Advanced AI Chatbots for Enhanced Customer Interaction: A Comprehensive Review of Emerging Trends and Future Directions

Advanced AI Chatbots for Enhanced Customer Interaction: A Comprehensive Review of Emerging Trends and Future Directions

Nurliyana Abas., Norafiza Mohd Hardi^*, Norfaizah Md. Nasir

Faculty of Business and Management, Universiti Teknologi MARA Cawangan Kedah, 08400 Merbok, Kedah, Malaysia

DOI: https://dx.doi.org/10.47772/IJRISS.2025.909000785

Received: 26 September 2025; Accepted: 03 October 2025; Published: 30 October 2025

ABSTRACT

The increasing adoption of artificial intelligence (AI) in customer service has positioned advanced AI chatbots as pivotal tools for enhancing interaction, engagement, and satisfaction. However, despite their rapid deployment across industries such as e-commerce, banking, hospitality, and education, challenges related to technological limitations, user experience, data privacy, and ethical accountability continue to constrain their full potential. Addressing these gaps requires a comprehensive understanding of the scholarly landscape that links technological innovation with strategic implementation and human-centric design.The aim of this study is to review and synthesize the existing body of literature on advanced AI chatbots in customer interaction, with specific objectives to: (i) examine the breadth of scholarly contributions, (ii) construct a concept map to visualize the intellectual structure of the field, (iii) highlight contributions from key topic experts, and (iv) identify emerging themes that define future research directions. Methodologically, the study employed a structured review using Scopus AI, accessed on 25 September 2025. Boolean search strings captured publications at the intersection of AI chatbots, customer interaction, and automation. The findings were synthesized using analytical layers: summaries, concept maps, topic expert analysis and emerging theme identification. The results reveal consistent emphasis on personalization, efficiency, and customer engagement, alongside rising themes such as advanced natural language processing, empathetic AI, and chatbot adoption in educational contexts. Novel themes, particularly data privacy and ethical safeguards, indicate the field’s shift towards responsible and trust-oriented chatbot design. Theoretically, this review contributes by mapping the intellectual and thematic evolution of chatbot research, while practically, it offers insights into strategic solutions for businesses seeking to optimize chatbot adoption. The study concludes by highlighting limitations of existing research and outlining directions for future empirical inquiry into empathy, trust, and ethical AI in customer interaction.

Keywords: AI chatbots, customer interaction, natural language processing, user experience, data privacy

INTRODUCTION

Artificial intelligence (AI) technologies have become central to redefining the dynamics of customer service and engagement across industries. Among these innovations, conversational agents, commonly known as chatbots, have emerged as transformative tools capable of delivering personalized assistance, 24/7 availability, and seamless multi-platform integration. By leveraging advanced techniques such as machine learning (ML) and natural language processing (NLP), these systems facilitate human-like interactions, reduce service costs, and improve organizational efficiency [1], [2].

In this study, the term “advanced AI chatbots” refers to conversational systems that extend beyond rule-based designs by incorporating machine learning, transformer-based models such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-trained Transformers (GPT), and multimodal capabilities. These chatbots are distinguished by their ability to personalize interactions, adapt to user contexts, and perform complex tasks with greater autonomy and accuracy [3], [4]. Their deployment spans diverse sectors including hospitality [5], banking [6], and e-commerce [7].

Despite these advancements, persistent gaps remain between the promises of chatbot technologies and actual user experiences. Studies highlight challenges such as limited empathy, inconsistent trust, and unresolved ethical concerns related to data privacy and algorithmic transparency [8], [9]. These tensions indicate that adoption depends not only on technological performance but also on human-centric factors such as transparency, accountability, and emotional resonance.

Given the fragmented and sector-specific nature of prior research, a comprehensive synthesis is required to consolidate insights, identify contradictions, and situate findings within established theoretical models. This study therefore adopts a structured narrative review design, drawing on Scopus AI as a supportive analytical tool for mapping research trends, conceptual linkages, and emerging themes. Unlike prior reviews that remain primarily descriptive, this paper advances a critical synthesis anchored in the Technology Acceptance Model (TAM), which emphasizes perceived usefulness, ease of use, and trust [10], and Service-Dominant Logic (SDL), which frames technology as a platform for value co-creation [11].

The contribution of this study is threefold. First, it provides a consolidated and critical understanding of the intellectual structure of chatbot research. Second, it identifies unresolved tensions—efficiency versus empathy, innovation versus ethics—that define the trajectory of future inquiry. Third, it offers both theoretical and practical implications by linking chatbot adoption to established models of technology acceptance and service co-creation, while outlining strategies for organizations to enhance trust, empathy, and responsible deployment.

Accordingly, Scopus AI was employed as a supportive tool to generate preliminary outputs—including summary, expanded summary, concept map, topic experts, and emerging themes—that guided the identification of thematic clusters and intellectual linkages. However, the interpretation, coding, and synthesis were conducted manually by the authors to ensure conceptual depth and academic rigor.

The remainder of this paper is structured as follows. Section II outlines the methodology and explains how Scopus AI was used as an analytical aid rather than the methodology itself. Section III presents the discussion, organized into four components: (i) an overview of publication trends and sectoral applications (Summary and Expanded Summary), (ii) thematic clusters identified through the concept map, including implementation challenges, user experience, and technological innovations, (iii) influential scholars and intellectual contributions (Topic Experts), and (iv) key emerging themes and future directions. Section IV concludes by summarizing the theoretical contributions, practical implications, limitations, and opportunities for future research.

METHODOLOGY

This study employed a structured narrative review design, which combines the interpretive depth of narrative synthesis with the systematic rigor of structured procedures [12].  Such an approach is well-suited for emerging domains like advanced AI chatbots in customer interaction, where research is fragmented across technical, managerial, and user-experience perspectives.

Data Source and Search Strategy

The Scopus database was selected as the primary source, given its comprehensive coverage of peer-reviewed journals and conference proceedings in computer science, business, and service research [1], [2].  The search was conducted on 25 September 2025 using the following Boolean string:

(“chatbot” OR “virtual assistant” OR “conversational agent” OR “dialogue system”)

AND (“customer interaction” OR “customer service” OR “user engagement” OR “client communication”)

AND (“artificial intelligence” OR “AI” OR “machine learning” OR “natural language processing”)

AND (“automation” OR “support” OR “feedback” OR “experience”)

This string was designed to capture a wide spectrum of scholarly work at the intersection of AI-driven conversational agents and customer interaction [13], [3], [4].

Inclusion and Exclusion Approach

Rather than adopting a systematic screening process, this review applied a conceptual inclusion approach. Publications were included if they examined AI-powered chatbots (including generative AI, NLP, or ML-driven systems), focused on customer interaction as a central theme, and were peer-reviewed outputs indexed in Scopus. Studies were excluded if they addressed only the technical design of chatbots without customer interaction implications, or if they were non-peer-reviewed sources such as blogs, reports, or opinion essays. This ensured that the review remained focused on the intersection of technology and customer engagement rather than purely technical innovation.

Analytical Procedure

To structure the synthesis, this study employed Scopus AI’s analytical features as supportive tools, without treating them as the methodology itself. The following outputs were used to guide interpretation: the Summary, which provided an overview of publication trends; the Expanded Summary, which clustered methodological and sector-specific applications such as hospitality, banking, and e-commerce [5]–[7]; the Concept Map, which visualized keyword co-occurrences and thematic linkages such as personalization, multimodality, and ethics [9]; the Topic Experts, which identified influential scholars advancing key debates [14], [15]; and the Emerging Themes, which pointed to novel directions such as empathetic AI, multimodal systems, and privacy-preserving frameworks [8], [16].

These outputs were not accepted at face value; rather, they were critically reviewed, coded, and synthesized by the authors using thematic analysis principles [17]. This ensured that the discussion was author-driven, with Scopus AI serving as a supportive analytical aid to enhance transparency and structure.

Rationale for Approach

This methodological design is appropriate for a conceptual review that seeks to map intellectual structures, highlight thematic linkages, and identify future research directions. Unlike systematic reviews that emphasize reproducibility through numerical screening, this structured narrative review prioritizes conceptual integration and theoretical framing, making it suitable for consolidating insights in an emerging and rapidly evolving research area.

As illustrated in Figure 1, Scopus AI provides five analytical layers—Summary, Expanded Summary, Concept Map, Topic Experts, and Emerging Themes. In this review, these layers functioned as supportive tools to structure the identification of thematic clusters and intellectual linkages. The interpretation, coding, and synthesis, however, were carried out by the authors, ensuring academic rigor and critical depth [17], [12].

Fig. 1: 5 Core elements of Scopus AI

Having established the methodological approach, the next section presents the results of the structured narrative review. The findings are organized around the multi-layered outputs of Scopus AI—namely the Summary and Expanded Summary, the Concept Map, the Topic Experts, and the Emerging Themes—which served as analytical aids to structure the literature. Importantly, these outputs were not treated as final results; rather, they provided a framework that was critically examined, reclassified, and synthesized by the authors. This ensures that the discussion remains author-driven, moving beyond description to offer a comparative evaluation of themes, identification of contradictions, and integration with established theoretical models such as the Technology Acceptance Model (TAM) and Service-Dominant Logic (SDL).

DISCUSSION

The results of this review are organized through a structured synthesis guided by the multi-layered outputs of Scopus AI—namely the Summary, Expanded Summary, Concept Map, Topic Experts, and Emerging Themes. These features served as analytical aids to organize the literature, but the thematic interpretation and conceptual integration were conducted manually by the authors. This approach ensures that the discussion remains author-driven, moving beyond descriptive outputs to provide a critical synthesis of the scholarly landscape on advanced AI chatbots and their role in enhancing customer interaction.

Accordingly, the discussion proceeds in four parts. Section III-A draws on the Summary and Expanded Summary to highlight overall publication trends and disciplinary scope, with critical reflection on sector-specific contributions. Section III-B introduces the Concept Map as a starting point for identifying three major thematic clusters—implementation challenges, user experience, and technological innovations—which are then critically examined in Sections III-B1 to III-B3. Section III-C reviews the contributions of influential scholars, not as isolated outputs but as part of the broader intellectual structure of the field. Finally, Section III-D explores Emerging Themes, with a particular focus on unresolved contradictions, theoretical implications, and directions for future research.

This structure allows the review to move from descriptive mapping to comparative evaluation and theoretical interpretation, thereby aligning the findings with both established models (e.g., the Technology Acceptance Model, Service-Dominant Logic) and ongoing debates in the literature [10], [11].

Summary and Expanded Summary

The Summary of the Scopus dataset shows a steady growth of scholarly contributions on advanced AI chatbots over the past decade, underscoring their increasing significance in digital customer service ecosystems [1], [2]. Publications are distributed across computer science, business management, and service marketing, reflecting the inherently interdisciplinary nature of this research area. While the growth trajectory indicates maturity in chatbot adoption, it also reveals fragmentation: studies are often sector-specific and vary widely in methodological rigor. This fragmentation highlights the need for integrative reviews such as the present one, which consolidate insights across domains to generate a more comprehensive understanding of chatbot research.

The Expanded Summary provides deeper insight by clustering contributions according to methodological approaches and sector-specific applications. For example, in hospitality, chatbots are deployed for bookings and personalized assistance, improving service efficiency and guest satisfaction [5], [18]. In the banking sector, AI chatbots support transactional queries and fraud detection, reinforcing customer trust while also raising concerns about data security [6]. In e-commerce, GPT-powered chatbots have bridged gaps between user expectations and service delivery, leading to measurable improvements in satisfaction and loyalty [7], [19]. These sectoral comparisons demonstrate both the versatility and the uneven performance of chatbot adoption, suggesting that contextual factors play a decisive role in shaping effectiveness.

Critically, while many studies highlight efficiency gains and personalization, they also reveal contradictions in customer experience outcomes. Some report heightened satisfaction through adaptive, real-time support [20], whereas others point to frustration arising from limited empathy or perceived intrusiveness [8]. This divergence reflects the tension between technical sophistication and human-centric expectations, a recurring theme across the literature. The Expanded Summary thus indicates that the success of chatbot adoption cannot be explained solely by technological capabilities but must be examined in relation to trust, empathy, and transparency—factors that remain underexplored in current research.

Concept Map

The Concept Map generated from the Scopus dataset (Figure 2) provides a visual overview of the intellectual linkages shaping research on advanced AI chatbots in customer interaction. The map highlights three dominant clusters: implementation challenges, user experience, and technological innovations. Sub-clusters within these categories emphasize topics such as personalization, multimodal interactions, ethical considerations, and data privacy [9]. While the visualization offers a useful structural overview, it was employed in this review as a starting point for analysis rather than as a definitive result.

The clusters identified in the concept map were critically examined and reclassified by the authors to ensure conceptual clarity and interpretive rigor [17]. For instance, the “user experience” cluster encompassed studies that reported both positive outcomes, such as improved satisfaction and personalization [20], [7], and negative experiences, such as frustration caused by limited empathy [8]. Similarly, the “technological innovations” cluster reflected advancements in transformer-based models [3], [4], but also raised unresolved questions about bias, transparency, and long-term trust.

By integrating these clusters into a narrative synthesis, the discussion moves beyond description to provide a comparative evaluation of findings across contexts. The following subsections (III-B1, III-B2, III-B3), expand on each cluster—implementation challenges, user experience, and technological innovations—drawing on sector-specific evidence, highlighting contradictions, and linking insights to established theoretical frameworks such as the Technology Acceptance Model (TAM) [10] and Service-Dominant Logic (SDL) [11].

Figure 2: Advanced AI chatbots in customer interaction

A review on Advanced AI chatbots in customer interaction

The adoption of advanced AI chatbots has reshaped customer interaction across diverse industries, particularly in e-commerce, banking, hospitality, and education. Studies consistently emphasize their ability to enhance efficiency, reduce response times, and deliver personalized support at scale [7], [6], [5]. In banking and finance, for instance, chatbots support transactional queries and fraud detection, thereby strengthening customer trust and operational efficiency [6], [21]. In e-commerce, GPT-powered conversational agents have bridged gaps between customer expectations and service quality, leading to higher satisfaction and loyalty [7]. Within hospitality, chatbots have streamlined bookings and guest services, significantly improving service quality [5].

At the same time, the literature reveals contradictions in user experiences. While many studies report improved engagement and satisfaction through adaptive, real-time responses [20], others note frustration when chatbots fail to demonstrate empathy or contextual understanding [8]. These findings suggest that technological sophistication alone does not guarantee positive customer outcomes; rather, factors such as emotional resonance, transparency, and trust play a decisive role. Ethical concerns surrounding privacy and data protection further complicate adoption, especially in industries handling sensitive customer information [9].

Taken together, the evidence indicates that advanced AI chatbots represent a double-edged innovation: they deliver measurable efficiency and personalization benefits, but also expose unresolved challenges related to empathy, ethics, and user trust. To capture these dynamics more systematically, the following subsections (III-B1 to III-B3) examine the thematic clusters of implementation challenges, user experience, and technological innovations, offering a critical synthesis that integrates sectoral insights with relevant theoretical frameworks.

B1:The relationship between Advanced AI Chatbots in Customer Interaction with Implementation Challenges

The implementation of advanced AI chatbots in customer interaction is shaped by a complex set of technological, organizational, and ethical challenges. On the technological front, several studies highlight improvements in natural language processing and large language models, which enhance contextual accuracy and conversational flow [3], [4]. Yet, despite these advances, organizations continue to report difficulties in achieving seamless integration across platforms, particularly in contexts that require interoperability with legacy systems [22], [23]. This contradiction—where chatbots are technically sophisticated yet operationally constrained—underscores the importance of aligning innovation with organizational readiness, as highlighted by the Technology–Organization–Environment (TOE) framework [23].

Equally pressing are challenges linked to user trust and experience. While chatbots are praised for their efficiency and scalability, studies show that customers disengage when interactions feel mechanical or lack empathy [8], [24]. This tension reflects a broader managerial dilemma: investing in efficiency through automation may inadvertently undermine customer confidence if emotional resonance and transparency are neglected. The problem is particularly acute in industries such as banking and e-commerce, where sensitive personal data are routinely exchanged [9]. Here, the absence of clear safeguards around privacy and algorithmic fairness becomes a barrier to adoption, even when technological performance is strong [25].

Taken together, these findings suggest that implementation challenges cannot be reduced to technical limitations alone. They must be understood as the outcome of organizational alignment, user expectations, and ethical accountability. From a theoretical standpoint, the TOE framework helps explain why some organizations successfully adopt chatbots—by balancing technological capability with managerial readiness and regulatory compliance—while others struggle despite access to similar tools. Future research should therefore explore how firms can develop integrated implementation strategies that address not only technical feasibility but also the human and ethical dimensions of adoption.

B2: The relationship between Advanced AI Chatbots in Customer Interaction with User Experience

The adoption of advanced AI chatbots has had a profound influence on user experience, particularly in sectors such as e-commerce and digital services where immediacy and personalization are valued. Studies consistently highlight that GPT-powered and context-aware chatbots improve satisfaction by delivering real-time, accurate, and personalized responses [7], [20]. From the perspective of the Technology Acceptance Model (TAM), these benefits directly reinforce perceptions of usefulness and ease of use, which are central predictors of technology adoption [10]. By minimizing response delays and enabling dynamic personalization, advanced chatbots are positioned as valuable service tools that enhance customer engagement.

However, contradictory findings complicate this optimistic picture. Several studies emphasize that users often disengage when chatbots fail to demonstrate empathy, contextual awareness, or emotional intelligence [8], [24]. This reveals a gap between the perceived usefulness of chatbots as efficient service agents and the perceived ease of trust, which depends on the system’s ability to convey transparency and emotional resonance. For instance, while some customers report satisfaction with highly responsive systems, others express frustration when interactions feel mechanical or intrusive, particularly in sensitive service contexts such as financial advice or healthcare [9].

The literature also suggests that demographic factors, especially generational differences, shape user perceptions. Younger users, such as Gen Z, often value speed and convenience but remain critical of limitations in empathy and conversational depth [26]. This highlights the need for more user-centered design, where chatbot systems incorporate sentiment analysis, emotional recognition, and context-awareness to meet evolving expectations [16].

Taken together, the evidence indicates that while advanced chatbots positively influence user experience by improving efficiency and personalization, they remain constrained by emotional and trust deficits. TAM provides a useful explanatory lens: adoption is strongest when chatbots are perceived as both functionally useful and trustworthy in interaction design. Future research should therefore explore how emotional intelligence, anthropomorphism, and hybrid human–AI service models can strengthen both dimensions, ensuring that chatbots are not only efficient but also empathetic companions in customer interaction.

B3: The relationship between Advanced AI Chatbots in Customer Interaction with Technological Innovations

Technological innovations have been central to the transformative role of advanced AI chatbots in customer interaction. Breakthroughs in natural language processing (NLP) and machine learning (ML) have enabled chatbots to handle increasingly complex queries and provide context-aware responses, moving far beyond the limitations of earlier rule-based systems [3], [4]. Multimodal systems, which integrate text, voice, and even visual recognition, have expanded the utility of chatbots across industries such as e-commerce, hospitality, and banking [5], [7]. These innovations not only reduce response times and operational costs but also create opportunities for new forms of customer engagement and data-driven insights [27].

Yet, the literature also reveals tensions between technological advancement and customer expectations. While transformer-based and generative models (e.g., GPT) provide greater conversational accuracy, studies highlight persistent concerns over bias, transparency, and the limited ability of chatbots to replicate empathy [9], [26]. In industries where sensitive data are exchanged, such as finance and healthcare, customers remain cautious, often prioritizing trust and ethical safeguards over technical sophistication [25]. This indicates that technological innovation alone cannot guarantee adoption unless accompanied by accountability frameworks and user-centric safeguards.

From the perspective of Service-Dominant Logic (SDL), these innovations should be seen not as isolated technical upgrades but as enablers of value co-creation between firms and customers [11]. In this view, chatbots become more than service tools; they act as platforms through which customers and organizations jointly create value via interactive experiences. However, the potential for co-creation is undermined if chatbots fail to address ethical concerns or align with user trust. The literature therefore suggests a paradox: the more technologically advanced chatbots become, the greater the pressure to design systems that are transparent, inclusive, and emotionally intelligent.

In sum, technological innovations in advanced AI chatbots are driving efficiency and personalization, but their long-term success depends on embedding them within frameworks that prioritize trust, accountability, and co-created value. Future research should investigate how hybrid service models—where human agents complement technologically advanced chatbots—can balance efficiency with empathy, thereby fulfilling the promise of SDL in the digital customer service context.

Topic Expert – Influential Scholars and Intellectual Contributions

An important dimension of the literature on advanced AI chatbots concerns the contributions of influential scholars who have shaped the intellectual trajectory of the field. Rather than treating these “topic experts” as isolated outputs, their work can be understood as reinforcing the thematic clusters identified earlier.

For instance, Santhoshkumar [14] has advanced the integration of machine learning techniques into chatbot development, focusing on scalable personalization for mobile environments. His work directly contributes to the cluster on technological innovations, demonstrating how adaptive models can enhance real-time responsiveness and expand the scope of chatbot applications. Similarly, Nalini [15] has emphasized the importance of personalization and adaptability, exploring how chatbots can dynamically adjust to user preferences. Her research aligns closely with the cluster on user experience, underscoring that perceived usefulness and trust are contingent on systems’ ability to mirror individual needs.

Other scholars, such as Mohamed Sithik [28], have targeted the intersection of implementation challenges and user experience, highlighting how machine learning can support the development of intuitive, human-like interactions that address customer frustration. These contributions reveal that while individual researchers often approach chatbots from specific technical or design perspectives, their collective insights converge on broader issues of scalability, empathy, and ethical accountability.

Taken together, the intellectual contributions of these topic experts illustrate the state-of-the-art in chatbot scholarship: a field that is rapidly evolving from technical experimentation toward integrated, user-centered solutions. Their work reinforces the importance of framing chatbot adoption not only as a technological upgrade but also as a managerial and ethical challenge that requires alignment with customer expectations and organizational strategy.

Emerging Themes

The synthesis of the reviewed literature reveals several emerging themes that extend beyond transactional efficiency and signal the future trajectory of advanced AI chatbots. A consistent theme across multiple studies is the role of chatbots in enhancing customer engagement and satisfaction by providing personalized, real-time support [7], [20]. This positions chatbots as integral components of customer service ecosystems rather than experimental add-ons. However, the persistence of contradictory findings—where some users report high satisfaction while others express frustration with mechanical or intrusive responses [8], [9]—indicates that personalization and speed alone are insufficient. This contradiction highlights the growing importance of empathetic AI capable of emotional recognition and sentiment-aware responses.

Another rising theme is the integration of advanced NLP and transformer-based models, such as GPT and BERT, which improve contextual understanding and conversational accuracy [3], [4]. While these innovations enhance the perceived usefulness of chatbots, they simultaneously amplify concerns about bias, transparency, and data privacy. This duality suggests that technological innovation must evolve hand-in-hand with ethical frameworks if it is to sustain user trust in high-stakes industries such as finance and healthcare [25].

Novel directions also include the deployment of chatbots in education, where they are increasingly used to support administrative services and student engagement [16]. These applications illustrate the versatility of chatbots but also raise questions about inclusivity, accessibility, and the balance between automation and human interaction in sensitive domains. From a theoretical standpoint, these developments challenge the sufficiency of existing adoption models such as TAM [10], pointing toward the need for frameworks that incorporate trust, emotional intelligence, and human–AI collaboration.

Taken together, the emerging themes suggest that the future of chatbot research and practice lies not in maximizing efficiency alone, but in designing systems that are empathetic, ethically accountable, and co-create value with users. This trajectory aligns with broader discussions in Service-Dominant Logic (SDL) [11], where technology is viewed as a platform for value co-creation rather than a one-way service tool. Future research should therefore explore hybrid service models, cross-cultural trust dynamics, and regulatory frameworks that can support the responsible evolution of advanced AI chatbots.

Limitations

Despite its contributions, this study has several limitations that should be acknowledged. First, the review relied primarily on the Scopus database, which, although comprehensive, may have excluded relevant publications indexed in other databases or grey literature sources. Second, while Scopus AI’s features—such as summary, expanded summary, concept map, topic experts, and emerging themes—were employed to assist in mapping the intellectual structure of the field, these outputs were not treated as final results. Instead, they served as analytical aids that supported the authors’ interpretive synthesis. A key limitation, therefore, lies in the potential bias or incompleteness of AI-generated outputs, which may overemphasize certain clusters while overlooking others. To mitigate this, the authors conducted manual review, coding, and theoretical framing [17], [12], ensuring that the discussion and conclusions remained author-driven.

Third, this review did not include a numerical screening process, as is typical in systematic reviews, but instead adopted a structured narrative approach designed to synthesize emerging and fragmented literature. While this is appropriate for a conceptual review, it limits the reproducibility of study inclusion in precise quantitative terms. Finally, given the rapid pace of advancements in AI and NLP technologies, some findings may quickly become outdated as new models and applications emerge. Future reviews should therefore update the synthesis with multi-database searches and consider triangulating AI-aided analysis with traditional bibliometric techniques for greater robustness.

CONCLUSIONS

This review has synthesized the scholarly landscape on advanced AI chatbots in customer interaction, integrating insights across diverse sectors and theoretical perspectives. The analysis revealed that chatbots have transformed customer engagement by enhancing personalization, responsiveness, and scalability, particularly in e-commerce, banking, and hospitality [7], [6], [5]. At the same time, persistent challenges—such as empathy deficits, ethical concerns, and trust barriers—highlight that technological sophistication alone does not guarantee user acceptance [8], [9].

Theoretically, this review contributes by linking chatbot adoption and interaction outcomes to established frameworks such as the Technology Acceptance Model (TAM), which emphasizes perceived usefulness, ease of use, and trust [10], and Service-Dominant Logic (SDL), which frames chatbots as co-creators of value with customers [11]. Practically, it highlights that organizations must balance efficiency with empathy and transparency, adopting hybrid human–AI service models and embedding accountability mechanisms to foster sustainable adoption.

Overall, while Scopus AI provided valuable analytical features that structured the review, the interpretive synthesis was conducted by the authors to ensure conceptual depth and critical evaluation. By explicitly acknowledging the limitations of AI-assisted review and reframing the methodology as a structured narrative design, this study enhances transparency and rigor. Future research should build on these insights through empirical validation, cross-cultural comparisons, and deeper exploration of empathetic AI, privacy-preserving design, and hybrid service frameworks.

ACKNOWLEDGMENT

The authors would like to show their deepest gratitude to Universiti Teknologi MARA (UiTM) Kedah Campus in their constant support and to the anonymous reviewers who have provided them with insightful feedback that has improved this article in its quality and clarity significantly.

REFERENCES

1. K. Patel, “Customer service bots: Enhancing support and personalization,” in Design and Development of Emerging Chatbot Technology, pp. 75–92, IGI Global, 2024. doi: 10.4018/979-8-3693-1830-0.ch009
2. T. Elavarasi, M. Jayaprakash, E. Shivakumar, N. Savitha, K. C. Gayathri, and M. Agoramoorthy, “Designing and developing intelligent chatbots with natural language processing through a conversational AI approach,” in Proc. 5th Int. Conf. on Pervasive Computing and Social Networking (ICPCSN 2025), pp. 1800–1805, IEEE, 2025. doi: 10.1109/ICPCSN65854.2025.11035776
3. M. S. Thasleem, K. M. Varma, and R. Sethuraman, “Customer support chatbot development using BERT,” in Proc. 7th Int. Conf. on Intelligent Sustainable Systems (ICISS 2025), pp. 189–194, IEEE, 2025. doi: 10.1109/ICISS63372.2025.11076409
4. S. K. Zelani, S. G. Begum, M. Chandana, and N. L. Tirupatamma, “Design of novel chatbot using generative artificial intelligence,” in Quantum Computing Models for Cybersecurity and Wireless Communications, pp. 251–258, Wiley, 2025. doi: 10.1002/9781394271429.ch16
5. N. Suhag, D. Sarkar, and A. Singh, “Exploring the impact of AI-driven chatbots on customer service in the hospitality industry,” in Integrating AI-driven Technologies into Service Marketing, pp. 211–224, IGI Global, 2024. doi: 10.4018/979-8-3693-7122-0.ch012
6. S. Singh, B. Sharma, and Y. Awasthi, “Chatbots in banking: Transforming customer interaction and service efficiency through AI,” in Cultural Sensitivity, Cross-Border Logistics, and E-commerce in Global Marketing, pp. 375–396, IGI Global, 2025. doi: 10.4018/979-8-3373-0528-8.ch018
7. A. Mohamed and I. Ismail, “Revolutionizing e-commerce customer service: Leveraging GPT-powered AI chatbots to bridge the user expectation gap,” in Lecture Notes in Electrical Engineering, vol. 1228, pp. 457–468, Springer, 2025. doi: 10.1007/978-981-97-4784-9_33
8. D. Lestarini and S. P. Raflesia, “Towards AI chatbot effectiveness: Crucial elements impacting user engagement and satisfaction,” in ICAAEEI 2024 – 1st Int. Conf. of Adisutjipto on Aerospace Electrical Engineering and Informatics, pp. 221–229, IEEE, 2024.
9. Z. Xia and R. Shannon, “Navigating the digital frontier: Exploring the dynamics of customer–brand relationships through AI chatbots,” Sustainability, vol. 17, no. 5, p. 2173, 2025.
10. F. D. Davis, “Perceived usefulness, perceived ease of use, and user acceptance of information technology,” MIS Quarterly, vol. 13, no. 3, pp. 319–340, 1989.
11. S. L. Vargo and R. F. Lusch, “Service-dominant logic: Continuing the evolution,” J. of the Acad. of Marketing Science, vol. 36, no. 1, pp. 1–10, 2008.
12. H. Snyder, “Literature review as a research methodology: An overview and guidelines,” Journal of Business Research, vol. 104, pp. 333–339, 2019. doi: 10.1016/j.jbusres.2019.07.039
13. P. V. Nagare and P. Shirke, “A comprehensive review of efficient multimedia smart chatbots for customer service: Leveraging artificial intelligence and machine learning,” AIP Conf. Proc., vol. 3327, no. 1, 020033, 2025. doi: 10.1063/5.0289454
14. S. P. Santhoshkumar, “Machine learning integration for personalized mobile chatbots: Improving user interaction and engagement,” Journal of Intelligent Systems and Applications, vol. 32, no. 3, pp. 145–158, 2024.
15. T. Nalini, “Personalized mobile chatbot development: Enhancing user preference adaptability in AI-driven systems,” in Proc. Int. Conf. on Machine Learning and Artificial Intelligence, pp. 201–210, IEEE, 2024.
16. T. Y. Zhou, Y. Liu, and M. Kumar, “AI-powered chatbots for improving interactive user experience: State-of-the-art,” in Proc. Int. Conf. on Computers and Industrial Engineering (CIE 2024), pp. 301–308, IEEE, 2024.
17. V. Braun and V. Clarke, “Using thematic analysis in psychology,” Qualitative Research in Psychology, vol. 3, no. 2, pp. 77–101, 2006. doi: 10.1191/1478088706qp063oa
18. B. Aggarwal, P. Kumar, and R. Verma, “Technological advancements in the hospitality sector: The role of intelligent chatbots,” in Impact of AI and Tech-driven Solutions in Hospitality and Tourism, pp. 282–301, IGI Global, 2024. doi: 10.4018/979-8-3693-6755-1.ch015
19. T. C. Mhiribidi, E. T. Chigodo, and V. Jowa, “Conversational commerce: Designing and evaluating a WhatsApp-based e-commerce chatbot,” in Proc. 3rd Zimbabwe Conf. on Information Communication and Technology (ZCICT 2024), IEEE, 2024. doi: 10.1109/ZCICT63770.2024.10958353
20. S. Uzan, D. Freud, and A. Elalouf, “Optimizing chatbots to improve customer experience and satisfaction: Research on personalization, empathy, and feedback analysis,” Applied Sciences, vol. 15, no. 17, p. 9439, 2025. doi: 10.3390/app15179439
21. A. Yella and M. Vivek Kumar, “Integrating AI chatbots in customer service for credit card companies,” in Proc. Int. Conf. on Networks and Cryptology (NETCRYPT 2025), pp. 1157–1161, IEEE, 2025. doi: 10.1109/NETCRYPT65877.2025.11102796
22. G. Parasa, D. K. Nayak, A. Gangopadhyay, M. Stanlywit, S. Gokulakrishnan, and P. Sharma, “Exploring the role of artificial intelligence in enhancing chatbot functionality,” in Proc. 1st Int. Conf. on Advances in Computing, Communication and Networking (ICAC2N 2024), pp. 1647–1651, IEEE, 2024. doi: 10.1109/ICAC2N63387.2024.10895958
23. R. Urbani, C. Ferreira, and J. Lam, “Managerial framework for evaluating AI chatbot integration: Bridging organizational readiness and technological challenges,” Business Horizons, vol. 67, no. 5, pp. 595–606, 2024. doi: 10.1016/j.bushor.2024.05.004
24. J. J. A. Ocxas, F. K. C. Martin, and A. N. H. Junes, “AI chatbot with emotional recognition for an improved customer service experience,” in Proc. 31st Int. Conf. on Electronics, Electrical Engineering and Computing (INTERCON 2024), pp. 411–416, IEEE, 2024.
25. S. V. Siva Kumar and S. Metta, “AI-powered customer service: The role of chatbots in enhancing e-commerce interactions,” in Navigating Data Science in the Age of AI: Exploring Possibilities of Generative Intelligence, pp. 67–82, Emerald Publishing, 2025. doi: 10.1108/978-1-83608-432-720251004
26. S. Saklani and D. Kala, “Perception of Gen Z customers towards chatbots as service agents,” Journal of Telecommunications and the Digital Economy, vol. 12, no. 1, pp. 356–376, 2024. doi: 10.18080/jtde.v12n1.781
27. C. Khneyzer, K. S. Rebeiz, and J. Touma, “A comprehensive review of economic and managerial factors: The implications of AI chatbots in customer relationship management,” in Navigating Business Through Essential Sustainable Strategies, pp. 217–237, IGI Global, 2024. doi: 10.4018/979-8-3693-5360-8.ch009
28. M. Mohamed Sithik, “Advanced conversational agents through machine learning: Enhancing customer interaction with personalized experiences,” International Journal of Artificial Intelligence Research, vol. 18, no. 2, pp. 211–224, 2024.