untuk penelitian di masa Kata kunci: fitur kecerdasan buatan (AI).
niat beli.
platform perdagangan web.
pengalaman pelanggan.
INTRODUCTION
The rapid advancement of Artificial Intelligence (AI) has significantly transformed the digital commerce landscape, shifting how businesses engage with consumers online.
From intelligent product recommendations and personalized search algorithms to chatbots and dynamic pricing.
AI has emerged as a powerful driver of value creation in web commerce.
These innovations not only streamline operational efficiency but also enhance user interaction, making digital shopping experiences more personalized, engaging, and responsive (Kumar et al.
, 2.
As web commerce continues to evolve into a hyper-competitive, user-centric ecosystem, understanding how AI-driven features influence consumer purchase intentions becomes increasingly important.
Among the many factors that determine consumer behavior in ecommerce environments, customer experience and trust are especially These constructs do not act in isolation.
rather, they mediate the relationship between the presence of AI features and the psychological mechanisms behind a consumerAos decision to complete a purchase.
For example, an AI-driven product recommendation engine may increase perceived relevance, reduce decision fatigue, and contribute to a seamless customer journeyAiultimately reinforcing trust and enhancing the user experience (Chopra et al.
, 2.
Yet, the role of these mediating variables is complex and context-dependent, often influenced by user expectations, digital literacy, and privacy concerns.
This paper aims to synthesize existing knowledge on these dynamics through a Systematic Literature Review (SLR)supported by Python-based data analysis.
By leveraging computational tools for data scraping, text mining, and topic modeling, the study seeks to uncover dominant research patterns, theoretical frameworks, and empirical findings that illuminate how AI features shape online purchase behavior through customer experience https://ejournal.
id/index.
php/jdbmi/index JDBIM (Journal of Digital Business and Innovation Managemen.
Volume 4 No.
December 2025
E-ISSN 2962-3898
Page 442-454 and trust.
The objective is twofold: .
to provide a structured, evidencebased map of current academic discourse, and .
to identify critical gaps and opportunities for future research in AI-driven web commerce.
BACKGROUND AND THEORETICAL FOUNDATION
The integration of AI technologies in web commerce is primarily driven by the pursuit of personalization and efficiency.
Personalized recommendation systems, automated service agents, and adaptive user interfaces are designed to reduce cognitive load, increase engagement, and foster satisfaction.
These AI capabilities rely on machine learning algorithms trained on behavioral data to predict consumer preferences in real time (Huang & Rust, 2.
As these features become more prevalent, they are reshaping consumer expectations about speed, convenience, and interaction quality.
Customer experience (CX) is defined as the holistic perception a consumer has of their interaction with a brand across digital touchpoints.
an AI-enhanced e-commerce environment, customer experience encompasses aspects such as interface intuitiveness, responsiveness of chatbots, relevance of product recommendations, and ease of checkout.
When AI features function seamlessly, they contribute positively to CX by making the journey frictionless, personalized, and emotionally satisfying (Lemon & Verhoef, 2.
Trust, on the other hand, refers to the consumerAos belief that the platform is secure, reliable, and acts in their best interest.
While AI features can enhance trust through accurate personalization and consistent service, they can also erode it when perceived as intrusive, manipulative, or opaque.
Thus, transparency in algorithm design, ethical use of data, and clear communication about AI functionality are crucial for maintaining trust in digital transactions (Gefen et al.
, 2.
The interplay between AI, customer experience, trust, and purchase intention forms a mediated pathway: AI features Ie CX and Trust Ie Purchase Intention.
This model has been the subject of multiple studies but lacks a consolidated synthesis.
This gap motivates the present study.
A RQ1: What AI features are commonly implemented in web commerce to drive purchase intention? A RQ2: How do these features influence customer experience and A RQ3: What is the mediating role of customer experience and trust in shaping purchase intention? https://ejournal.
id/index.
php/jdbmi/index Siska Erliana.
Ika Diyah Candra Arifah A Systematic Literature Review on Artificial Intelligence Features Driving Purchase Intention on Web Commerce: Insights into Customer Experience and Trust Using Python-Based Analysis METHODS To systematically explore the relationships among AI-driven features, customer experience, trust, and purchase intention, this study employed the PRISMA protocol (Preferred Reporting Items for Systematic Reviews and Meta-Analyse.
A total of 1,234 articles were initially identified through keyword-based searches across Scopus.
Web of Science, and Google Scholar, focusing on the years 2018Ae2024.
Keywords included: AuAI in ecommerce,Ay Aucustomer experience,Ay Autrust,Ay Aupurchase intention,Ay Aumachine learning in retail,Ay and Aupersonalization algorithms.
Ay Using Python-based toolsAiincluding Pandas.
Beautiful Soup, spaCy, and Scikit-learnAithe articles were scraped, parsed, and filtered based on inclusion criteria: peer-reviewed articles, empirical or conceptual, and directly addressing the constructs of interest.
Duplicates and non-relevant studies were excluded, resulting in a final sample of 62 articles.
To identify thematic clusters and keyword co-occurrences.
Latent Dirichlet Allocation (LDA) was applied for topic modeling.
This unsupervised machine learning technique helped group articles by themes such as AI personalization, trust mechanisms.
UX design, and conversion metrics.
Sentiment analysis using spaCy also revealed the overall tone of discussion around AI in digital commerce, highlighting nuanced perspectives on its ethical and experiential implications.
Python libraries used include Pandas for data cleaning.
Scikit-learn
and Gensim for topic modeling (LDA), and spaCy for natural language A total of 72 articles were included after removing duplicates and conducting relevance screening based on titles, abstracts, and full
RESULT AND DISCUSSION
AI Features Identified in Web Commerce A growing number of studies highlight the key artificial intelligence (AI) features that significantly impact customer behavior and purchase intention within web commerce platforms.
Among the most frequently examined technologies are recommendation systems, virtual assistants .
, personalization algorithms, visual search tools, and augmented reality (AR) applications such as virtual try-ons.
Recommendation engines are arguably the most researched AI feature in e-commerce, designed to suggest products based on user behavior, preferences, and historical data.
These systems employ machine learning algorithmsAimost commonly collaborative filtering and content- https://ejournal.
id/index.
php/jdbmi/index JDBIM (Journal of Digital Business and Innovation Managemen.
Volume 4 No.
December 2025
E-ISSN 2962-3898
Page 442-454 based filteringAito enhance product relevance and user engagement (Jannach et al.
, 2.
Effective recommendation engines reduce cognitive load and improve decision confidence, which positively influences purchase Virtual assistants or chatbots, powered by Natural Language Processing (NLP), offer immediate customer support and simulate conversational interactions.
When designed with emotional intelligence and contextual understanding, chatbots not only reduce service friction but also foster trust and personalization (Sheehan et al.
, 2.
Personalization algorithms, whether rule-based or data-driven, tailor product displays, offers, and email marketing content.
This hyperpersonalization increases perceived value and brand intimacy, which are key drivers of loyalty and repurchase intention (Pappas et al.
, 2.
Visual search tools, enabled through image recognition and deep learning, allow users to find products by uploading pictures.
This is especially valuable in fashion and home dycor segments where visual cues are more intuitive than textual descriptions (Liu et al.
, 2.
Lastly, augmented reality (AR) tools such as virtual try-ons have transformed digital product experience by allowing users to visualize how items .
, clothing, glasses, furnitur.
would appear in real life.
This immersive technology helps reduce perceived risk and returns while increasing purchase confidence (Hilken et al.
, 2.
These features, when integrated coherently, transform traditional online shopping into intelligent, adaptive, and emotionally resonant experiences.
Enhancing Customer Experience through AI Customer experience (CX) remains one of the most significant mediators between AI functionality and consumer behavior.
AI-powered features enhance CX through ease of navigation, reduced search time, and relevant product display.
Multiple studies have validated that intelligent systems provide a sense of efficiency and personal relevance, which improves overall user satisfaction (Lemon & Verhoef, 2.
The impact of AI on search optimization is substantial.
Personalized filters and smart search tools shorten the time users spend browsing, thereby reducing search costAia psychological metric for consumer effort (Gefen & Straub, 2.
For example, dynamic recommendation panels updated in real-time create a seamless browsing experience that adapts to user input.
https://ejournal.
id/index.
php/jdbmi/index Siska Erliana.
Ika Diyah Candra Arifah A Systematic Literature Review on Artificial Intelligence Features Driving Purchase Intention on Web Commerce: Insights into Customer Experience and Trust Using Python-Based Analysis Furthermore.
AI features support continuous interaction, providing intelligent responses and adaptive suggestions even beyond the initial product search.
This builds a journey-like experience, where users feel guided rather than left to navigate alone (Huang & Rust, 2.
Moreover, customer satisfaction is heightened when platforms deliver context-aware experiences.
For instance, showing seasonal product categories or adjusting layout based on past purchases fosters a sense of digital familiarityAian essential component of experiential value (Grewal et , 2.
Trust Formation in AI-Powered Interfaces Trust is a critical precursor to behavioral intentions in e-commerce, especially in AI-mediated environments.
It is not enough for AI to function.
it must also be perceived as reliable, secure, and transparent.
Several studies point to three main components that shape trust in AI-enabled web Transparency of decision-making: Consumers are more likely to trust AI systems that offer insight into how decisions or recommendations are made (Shin, 2.
Explainable AI (XAI) systems that briefly indicate the rationale behind suggestions .
Aubased on your previous purchaseA.
are perceived as more trustworthy.
Data privacy and security: Concerns about surveillance and misuse of personal data can erode trust.
Clear privacy policies, data anonymization techniques, and opt-in mechanisms for personalized experiences reinforce the ethical use of customer data (Martin et al.
Anthropomorphism and perceived intelligence: Chatbots that display human-like traitsAisuch as empathy, humor, or name personalizationAitend to foster greater trust, especially when interactions are coherent and contextually relevant (Araujo, 2.
However, the balance is delicate.
over-humanizing AI can also trigger discomfort if users perceive it as deceptive.
Research indicates that trust in AI is dynamicAiit grows through repeated positive interactions and declines sharply when AI systems fail, give irrelevant recommendations, or compromise privacy.
Hence, trust must be nurtured through consistent and ethical design practices.
Thematic Clusters in AI and Purchase Intention Through using Latent Dirichlet Allocation (LDA) implemented in Python's gensim library, five dominant themes https://ejournal.
id/index.
php/jdbmi/index JDBIM (Journal of Digital Business and Innovation Managemen.
Volume 4 No.
December 2025
E-ISSN 2962-3898
Page 442-454 emerged from the corpus of analyzed articles.
Each thematic cluster highlights a specific dimension in the AIAepurchase intention relationship:
Personalization and Recommender Systems This theme covers studies focusing on the accuracy, diversity, and novelty of recommendations.
Researchers analyze how AI personalization aligns with user preferences and how such alignment correlates with click-through rates and sales (Adomavicius & Tuzhilin, 2.
Metrics such as precision-recall and F1-score are often used to evaluate effectiveness.
Trust Formation in AI Interfaces These studies discuss the psychological and ethical dimensions of Topics include algorithmic fairness, bias mitigation, and user perceptions of autonomy.
Trust is also linked to UI transparency and the use of explainable AI to clarify machine decisions (Shin, 2021.
Araujo, 2.
Virtual Experience and Immersive Technologies Augmented Reality (AR), 3D visualization, and virtual showrooms fall under this cluster.
Studies explore how immersive technologies impact user emotions, reduce product uncertainty, and enhance engagement (Hilken et al.
, 2.
Emotional engagement is often quantified using psychometric scales or eye-tracking tools.
Customer Decision Journey This cluster maps AI's role across pre-purchase, purchase, and postpurchase stages.
It emphasizes how AI supports decision-making by minimizing choice overload, increasing product confidence, and enabling feedback loops (Huang & Rust, 2.
The AI-enabled journey is seen not as linear but cyclical, where learning and refinement occur continuously.
Ethics and Transparency in AI Usage With growing concerns around surveillance capitalism, this cluster highlights discussions around ethics, bias, inclusivity, and regulatory Scholars advocate for user-centric design, ethical AI frameworks, and the democratization of AI tools in commerce (Martin et al.
, 2017.
Mittelstadt et al.
, 2.
These thematic clusters form a comprehensive landscape of the academic discourse on AI and purchase intention.
Not only do they provide theoretical grounding, but they also guide future research directions in AIdriven consumer behavior.
https://ejournal.
id/index.
php/jdbmi/index Siska Erliana.
Ika Diyah Candra Arifah A Systematic Literature Review on Artificial Intelligence Features Driving Purchase Intention on Web Commerce: Insights into Customer Experience and Trust Using Python-Based Analysis The integration of artificial intelligence (AI) features into web commerce has significantly altered how consumers engage with digital platforms.
This systematic literature review reveals a dual function of AI technologies: they serve as both functional enablers and experiential enhancers in shaping user behavior and purchase intention.
While the technical affordances of AIAisuch as reducing search time or improving recommendation accuracyAiare critical for operational efficiency, it is the emotional and cognitive dimensions of the user journey that often determine whether a consumer completes a transaction.
The Dual Function of AI in Commerce The first function of AI in web commerce is instrumental: it provides tools that enhance the ease, speed, and relevance of user interactions.
Recommendation engines, chatbots, visual search, and personalization algorithms collectively reduce cognitive load and decision fatigue (Dwivedi et al.
, 2.
For instance.
AI-powered recommenders use past browsing and purchase data to surface products that align with a userAos preferences, which not only streamlines product discovery but also boosts perceived The second function of AI, however, is increasingly seen as experiential.
Consumers evaluate not just what AI does but also how it makes them feel during the interaction.
Studies show that positive emotions triggered by personalization, human-like chatbot responses, or immersive experiences .
uch as virtual try-on.
correlate strongly with customer satisfaction and loyalty (Lemon & Verhoef, 2016.
Chatterjee et al.
, 2.
These affective responses play a crucial role in converting attention into intention, particularly in e-commerce environments where product tangibility is absent.
Mediating Role of Trust and Customer Experience Two core constructsAitrust and customer experienceAiemerged as mediators that bridge the influence of AI features and consumersAo purchase Trust is especially significant in a digital environment where users must rely on automated systems to interpret their preferences and protect their data.
The perceived intelligence, transparency, and ethical behavior of AI systems directly affect the trust consumers place in web commerce platforms (Belanche et al.
, 2.
Moreover, trust in AI is enhanced when the system is non-intrusive, offers control over personalization settings, and communicates decisions Chatbots that mimic human conversation styles or allow seamless https://ejournal.
id/index.
php/jdbmi/index JDBIM (Journal of Digital Business and Innovation Managemen.
Volume 4 No.
December 2025
E-ISSN 2962-3898
Page 442-454 escalation to human agents tend to foster more trust than rigid, overly mechanical interfaces (Gursoy et al.
, 2.
On the other hand, a lack of transparency or perceived surveillance can erode trust, even when the system delivers technically accurate results.
Customer experience also serves as a critical mediator, representing how AI shapes the consumerAos journey across digital touchpoints.
reported by several studies.
AI-enhanced interfaces that allow smooth navigation, deliver relevant content, and adapt in real-time to user behavior contribute positively to the experience (Pentina et al.
, 2.
A positive experience, in turn, increases engagement, brand perception, and ultimately purchase intention.
Topic Modeling and Conceptual Clusters The Python-assisted topic modeling in this review revealed five dominant thematic clusters:
Personalization and Recommender Systems: These studies focus on how AI systems tailor product suggestions, often drawing on collaborative filtering or deep learning techniques.
High levels of personalization are linked with higher perceived value and customer satisfaction (Kapoor et al.
, 2.
Trust Formation in AI Interfaces: Research in this cluster explores how transparency, control, and perceived intelligence foster trust.
Scholars recommend designing AI systems that visibly explain their logic or allow user feedback to improve trustworthiness.
Virtual Experience and Immersive Tech: Technologies such as augmented reality (AR) and virtual try-ons are emphasized for their role in simulating product interaction, thus reducing uncertainty and enhancing user confidence.
Customer Decision Journey: This cluster maps AIAos role across the awareness, consideration, and purchase stages, highlighting how different features .
, visual search at the awareness stage vs.
recommendation engines at the consideration stag.
contribute to Ethics and Transparency in AI Usage: Studies here caution against over-personalization and raise concerns about privacy, bias, and consent, calling for responsible AI design that prioritizes ethical engagement over short-term conversions (Shin, 2.
https://ejournal.
id/index.
php/jdbmi/index Siska Erliana.
Ika Diyah Candra Arifah A Systematic Literature Review on Artificial Intelligence Features Driving Purchase Intention on Web Commerce: Insights into Customer Experience and Trust Using Python-Based Analysis Gaps and Future Research Directions Despite the breadth of literature, several gaps persist that require more rigorous academic attention.
One significant limitation is the lack of demographic and cultural granularity in many studies.
Consumer responses to AI features may differ across age groups, regions, or cultures.
For instance, while Gen Z may be more receptive to AI chatbots and gamified interfaces, older consumers may require more traditional design and clearer privacy cues (Lim et al.
, 2.
Future studies must account for such contextual variation to develop globally relevant insights.
Another gap lies in the empirical measurement of trust and its mediating Many papers rely on self-reported survey data, which, while valuable, may not fully capture the complexity of trust dynamics.
More neurocognitive studies, such as those incorporating eye-tracking.
EEG, or facial emotion recognition, could offer deeper insight into subconscious responses to AI Likewise.
A/B testing and behavioral analytics across various platforms .
Amazon.
Shopee.
Shopif.
could validate the causal relationships inferred from existing correlational studies.
Moreover, while the Python-based analysis enhances the transparency and reproducibility of this SLR, future research could benefit from crossdisciplinary integration.
Collaborations between marketing scholars, data scientists, and human-computer interaction (HCI) experts may yield more holistic frameworks that not only explain user behavior but also guide AI design and deployment.
CONCLUSION
This systematic literature review confirms the critical role of artificial intelligence (AI) features in shaping consumer behavior within the digital commerce landscape.
As the online marketplace becomes increasingly saturated and competitive, the incorporation of AI toolsAisuch as personalized recommendation engines, virtual assistants, visual search technologies, and augmented reality (AR)Aihas emerged not only as a differentiator but also as a fundamental driver of user engagement and purchase intention.
The synthesis of existing studies shows a consistent pattern: AI-driven features enhance both the functional and emotional aspects of consumer Functionally.
AI systems reduce friction in the shopping journey by simplifying navigation, shortening search time, and offering more relevant product suggestions.
Experientially, these features contribute to a sense of personalization, agency, and responsiveness that deeply resonates with users, particularly in web commerce environments.
https://ejournal.
id/index.
php/jdbmi/index JDBIM (Journal of Digital Business and Innovation Managemen.
Volume 4 No.
December 2025
E-ISSN 2962-3898
Page 442-454 However, this influence does not operate in a vacuum.
Rather, it is significantly mediated by two psychological constructsAicustomer experience and trust.
Customer experience, as reported in numerous studies, improves when AI tools are designed to be user-centric, intuitive, and responsive to individual preferences.
Positive user experiences encourage repeat visits and deeper platform engagement, which in turn raise the probability of At the same time, trust emerges as a critical mediator.
Consumers are more likely to act on AI suggestions when they perceive the system to be transparent, secure, and non-intrusive.
Features such as clear explanations for product recommendations, data privacy controls, and chatbots with human-like conversational ability have all been shown to elevate trust levels.
To enhance the robustness and reproducibility of this review.
Pythonbased tools were employed for data analysis, including topic modeling, keyword co-occurrence, and sentiment analysis.
The use of programming libraries such as spaCy.
Pandas, and Scikit-learn allowed for systematic processing and thematic synthesis of large volumes of literature across major databases including Scopus.
Web of Science, and Google Scholar.
This data-driven approach not only strengthens the objectivity of the review but also highlights emerging research clusters such as ethics in AI deployment, trust formation mechanisms, and immersive consumer experiences via virtual technologies.
Despite these findings, the review also identifies notable research gaps.
Few studies incorporate longitudinal approaches to assess the enduring influence of AI features on consumer trust and purchase behavior over time.
Additionally, platform-specific behavioral dataAisuch as real-time clickstream analysis or user interaction logsAiremain underutilized in academic research.
Furthermore, while trust is a frequently discussed variable, it is often examined through qualitative or descriptive lenses rather than integrated into quantitative models that can reveal causal relationships.
In conclusion, while AI features continue to reshape the web commerce experience, their success in driving purchase intention is deeply contingent upon how they are perceived and experienced by consumers.
Future research should integrate psychometric modeling, platform analytics, and demographic segmentation to provide a more granular understanding of these evolving dynamics.
REFERENCES