Abstract
Artificial Intelligence (AI) technologies have become increasingly integrated into travel experiences. Understanding how and why travelers adopt these technologies and whether they will continue to use them is essential. This study adopts a qualitative approach to examine how travelers interact with AI during their journeys. Data were collected through semi-structured interviews with 50 Omani travelers who had used AI tools while traveling and analyzed using thematic analysis guided by abductive reasoning. The findings reveal a range of motivations for adoption, including functional, social, and emotional drivers, as well as key challenges such as impersonal recommendations, technical limitations, and privacy concerns that influence satisfaction and future use. Theoretically, the study develops the MAO (Motivation–Adoption–Outcome) model, which extends existing technology adoption theories. The study offers practical implications for developers and tourism providers aiming to design AI tools that are more human-centered, culturally relevant, and emotionally engaging.
Keywords
Introduction
The rapid progress of artificial intelligence (AI) will result in a major change in the travel, tourism, and hospitality sectors (Ling et al., 2025). AI is defined by Gretzel et al. (2015) as the simulation of human intelligence processes by computer systems, enabling higher levels of automation, efficiency, and personalization. AI applications have evolved significantly since the mid-20th century, when they exhibited only rudimentary functions, to today’s advanced systems that incorporate technologies such as robots, facial recognition, and complex problem-solving capabilities (Ivanov et al., 2019; Samala et al., 2022). While AI has been used for decades, it has recently emerged as a significant component across various sectors—particularly tourism—due to its rapid advancement in transforming how travelers plan and experience their journeys (Kang et al., 2026).
AI is rapidly changing how travelers interact with destinations and how they make travel-related decisions (Kang et al., 2026; Ling et al., 2025). AI tools—such as virtual assistants, chatbots, predictive analytics, and robotic service agents—continue to evolve as they become integrated into the digital travel ecosystem (Sun et al., 2025; Tussyadiah, 2020). These tools play a crucial role in the pre-travel stage by enabling access to timely and personalized recommendations that support travelers’ decision-making (C. Koo et al., 2021). Consequently, AI technologies enhance travelers’ overall destination experience by providing seamless convenience and improved service delivery (Sun et al., 2025).
Despite these benefits, much of the existing research on AI in travel reflects a pro-innovation bias, emphasizing adoption drivers while giving limited attention to travelers’ concerns, hesitation, and resistance (Seyfi et al., 2026; Seyfi et al., 2025; Seyfi et al., 2025). This tendency assumes that new technologies will naturally be embraced, overlooking the emotional, cultural, and psychological factors that may impede adoption. Recent studies challenge this assumption by showing that travelers’ interactions with AI are not universally positive. Seyfi et al. (2026) demonstrate that many tourists experience functional and psychological barriers when considering generative AI for travel planning, including concerns about usage difficulties, perceived risks, and conflicts with cultural traditions. These concerns align with broader drawbacks identified in the literature, where AI tools raise issues related to privacy risks, trustworthiness, data breaches, ethics, and transparency (Hu & Min, 2023; Knani et al., 2022; Shuqair et al., 2024; Yhee et al., 2025). Additionally, many users remain hesitant to adopt AI tools due to the perceived lack of warmth, authenticity, and personal connection within tourism experiences (Liu et al., 2024; Roy et al., 2024). Issues related to the quality and relevance of information provided by AI tools also represent a significant barrier (Carvalho & Ivanov, 2024). J. H. Kim et al. (2025) further note that when inaccurate information is generated by AI tools such as ChatGPT, travelers’ willingness to rely on such solutions may decline. Supporting this viewpoint, Seyfi, Gorji, et al. (2025) show that many travelers perceive AI-generated travel advice as untrustworthy or lacking authenticity, revealing resistance patterns such as “rejecters,” “postponers,” and “opinion leaders.”
These persistent limitations highlight the need for a more comprehensive understanding of AI adoption in travel, particularly regarding user motivations, satisfaction, and continuance usage intention. Building on this, the present study advances knowledge in this area by addressing the following critical gaps in the literature. First, although the adoption of AI-driven technologies is increasing in service settings, including travel and tourism (Gursoy et al., 2019; Ivanov et al., 2019), most previous studies (e.g., Chi et al., 2022; I. Koo et al., 2025; Sundar et al., 2016) rely on traditional technology adoption theories, such as the Technology Acceptance Model (TAM; Davis, 1989) and the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2; Venkatesh et al., 2012), to explain this behavior. However, recent research shows that travelers’ decisions around AI are shaped by more complex dynamics. For example, Seyfi et al. (2026) reveal that AI adoption in travel contexts is negatively impacted by functional barriers such as risk and usage difficulty, as well as psychological barriers linked to image and cultural tradition. Similarly, Seyfi, Kim, et al. (2025) demonstrate that individual personality traits influence travelers’ openness or resistance to AI-generated travel recommendations, pointing to emotional and identity-based mechanisms not captured by traditional models. These insights highlight that AI adoption is fundamentally different from the adoption of non-intelligent technologies because AI systems continuously learn, adapt, and interact with users in personalized ways (Gursoy et al., 2019; D. Kim et al., 2026). Therefore, the present study addresses this gap by extending traditional technology adoption models through the integration of additional factors that better capture the complexities of AI adoption in travel.
Second, a significant gap in previous research is related to the limited attention given to measuring the outcomes of technology adoption. Most of the previous studies focused on the initial adoption of technology while overlooking the outcomes of that adoption, such as satisfaction and continuance usage intention (Bhattacherjee, 2001; C. Koo et al., 2021). This is particularly important for AI technologies, as their adaptive, interactive, and learning capabilities make post-adoption experiences critical in shaping user satisfaction, and continuance usage intention (Ku & Chen, 2024). In addition, the rapid advancement of AI technologies and the continuous emergence of new types and models require AI developers to place greater emphasis on satisfaction and continuance usage intention, since consumer preferences can change quickly (Huang et al., 2024; Ku & Chen, 2024). To address this gap, the current study aims to develop a conceptual model, conceptualizing AI adoption as a three-phase process encompassing motivation, adoption, and outcomes.
Third, a further research gap is related to the context of the study. While existing research has examined traveler acceptance of AI tools, it has largely focused on Western and East Asian contexts (e.g., Sadiq et al., 2025; Seyfi et al., 2026). In contrast, studies specific to the Middle East in general and to Oman in particular in particular remain relatively limited, particularly in understanding the cultural, emotional, and motivational factors that influence the use of AI tools in travel. This gap limits the ability of travel providers and policymakers to deliver culturally aligned, effective, and personalized AI-based services (Ali et al., 2023).
To address these research gaps, this study aims to achieve three main objectives. First, to explore the types of AI tools and the motivations behind travelers’ adoption of these tools during the pre-travel and during-travel stages. Second, to identify how AI tools are used, the challenges that negatively affect travelers’ satisfaction and continuance usage intention, and the improvements that could enhance satisfaction and promote sustained use. Third, to develop a theoretical model that explains the relationships among motivation, adoption, satisfaction, and continuance usage intention of AI tools in travel, building on and extending existing technology adoption theories. Specifically, the study addresses three key research questions:
To answer these research questions, the study adopts a qualitative methodology, drawing on semi-structured interviews with Omani travelers who have experience with AI tools in travel.
This study contributes to both theoretical and practical understanding of AI adoption in travel. The key contribution lies in extending traditional technology adoption constructs to account for the unique characteristics of AI tools, while integrating them with post-adoption outcome measures such as satisfaction and continuance usage intention. By doing so, the study demonstrates how traditional technology adoption frameworks can be expanded to capture the complex, adaptive, and dynamic nature of AI adoption behavior. Theoretically, the study develops the Motivation–Adoption–Outcome (MAO) model as a comprehensive framework to guide future research on technology adoption in AI-enabled tourism contexts, while considering users’ preferences and the distinctive nature of AI tools compared with traditional, non-intelligent technologies. Practically, the study provides actionable insights for AI developers and tourism service providers on designing AI tools that are personalized and better aligned with travelers’ needs, ultimately enhancing user satisfaction and promoting continuance usage intention.
Literature Review
Artificial Intelligence in Travel and Tourism
The tourism industry has experienced major changes due to the growth of digital technologies (Ling et al., 2025). Historically, travelers relied on travel agencies that provided simple and non-engaging flyers which contained basic and limited travel information (Joyce, 2013; Xiang, 2018). Early digital tools lacked personalization and most of the travel services were handled manually (Joyce, 2013). As internet technologies evolved, tourism services became more dynamic, personalized, and customer-centric (Xiang, 2018). Recent technologies, such as AI-powered virtual assistants and recommendation systems, have taken interaction and engagement with tourists and travelers in real-time to a new level, meeting the growing demand for seamless and relevant experiences (Ali et al., 2023; Dwivedi et al., 2024).
AI systems and tools have become a standard part of the operational and service landscape of the tourism and travel industry (Prentice et al., 2020; Roy et al., 2024; C. X. Wang et al., 2020). For tourists themselves, AI tools can improve the quality of tourists’ experiences, and offer personalized assistance for itinerary recommendations, translation services or give tourists cultural cues or hints (Gretzel et al., 2015; Q. Zhang & Ahmad, 2024). For tourism service providers AI can also provide personalization or customization of services and experiences for individual tourists, based on that individual tourist’s preferences. AI systems and tools can also assist service providers operationally, by analyzing the vast amounts of information created by tourists to enhance customer experiences and optimizing tourism services (Lopes & Estevão, 2024; Sorokina et al., 2022; Traversa, 2024).
In recent years, as social robotics are integrated with AI technology, there is great potential for innovation in the tourism industry with technology (C. Koo et al., 2021). This integration has enhanced the way people experience service delivery through intelligent robots, allowing for faster and more personalized experiences (Skubis, 2025). The social robotics use within tourism mainly refers to personal, social, and interactive technologies that are driven by AI and can bring value through social and interactive competencies to service delivery and consumer interaction (Sun et al., 2025; Tussyadiah, 2020). In the tourism industry, service robots vary in their level of intelligence and can therefore be classified into different categories. For instance, some robots perform basic tasks, while others are more advanced and capable of interacting directly with customers (Tussyadiah, 2020).
AI-Driven Tools Used by Travelers
More and more travelers today are using AI tools to make their travel experiences easier and more enjoyable (H. Zhang et al., 2025). There are many types of AI tools, but they can range in function from planning and booking to navigation and post-travel reviews (J. Wang, 2025). A popular type of AI system is personalization engines (Della Corte et al., 2025). Personalization engines take customer data such as individual preferences, previous choices, and past customer behavior to generate personalized accommodation, activities, dining options, and overall travel experiences (Della Corte et al., 2025). To do this these engines rely on natural language processing (NLP), deep learning (DL), and machine learning (ML) technology (Della Corte et al., 2025; Traversa, 2024). For instance, Airbnb, Expedia, Booking.com, and TripAdvisor offer accommodation, restaurant, attraction, and travel recommendations based on previous bookings (Ali et al., 2023).
Nowadays, travelers are increasingly turning to generative AI tools (H. Zhang et al., 2025). Tools such as ChatGPT, Gemini, Claude, Cohere, and Bard are used more and more frequently by travelers (Sigala et al., 2024). These tools assist travelers with all types of tasks, including personalized itinerary planning, responding to questions, and assisting with hotel, flights, and ground transfer arrangements (Sigala et al., 2024; H. Zhang et al., 2025). Additionally, Chatbots and virtual assistants are among the most used tools. They can automate repetitive tasks, provide answers to frequently asked questions, offer customer support, assist with bookings and reservations on websites, mobile applications, and messaging platforms, and conduct small transactions (S. Gupta et al., 2023).
Travelers use a number of other AI tools to facilitate certain tasks related to their travel experiences. For example, search and booking engines are commonly used by travelers for information and reservations (Yin et al., 2024). Also, tourism service providers have started offering self-service screens which include facial recognition to facilitate the check-in, payment and other interactions by customers (Yin et al., 2024). Similarly, augmented reality (AR) and navigation tools enable tourists to navigate unfamiliar locations using a digital overlay of information over the real world (Della Corte et al., 2025).
Applications of AI Before and During Travel
AI applications span across the different stages of the travel cycle from planning the trip to post-trip reflection and engagement (Manoharan et al., 2024). During the pre-travel stage, travelers rely heavily on online sources such as search engines to gather and evaluate travel information (Xiang et al., 2015). However, at this stage, travelers often face challenges due to the large amount of information available (Xiang & Gretzel, 2010). Nowadays, many intelligent tools help travelers access relevant information and reduce information overload (Qassimi & Rakrak, 2025). For instance, chatbots are increasingly used to assist travelers in handling pre-trip inquiries and providing recommendations (Li et al., 2025). Chatbots function like digital shadows of human, capable of understanding and processing natural language to provide the appropriate and personalized responses while handling basic tasks (Deepika et al., 2020).
In the pre-travel phase, AI chatbots such as ChatGPT are becoming increasingly beneficial for travelers seeking personalized and efficient support. According to Dwivedi et al. (2024) when users ask questions, for example “places to visit in Rome,” ChatGPT not only provide popular suggestions to visit in Rome but also highlights the hidden gems, local cuisine, restaurant’s locations, the best time to visit, and the price ranges. Moreover, ChatGPT can guide users to websites for making reservations and provide suggestions for different types of booking (Chen et al., 2021; Mich & Garigliano, 2023). Despite the advantages that platforms such as ChatGPT provide, Carvalho and Ivanov (2024) highlighted that travelers should remain careful as the platform depends on existing data that may sometimes be outdated and inaccurate which leads to unreliable recommendations. Moreover, AI tools, such as those using simulated annealing techniques can streamline the creation of detailed travel itineraries, saving time and improving organization (Abram & Tedjojuwono, 2020).
During the travel stage, AI enhances real-time decision-making (Li et al., 2025; Wei, 2022). For example, AI-based recommendation systems (such as those of Airbnb, Booking.com and Expedia) provide personalized recommendations based on user data (Ali et al., 2023). Similarly, AI applications, such as facial recognition and biometric checks, assist with automated border control (Ioannou et al., 2020). With cultural and recreation settings, we are seeing AI robots and digital tour guides being used more frequently in museums, galleries and other tourist attractions to deliver reliable and engaging information to tourists (Li et al., 2025; J. Wang, 2025).
Theoretical Foundations
Understanding the adoption and use of AI tools among travelers calls for a theoretical approach that captures both the motivations for their initial use and the experiences that influence continuance usage intention. To inform this understanding, the present study draws on several well-established perspectives from technology adoption research, including the Technology Acceptance Model (TAM; Davis, 1989), Unified Theory of Acceptance and Use of Technology 2 (UTAUT2; Venkatesh et al., 2012) and Expectation Confirmation Model (ECM; Bhattacherjee, 2001). These three models have been widely applied to various service industries, including tourism and hospitality, to examine both technology adoption behavior (TAM and UTAUT2) and post-adoption outcomes such as satisfaction and continuance usage intention (ECM).
TAM and UTAUT2 remain relevant for understanding the cognitive and motivational processes affecting travelers’ initial perceptions of new technologies such as artificial intelligence. TAM focuses on perceived usefulness and perceived ease of use as important predictors of intention, whereas UTAUT2 extends this perspective by incorporating additional constructs such as habit, hedonic motivation and facilitating conditions. Their relevance in tourism contexts has been demonstrated in several studies. For example, T. Zhang and Xiong (2024) expanded TAM by integrating expectancy theory to investigate the adoption of virtual reality (VR) in tourism by Chinese tourists, while A. Gupta et al. (2018) used UTAUT2 to explain the use of travel-related mobile applications by tourists. Similarly, Radic et al. (2022) combined TAM and UTAUT2 to study cryptocurrency payment adoption by travelers in South Korea and China. ECM complements these models by providing a framework for the understanding of post-adoption evaluation. Specifically, ECM explains how confirmation of expectations and perceived performance influence satisfaction and continuance usage intention. In the context of tourism, Huang et al. (2024) demonstrated that customer satisfaction with AI tools is a significant predictor of their future intentions to use them. Given that generative AI tools improve through repeated interaction, understanding post-adoption experiences becomes particularly important for explaining long-term engagement.
While these theories continue to provide valuable insights into the cognitive evaluation of technologies, recent research suggests that AI adoption involves additional complexities that extend beyond the traditional constructs of technology adoption theories. Functional barriers such as accuracy, privacy, risk, and complexity of use have been repeatedly documented not only in tourism (Seyfi et al., 2026) but also in education (Lin et al., 2025), workplace environments (Al-Emran et al., 2025) and service organizations (R. Gupta & Rathore, 2024). These findings suggest that travelers evaluate AI tools using a broader set of criteria than those traditionally captured by established adoption models. Psychological and cultural factors also play an important role. For instance, concerns related to image and tradition have been found to influence travelers’ interactions with AI technologies (Seyfi et al., 2026). Similar patterns have been found in other industries where technological anxiety, confidentiality concerns, and loss of human interaction have been found to affect acceptance (Chow & To, 2025; de la Fuente Tambo et al., 2025).
Based on the above discussion, it is evident that complementing TAM, UTAUT2 and ECM with an additional perspective is valuable to understand the reasons that travelers may hesitate or resist using AI tools. Innovation Resistance Theory (IRT), first proposed by Ram (1987) and subsequently developed by Ram and Sheth (1989) provides such a lens. Rather than focusing on what motivates people to adopt technology, IRT focuses on what are the barriers preventing and delaying adoption. It makes a distinction between functional barriers (related to use, value and risk) and psychological barriers (related to image, tradition and social expectations). Recent tourism research supports the relevance of this perspective. For example, by applying the lens of the IRT theory, Seyfi, Kim, et al. (2025) discovered that many travelers are still reluctant to adopt generative AI for travel advice and identified five key barriers: usage, value, risk, image, and tradition. Although IRT is not applied as a full explanatory model in this study, its concepts provide an important lens for interpreting travelers’ concerns and resistance toward AI tools.
Taken together, these four theoretical perspectives provide a comprehensive lens through which to examine travelers’ adoption and continuance usage intention of AI tools, as well as barriers that may hinder their use. TAM and UTAUT2 offer insight into the cognitive and motivational drivers of initial adoption; ECM explains post-adoption evaluation and continued usage intentions, and IRT contributes an understanding of the resistance factors that may shape adoption decisions. By integrating these perspectives, the present study aims to develop the Motivation–Adoption–Outcome (MAO) model, which extends traditional technology adoption frameworks by incorporating travelers’ motivations, resistance factors, and post-adoption outcomes in the context of AI adoption in travel. A visual representation of the conceptual model guiding the study is presented in Figure 1.
Conceptual model.
Methodology
Research Design and Approach
This study adopted a qualitative research design, as it allows for capturing the depth and complexity of participants’ actual experiences (Creswell & Poth, 2016). The data collection and analysis were guided by thematic analysis informed by abductive reasoning (Braun & Clarke, 2006; Timmermans & Tavory, 2012). Thematic analysis enables researchers to identify and interpret patterns emerging from interview data, while abductive reasoning supports an iterative process of moving between empirical observations and existing theoretical perspectives (Reichertz, 2009; Timmermans & Tavory, 2012). In this study, abductive reasoning was adopted to allow themes to emerge from the data while also considering existing theories in technology adoption. This iterative movement is widely recognized as a defining feature of abductive analysis, enabling researchers to refine explanations when empirical observations do not fully align with existing frameworks (Reichertz, 2009; Timmermans & Tavory, 2012). In the context of this study, abductive reasoning enabled the researchers to interpret travelers’ perceptions not only through the lens of existing technology adoption theories (TAM, UTAUT2, ECM, and IRT), but also by identifying new themes specific to AI tool adoption in travel.
A summary of the methodological steps followed in this study is presented in Figure 2, outlining the sequential flow from research design through to data analysis.
Methodological framework for exploring AI adoption in travel.
Context of the Study
This research was conducted in Oman, a country marked by its increasing global mobility, growing outbound travel trends, and rising engagement with digital technologies. Oman recorded 5.19 million outbound trips in 2022 (National Centre for Statistics & Information, 2024), with 3.49 million made by Omani nationals, accounting for over 67% of all outbound tourism that year (Ministry of Heritage and Tourism, n.d.). These figures reflect a strong preference among Omanis to spend their holidays abroad. As Omani travelers become increasingly reliant on AI tools during international trips, the country offers a unique context in which to explore the intersection of AI and travel behavior. Despite this trend, there has been a lack of academic focus on how travelers from Oman perceive and interact with AI technologies throughout the travel cycle. Thus, Oman was purposefully selected as a culturally and contextually relevant setting to generate in-depth insights that are underrepresented in the current body of tourism and technology literature.
Sampling and Ethical Considerations
A purposive sampling strategy was employed as the primary method to recruit participants who had relevant experiences with AI tools in the context of travel. This sampling method is commonly used in qualitative research to ensure that participants possess specific knowledge or experiences relevant to the research objectives (Patton, 2015). In this study, purposive sampling was applied to ensure that the participants were Omanis, representing different ages, genders, and educational backgrounds, had interacted with AI tools during the pre-travel and travel stages, and were able to recall the interaction. To broaden and diversify the sample while maintaining relevance to the study’s criteria, snowball sampling was also employed to recruit additional respondents through participant referrals. To minimize potential snowball sampling bias, the same eligibility criteria used for purposive sampling were applied to all referred participants. This ensured that referrals met the study requirements while contributing diverse perspectives. Ethical approval was obtained in accordance with institutional guidelines. The study was classified at Level 0 (minimal risk) under university policy, and therefore ethical approval was granted by the Head of Department. Participants were informed about the purpose of the study, the confidentiality of their information and their rights to withdraw at any time. Verbal consent to participate in the study was obtained prior to each interview.
Data Collection
Data Collection Tool
Semi-structured interviews were used as the primary data collection method. Semi-structured interviews were chosen because they enable the researchers to explore predefined themes while also allowing for deeper insights to emerge organically from participants’ experiences (Kallio et al., 2016). This flexibility aligns well with the study’s methodological approach, which combines thematic analysis with abductive reasoning, enabling researchers to remain open to new themes emerging from the data while also interpreting them in light of existing theoretical perspectives (Kallio et al., 2016; Timmermans & Tavory, 2012). The interview guide consisted of open-ended questions that explored the demographics of the participants, the type and timing of AI tool usage (e.g., planning vs. during travel), motivations for using AI tools, how travelers interacted with AI tools, challenges and limitations experienced, recommendations for improvement, and satisfaction and continuance usage intentions. Table 1 summarizes the key interview questions and their alignment with the study’s objectives.
Alignment Between the Research Objectives and the Interview Questions.
Data Collection Process
To ensure contextual relevance and reach participants experienced with AI tools, participants were carefully selected from settings where interaction with such technology was expected. First, universities were visited to interview both academic staff and students, recognizing their frequent international travel for educational, professional, or leisure purposes. Second, Oman International Airport was also chosen as a strategic site to engage with Omani travelers and professional staff, where participants shared their personal travel experiences. To broaden and diversify the sample while maintaining relevance to the study’s criteria, snowball sampling was employed to recruit additional respondents through referrals. Prior to each interview, participants were asked a set of screening questions to confirm that they met the study’s eligibility criteria, including being Omani travelers who had previously used AI tools during the pre-travel or travel stages, and were able to recall the interaction. Individuals who did not meet these criteria were thanked for their time, and the interview was not continued. The data collection process lasted for 2 months from May to June 2025. A total of 50 semi-structured interviews were conducted with an average time of 20 min per interview. Table 2 provides a summary of the demographics of the participants.
Demographic Profile of Participants (N = 50).
Data Analysis and Validation
After completing the interviews, they were transcribed and analyzed. The analysis followed Braun and Clarke’s (2006) six-phase framework. Since the current study adopts thematic analysis guided by abductive reasoning for data analysis, Braun and Clarke’s (2006) framework is particularly suitable, as its phases are iterative and allows for both theory-driven and open coding (Wilson et al., 2021). In the first phase, the research team became acquainted with the dataset by reading and then re-reading the transcripts and then discussing the preliminary observations. In the second phase, the transcripts were imported into NVivo 15 for coding and analysis using a combination of a priori and open coding approach (Braun & Clarke, 2006). Throughout the coding process, each code was linked to specific transcript references.
In the third phase, codes were further refined to identify relationships between them and organize them into meaningful themes and sub-themes. In phase four, the themes and sub-themes were further refined and reviewed to ensure they accurately represented the data. In phase five, the themes and sub-themes were finalized and validated by the research team, and a thematic structure was developed for each theme. In the final phase, the findings were presented under the validated themes, and a theoretical model was developed to represent the relationship between the different themes.
In keeping with the abductive approach, the analytical process involved going back and forth between the emerging codes from the data and existing theoretical concepts. This became particularly important when participants described experiences that went beyond a priori codes. For example, when coding the motivations to use AI, there were many emerging insights that aligned well with established technology adoption constructs such as perceived usefulness, ease of use, enjoyment during interaction with AI tools, and the influence of friends or family members. However, several participants described motivations not fully captured by these traditional frameworks, such as trust in AI generated responses, value of getting personalized recommendations, curiosity or novelty-seeking tendencies, and feelings of comfort when relying on AI during travel planning.
Through this iterative movement between theoretical perspectives and empirical insights, the final themes on functional, emotional, and social motivations, adoption behaviors, challenges to adoption, suggested improvements, and post-adoption outcomes were shaped by both participants’ experiences and the theoretical foundations of the study. This ensured that the resulting theoretical model was both theoretically informed and based on the experiences of travelers using AI tools.
Results
Theme 1: Usage of AI Tools Across the Travel Journey
This theme explores the various AI tools that travelers interacted with before and during their trips. It also examines how these tools were used in different phases of the travel journey. Rather than focusing on the tools adopted by travelers, the analysis reveals how these tools were used during the pre-travel and travel stages of the travel cycle.
Pre-Travel Use of AI Tools
Types of AI Tools
During the pre-travel phase, participants mentioned several AI tools used to support their travel planning. ChatGPT was the most widely used tool, pointing to its popularity among Omani travelers. Chatbots were the second most common AI tool used by individuals. Other tools such as Copilot, Google Gemini, Blue Ocean, Hopper, and Google assistance were mentioned, representing the less commonly used tools but still assist Omani travelers in planning their travel experiences. Thus, despite the wide range of AI-Driven tools, Omani travelers still prefer to use ChatGPT compared to other AI due to the advantages it provides to travelers.
Usage of AI Tools
The findings show that participants actively used AI tools during the pre-travel phase for many reasons. The main objective of using AI tools was trip itinerary planning. One participant shared, “I used ChatGPT to plan the whole trip itinerary.” Another reason was that they used AI tools to explore new countries and potential destinations. As one participant suggested, “I used it to help choose destinations.” Another key use was for flights and hotels selection. One participant explained, “I use ChatGPT to search for tickets on different airlines, find best prices, and get recommendations.” Additionally, many participants used AI-enabled chatbots to communicate with airlines and hotels. One shared, “I used chatbots mostly to communicate quickly and effectively with the airline or hotel for my reservations.” Quick access to information was another cited reason, as highlighted by one participant “saved time by providing comprehensive information all in one place.”
Preference for Voice-Based AI Tools Over Text in Travel Planning
An important pattern identified from the interviews is the types of communication used while interacting with AI tools. Most travelers pointed out that speaking to AI tools, rather than typing, significantly enhanced their communication experience. For instance, one participant explained, “Voice communication makes things easier because I can just talk to it, and it responds through sound, so I don’t have to type,” suggesting that voice interaction reduces effort required and simplifies the process, especially when multitasking or during organizing travel plans. Specific tools such as Google Gemini and ChatGPT voice features were frequently referenced. A participant noted, “I ask Google Gemini voice questions, and it answered me.”
During Travel Use of AI Tools
Types of AI Tools
Participants shared the types of AI tools that they used during travel, including ChatGPT, Google Assistance, Perplexity.ai, facial recognition systems, Google Gemini and AI-powered robots. Among these, ChatGPT emerged as one of the most commonly used tools, similar to its popularity in the pre-travel phase. Perplexity.ai was a less commonly used tool during travel; however, it was mentioned for its ability to rapidly process and summarize information easily. As one participant stated, “I used Perplexity.ai; it summarized everything I need.” Other AI tools that were less frequently mentioned by participants included real-time language translators such as Google Translate, personalized recommendation engines like Hopper, navigation tools such as Google Maps, AI-powered photo editing apps such as Lensa and AI-powered chatbots from airlines or hotels.
Usage of AI Tools
The results show that AI tools played an important role in enhancing Omani travelers’ experiences during the travel phase. One of the most practical uses was for Real-time recommendations and local exploration, where tools such as ChatGPT, Perplexity.ai, and Hopper were used for comparing prices, exploring options and identifying the best booking applications. These tools support users in making informed decisions quickly. This was supported by the statement, “I would tell it my mood. . .and asked for halal restaurants, and it gave me recommendations.”
Furthermore, travelers utilized AI tools for real-time spoken responses and translation support. Tools such as ChatGPT’s voice chat functionality and instant translators have helped participants to overcome the language barriers. One of the participants stated, “In China, there was an audio tour guide device that translated the guide’s Chinese into our preferred language in real-time through earpieces.” Another highlighted usage was facial recognition systems for check-in. One participant described the experience: “In China, I also used a facial recognition check-in system at a hotel; it was fast, and I didn’t need any human interaction.” Additionally, AI’s ability to summarize travel information was among the key reasons individuals used these tools.
Other less frequently referenced uses of AI tools during travel included instant problem-solving and emergency help. As one participant stated, “I once found a bug in my hotel room and wasn’t sure if it was harmful. I uploaded the picture to ChatGPT, and it gave me the answer in less than 2 minutes.” Cultural and safety guidance was another use; some travelers relied on AI tools for cultural tips, dress codes, and safety advice while abroad. One participant stated, “I asked about the weather, local culture, and currency, to know how to interact with people, and ChatGPT made a plan for me.”
Usage of AI-Powered Robots
The use of AI-powered robots emerged as a unique sub-theme, as participants revealed the growing adoption of robotic systems in hotels, airports, and restaurants. The analysis revealed multiple uses of AI-powered robots. First, service delivery automation. One participant clarified, “the delivery person gives the robot the order and enters the room number, these delivery robots were able to navigate and know exactly where the room is located.” Personalized robot interaction was also referenced, with multiple respondents pointing out how robots identified them by name, gave greetings and interacted with the guests in a humanized way. One respondent explained, “After ordering, the robot notified me when my food was on its way and say goodbye after it left.”
The analysis also reveals that AI-powered robots were used for real-time support and navigation assistance, especially in airports. One participant shared, “I interacted with a robot called ‘I help you’ at an airport in Korea. I got lost, and the robot provided assistance.” Respondents commended the strong communication capabilities of robots, with one participant saying, “I felt like I was talking to a human face to face.”
Several participants referenced the role of AI-powered robots in facilitating their check-in and check-out procedures, especially in hotels. One participant mentioned that he was informed in advance that he would be served by an AI-powered robot upon arrival at the hotel. This participant noted, “I received an email from the hotel introducing a service robot check-in to ensure hassle-free arrival.” Another participant highlighted how facial recognition is integrated in service robots to make the check-in process faster, stating, “I placed my passport, and the robot took a picture of me using facial recognition. Based on that, the robot confirmed my room number.”
Theme 2: Motivational Drivers for AI Tool Adoption in Travel
This theme explores the various motivations that encouraged participants to use AI tools in travel. These motivations were categorized into three sub-themes: functional, social, and emotional drivers.
Functional Drivers
In the context of this study, functional drivers can be defined as practical and task-oriented reasons for travelers’ usage of AI tools. These drivers are linked to how AI helps travelers plan their trips and manage them more efficiently in terms of saving time, effort, cost, and performing multiple travel-related functions in one platform. The analysis revealed multiple functional drivers that encouraged travelers to adopt AI in travel-related contexts such as efficiency and time saving, ease of use, perceived value and multifunctionality.
Efficiency and Time Saving
This motivation reflects participants’ motivation to use AI tools, primarily driven by the need for speed and efficiency. Participants repeatedly emphasized quick responses, fast service delivery, and the time-saving capabilities of AI tools, especially when compared to traditional communication methods. For instance, one participant noted, “I got my itinerary in minutes,” while another added, “It gives me information in less than two minutes.” Others described the tools as “very efficient; it’s better than going to a travel agency.”
Ease of Use and Accessibility
The results revealed that most participants found AI tools easy to use and navigate, and available in many travel applications. Accessibility was frequently mentioned as a key motivator, with one participant stating, “it’s free to access any time and everywhere.” Others describe the experience as “smooth” and said it “feels like talking to a person,” pointing to the human-like interaction which eases the communication between the user and AI tools. Another factor that encouraged perception of ease of use was the voice feature, as one participant noted: “I mostly speak to the tool rather than typing, the availability and easy access encouraged me to use.”
Cost Effectiveness/Perceived Value
Many respondents highlighted the appeal of free access and unlimited usage, with statements such as “It’s free” and “Why pay a travel agent when AI can provide similar information?” indicating how AI tools are not only seen as accessible but also financially beneficial. Budget-conscious travelers particularly appreciated AI tools such as Hopper, which helped them to compare prices and choose cheaper options. When asked about the reason for using Hopper, one participant commented, “It analyzes billions of daily prices for flights and hotels, which helped me compare and choose based on my budget.” Participants also expressed how the AI tools replaced paid services like translation and travel agent services, as one participant explained, “I used it for translation purposes.”
Multifunctionality
Multifunctionality refers to participants’ motivation to use AI tools due to their ability to serve multiple purposes and offer an integrated experience. Several interviewees described AI tools as a “companion” throughout the travel journey that perform diverse tasks for them. Interviewees emphasized how AI tools function as all-in-one platforms and thus eliminating the need for multiple applications. One respondent explained, “I asked about the weather, local culture, currency, and to know how to interact with people, and ChatGPT made a plan for me.”
Social Influence and Trust Drivers
In the context of this study, social drivers refer to the influence of other people and previously gained social experiences on the decision of travelers to use AI tools. In this study, social drivers encompass recommendations from friends, family members or online communities, as well as familiarity achieved from the use of AI in everyday life. Trust in AI also proved to be a major factor in AI adoption that was primarily influenced by perceived accuracy and users’ previous experiences.
Social Influence and Peer Recommendations
Social motivation to use AI appeared from word of mouth, media influence, and social behavior. Travelers’ quotes such as “everyone uses it and recommend it” and “I saw many influencers recommending them for travel purposes,” which act as evidence that social validation and word of mouth are important factors in shaping usage behavior. Thus, social effects built a sense of trust and curiosity, especially for first time AI users. Social trust was also reflected in the form of influence from family and friends as one participant mentioned, “My wife brought it to my attention” showing that recommendation from trusted individuals encourage the AI usage.
Familiarity From Prior Use
Some participants mentioned they had previously used AI tools in non-travel contexts, such as for academic assistance or daily activities. This prior familiarity lowered the barrier to using AI for travel purposes. As one participant stated, “I’m already used to the tool, and I decided to use it for travel as well. It made everything easier for me.”
Trust
Trust was one of the frequently mentioned factors that encouraged the usage of AI tools in travel. It was influenced by the perceived accuracy of the tools, users’ previous experiences and their human-like intelligence. Participants described how reliable the tools were, with statements such as, “accurate information 90% of the time,” and “It provided assurance, for some people it’s a hassle to call.”
Experiential and Emotional Drivers
In the context of this study, experiential and emotional drivers reflect the emotional and experiential motivations that influence the engagement of travelers with AI tools, apart from the functional benefits. Based on the analysis, such drivers include enjoyment when interacting with AI, curiosity and novelty seeking, appreciation for personalized recommendations, and comfort or reduced anxiety when planning and making travel decisions.
Perceived Enjoyment
Participants expressed that interacting with AI tools in travel was enjoyable and engaging. Several participants described these tools as “easy to use” and “quick,” which enhanced their perceived enjoyment. One participant highlighted how their children enjoyed interacting with an AI-powered robot, stating, “For my kids, it was fun and an attraction point for the family.” Others commended the voice feature of AI and appreciated being able to have a voice conversation, which made the experience more enjoyable than typing or reading: “More fun to have a conversation than just reading.”
Personalized and Tailored Recommendations
Participants valued the ability of AI tools to provide personalized experiences. The tools were seen as systems that could tailor suggestions in real time. As one participant noted, “I described exactly what I want, and it gave me ideas.” The analysis further highlights that many participants appreciated the adaptive nature of AI, with many recognizing how these tools adjusted to their mood, needs, and trip preferences. As one participant stated, “I would tell it my mood and the kind of vibes I wanted and gave it a few destination ideas. We basically had a conversation.”
Curiosity and Novelty-Seeking
A few participants described engaging with AI as a way to enjoy new technologies and explore digital interactions that differ from traditional tourism services. One participant recalled his experience with an AI-powered robot: “I was amazed by the advancements there. One thing I remember was an AI-powered robot in the lobby that would deliver food to your room. When it arrived, it greeted me by name and said goodbye after it left.”
Comfort and Reduced Anxiety
AI tools were perceived to be helpful for reducing stress and anxiety throughout the travel journey. Many participants referred to automated check-in, facial recognition, seamless booking opportunities and avoiding manual processes, and remarked that these experiences contributed to travel being more seamless and less stressful; as one participant described, “In China, I also used a facial recognition check-in system at a hotel. It was very fast, and I didn’t need any human interaction to complete the process. This helped avoid the awkwardness of having to explain things with hand gestures.”
Theme 3: Challenges to Effective Use of AI Tools in Travel
This theme presents the different challenges that negatively affected the perception of AI tools in travel. Overall, participants raised concerns related to the quality and reliability of information generated with AI, limitations in personalization and functionality, technical and communication challenges, and other challenges related to privacy, trust, and lack of human interaction.
Generic and Impersonal Recommendations
Many participants felt frustrated with their communication with AI tools, describing it as generic, and giving information that had no relevance to their personal travel needs and circumstances. Many participants described frustrations with lack of personalized recommendations. One participant stated, “The responses weren’t related to the country I was in,” while another mentioned, “It gave me a list of random attractions that didn’t suit my family and kids.”
Inaccurate Information and Verification Efforts
Several participants reported receiving inaccurate opening hours, outdated information, or incorrect locations. For instance, one participant noted, “ChatGPT told me a place was open, but we drove there and found it closed, it wasted our time.” Another shared “The app said the restaurant had halal food, but when I got there, they didn’t know what halal was.” Moreover, some participants expressed dissatisfaction with the effort required to verify the accuracy of the information. As one participant explained, “It’s also not always accurate, and I needed to double-check, especially if I was going to spend money on something.”
Communication Barriers and Input Misunderstanding
Several respondents highlighted communication barriers, reporting that the AI tool “took words too literally” and often misunderstood their intent, particularly when the user input few details. A few participants shared that they had to rephrase their questions multiple times to get to the desired answer; otherwise, the AI tool would provide incomplete information. As one participant stated, “sometimes I had to ask the same things in different ways until I got a useful answer.” Many participants also faced challenges with translation, where AI tools misunderstood the input or responded with incorrect phrases, especially when dealing with less commonly used expressions. As one participant noted, “It translated word for word, not the real meaning.” Furthermore, some participants noted that some suggested activities did not match the travel season, indicating a lack of contextual awareness in AI features. One participant highlighted that issue by stating, “It recommended parks, but it was winter.”
Integration and Functionality Gaps
While AI tools were helpful for research and getting recommendations, many participants reported that they didn’t really support completing tasks such as booking tickets or making reservations. As one participant stated, “It couldn’t help me to book the tickets,” and another noted, “It doesn’t book for you and doesn’t fully accommodate your needs, so it requires extra work.” Others mentioned that AI tools are inflexible and have limitations in functionality. For example, one participant stated, “The hotel robot couldn’t scan my child’s passport, and staff had to help,” while another said, “It didn’t recognize my ID.”
Technical and Operational Challenges
Technical issues such as weak internet connections and glitches were mentioned as challenges to receiving quick answers. One participant said, “The internet connection was weak sometimes, which made the responses slow and unusable.” Facial recognition errors were also reported when processing check-in information, with another sharing, “Some of my family members had trouble with facial recognition, so they had to check-in manually.” Additionally, some users faced language barriers while communicating with robots, especially in places where English was not supported. As one participant noted, “There was a language challenge because some robots didn’t speak English.”
Privacy and Safety Concerns
A degree of hesitation was noted among respondents when it came to sharing personal information with AI tools and robotics. As one participant stated, “I faced challenges from a safety perspective; I was hesitant to share my personal details.” The findings highlight the need for improvements in data privacy assurance and how AI systems handle users’ travel information.
Lack of Human Interaction
Some of the participants quoted lack of human interaction as one of the limitations especially when dealing with AI-powered robots, as one participant stated, “sometimes you travel and interact with people and learn about their culture, which doesn’t happen with a robot” and “I missed the human touch while checking in,” another participant recommended do have a balance between human touch and technology, “I think a mix between AI and human touch is best.”
Theme 4: Travelers’ Recommendations for Improving AI in Travel
This theme includes recommendations from travelers on how to improve AI tools in travel. Overall, participants highlighted a need for improved personalization and context knowledge, broader and more integrated functionality, and improved communication and language functionality. They also demonstrated the need for better accessibility, inclusive design, improved ethical standards and transparency, increased accuracy and reliability, and integration of visual elements to support decision-making along the travel journey.
Better Personalization and Contextual Understanding
A large percentage of the respondents expressed strong desires for AI tools that could provide more personalized responses that align with their preferences, moods and personalities. Accordingly, one participant shared, “AI tools should personalize the experience and provide more sufficient information.” Several participants wanted AI tools to learn from their past behavior, with one suggesting, “suggest things that fit the user’s personality and travel preferences.” Seasonal tips were another idea raised, with one respondent suggesting that “specifying seasons when recommending activities” would ensure more relevance.
Comprehensive Functionality
Participants view AI tools as more than just a way to get information; they imagine them as full-service travel assistants capable of managing bookings and transactions. This desire was clear in comments like, “I want everything related to travel in one place so that it doesn’t get messy.” Real-time updates, local tips, and faster information delivery were also recommended. For example, one participant suggested, “I suggest that each country develop its own tourism focused Chatbot with real time, accurate information, like something called Chatman.”
Enhanced Communication Capabilities
A recurring request across interviews was the need for improved communication features in AI travel tools. The ability to communicate with different languages, especially when dealing with AI-powered robots, was seen as important for improving communication between travelers and AI tools. One participant noted, “The robot should be able to speak different languages and allow me to choose the language. Also, the robot’s speed could be improved.” Language adaptation was also important. Respondents suggested improvements, such as the ability to understand different accents and incorporate slang to avoid misunderstandings. This desire was emphasized by one respondent recommending, “It would be helpful to add some enhancements regarding language such as adding different slangs to avoid misunderstandings.” The integration of real-time translation emerged as an important recommendation, with one participant highlighting how this feature could break linguistic barriers and help destinations attract new tourist markets from markets where language differences often pose challenges “There was also an audio tour guide device that translated the tour guide’s Chinese sentences into our preferred language in real-time through earpieces. This tool is very convenient, especially for tourism, as it breaks language barriers.”
Improved Accessibility and Inclusive Design
Responses highlighted the need for AI tools to be accessible to all users including those with special needs. Several recommendations focused on inclusive design and offline functionality to ensure these tools remain useful in diverse travel scenarios. One user mentioned “I think AI delivery robot services should be available for people with special needs, not just for regular people,” suggesting the need for inclusive design in service robotics and user interfaces. Multiple respondents stressed the need for offline-capable AI tools, noting that internet availability is limited during travel. As one participant explained, “these tools should have offline functionality. Sometimes, during travel, you don’t always have internet access.”
Ethical Considerations and Transparency
The findings indicate an expectation that AI tools should operate with ethical boundaries and cultural sensitivity differences. Users prefer AI tools to “Consider ethics while providing information.” These reflections point to the need for AI to understand the user’s culture and beliefs, making the communication experience feel more culturally respectful. Participants also wanted more transparency, with one suggesting “it would be helpful if the tools always showed the source of the information without needing to ask.”
Enhanced Accuracy and Reliability
Participants emphasized the need for AI tools to provide more accurate and reliable information. The issue of accuracy emerged as a recurring challenge. To improve accuracy, several respondents recommended structural changes, including implementing more targeted query mechanisms that segment and categorize travel inquiries to generate more precise responses. As one participant shared, “I think the tool should be divided into multiple prompts based on what I’m looking for. This would help generate more accurate answers.”
Integration of Visual Elements
Participants emphasized the value of incorporating visual features, such as images and maps (locations), to improve the functionality and user experience of AI tools. A few participants noted that visual elements would be especially beneficial in supporting travelers’ decisions, particularly when searching for information about unfamiliar destinations. As one participant suggested, “AI tools should provide pictures and exact location when recommending activities.” Another noted that it is “better to have pictures along with recommendations to visualize the options.”
Theme 5: Satisfaction and Continuance Usage Intention
Despite the challenges and the recommendations for improvement, almost all travelers expressed their general satisfaction with AI tools and their willingness to continue using them in travel. Participants highlighted many reasons for their continuance usage intention. For example, some viewed AI tools as more efficient than traditional search engines. As one participant explained, “Using AI tools now has become better in preparing for a trip as it gives me direct results of what I want instead of using Google.” Participants also indicated that AI tools help them avoid cultural misunderstandings. As one participant noted, “I will continue to use AI because these tools can help me understand what things I am supposed to avoid doing. . . so that I don’t offend the local communities and instead embrace and appreciate their cultures.” Others mentioned that they will continue to use AI tools because of their ability to provide a comprehensive itinerary for the trip, identify top attractions, offer local tips, recommend better offers/ deals and provide real-time assistance and instant translation.
Discussions and Implications
A Theoretical Model Based on Travelers’ Motivations, Adoption, and Outcomes (MAO)
Based on the findings, a theoretical model was developed to summarize and discuss the results and propose relationships between the emerging themes. Accordingly, we introduce the Motivation–Adoption–Outcome (MAO) model, a theoretical framework that explains AI adoption in travel as a process shaped by travelers’ motivations, adoption behaviors, and post-adoption outcomes (See Figure 3). In doing so, the MAO model enriches traditional technology adoption frameworks by extending their focus beyond functional adoption drivers to incorporate travelers’ experiential and social motivations and post-adoption outcomes. Moreover, it is also extended to consider challenges that hinder adoption through the lens of IRT (Ram, 1987; Ram & Sheth, 1989) and to reflect the specific context of the study, which is AI tools’ usage in travel. The proposed MAO model consists of four groups of factors: (1) travelers’ motivations to adopt AI tools, (2) types of AI tools adopted and their usage, (3) challenges to effective use and required improvements, and (4) post-adoption outcomes.
A theoretical model for AI motivations, adoption, and outcomes in travel.
The first group of factors, travelers’ motivations to adopt AI tools, refers to the different reasons that encourage travelers to adopt AI tools either during the planning stage or while traveling. Three categories of factors are identified under this group: functional, social and experiential and emotional drivers. The first category of factors includes four motivators under functional drivers: efficiency and time saving, ease of use and accessibility, cost effectiveness, and multifunctionality. These align with similar constructs in UTAUT2, including performance expectancy, effort expectancy, and price/value (Venkatesh et al., 2012). Efficiency and time saving, along with ease of use and accessibility, also correspond to perceived usefulness and perceived ease of use in TAM (Davis, 1989). However, within the functional drivers, multifunctionality emerged as a new theme that is not addressed in previous technology adoption theories. This refers to the ability of AI tools to perform multiple functions and provide an integrated experience. For example, AI tools were used for translation, real-time assistance, tour guidance, and to provide information on weather, local culture, and currency, which encouraged people to use them (Li et al., 2025).
The second category of factors under this group that motivated travelers to adopt AI tools during travel includes three motivators related to social drivers: social influence, familiarity from prior use, and trust. Social influence is a well-established motivator in technology adoption (Venkatesh et al., 2012), and in the context of AI usage in travel, it refers to the influence of others such as family and friends to adopt these tools. Familiarity from prior use, which corresponds to “habit” in the UTAUT2, explains how individuals’ use of AI in other contexts such as work or daily life encourages them to continue using it in travel-related contexts. Moreover, trust emerged as a distinct motivator for AI usage in travel. Previous studies (e.g., Sarker et al., 2019; Vimalkumar et al., 2021) have highlighted its importance in reducing perceived risks and increasing confidence in the reliability of technology.
The MAO model introduces a third category of motivators related to experiential and emotional drivers, including perceived enjoyment, personalized recommendations, curiosity and novelty-seeking, and comfort and reduced anxiety. Except for perceived enjoyment, emotional and experiential drivers are often ignored in traditional technology adoption models such as TAM (Davis, 1989) and UTAUT2 (Venkatesh et al., 2012), which primarily focus on functional benefits and social influence. In contrast, the current study introduces additional affective dimensions of user experience which are particularly important in the context of travel and tourism (Miao et al., 2025). For example, the ability of AI tools to provide travelers with personalized recommendations was highly valued by participants and has been supported in prior research. Sadiq et al. (2025) found that AI-powered personalized product recommendations positively impact consumers’ behavioral intention to participate in social commerce. Likewise, intrinsic motivators such as curiosity and novelty-seeking inspire travelers to explore and experiment with AI tools. This aligns with consumer behavior theories, which suggest that novelty-seeking enhances openness to adopting new technologies (Hirschman, 1980; Kashdan et al., 2004). Thus, it is argued that in the context of travel, AI adoption is driven by a combination of functional, social, and experiential-emotional drivers; therefore, we present the following proposition:
The second group of factors in the MAO model relates to the types of AI technology used and their application and includes two types. The first type includes the AI tools used in the pre-travel and travel phases. During both travel phases, the most common tool was ChatGPT (Sigala, 2024). The second group refers to how AI technology was used. In the pre-travel phase, the AI tools were largely used for information gathering, exploring destinations, building travel itineraries, and communication to support travel. These findings are consistent with earlier research highlighting the role of AI in supporting pre-travel activities (Li et al., 2025; Xiang et al., 2015). In the during-travel phase, AI tools are commonly used for real-time navigation, translation, transportation, safety guidance, and personalized recommendations. These findings align with previous studies showing how AI tools support travelers during their trips (Ali et al., 2023; Li et al., 2025; Rashid & Kausik, 2024; Wei, 2022). In the developed MAO model, it is argued that the adoption of AI tools during the pre-travel and travel stages will influence travelers’ satisfaction; this satisfaction level will differ based on the ability of the adopted AI tools to meet the travelers’ use expectations, such as information access, itinerary planning, booking, helpful cultural tips (Bhattacherjee, 2001; Huang et al., 2024). Based on this, we present the second proposition:
The findings of the study showed that customers who were satisfied with AI tools planned to continue using them in the future; therefore, this satisfaction– continuance usage intention relationship was depicted in the MAO model. Nearly all participants expressed satisfaction with AI tools and a strong intention to reuse them in future travel contexts. This finding supports the confirmation-satisfaction loop suggested by the ECM (Bhattacherjee, 2001) and reinforces the role of key predictors in TAM and UTAUT2 (Davis, 1989; Venkatesh et al., 2012), which are extended in the current study by introducing additional factors such as multifunctionality and emotional drivers. Therefore, in our study, we propose that creating satisfactory experiences with AI tools that meet travelers’ functional, social, and experiential needs will encourage continuance usage of these tools in future interactions. Accordingly, the following proposition is presented:
Besides motivations and adoption behaviors, the MAO model introduces a third category of factors, the “moderating factors” that characterize the adoption - continuance usage relationship. These moderators fit into two broader categories: those that negatively impact the relationship, called challenges to effective usage, and those that positively improve it, called improved AI capabilities. These two groups are interrelated, because often the challenges users face highlight aspects of the AI tool that need improvements. Therefore, we will discuss both groups together. Focusing on the challenges of adoption addresses a significant gap in current research on AI in travel, which is characterized by pro-innovation bias that emphasizes adoption drivers while giving limited attention to travelers’ concerns, hesitation, or resistance (Seyfi et al., 2026; Seyfi, Gorji, et al., 2025; Seyfi, Kim, et al., 2025).
One particular challenge that was referred to by many participants was generic and impersonal recommendations. Melián-González et al. (2021) discussed this challenge and noted that the lack of contextual awareness can negatively affect users’ trust and satisfaction. In response to this challenge, participants noted the requirement for AI tools to become more contextualized and personalized (Sadiq et al., 2025). Another challenge concerning the incorrectness of information generated with an AI tool was also reported (Carvalho & Ivanov, 2024; J. H. Kim et al., 2025). This concern aligns with findings by Seyfi, Gorji, et al. (2025), who found that travelers may perceive AI-generated travel advice as untrustworthy or lacking authenticity. In response, participants recommended that AI tools should be made more accurate and reliable by using real-time and verified data sources (J. H. Kim et al., 2025).
In addition to the challenges discussed earlier, participants also highlighted several other challenges including the lack of ability to complete bookings or transactions, communication barriers, technical and operational problems such as slow responses due to weak internet or errors in facial recognition, privacy and safety concerns and missing the human interaction. This result aligns with the work of Seyfi et al. (2026) who found that many tourists experience functional and psychological barriers when considering generative AI for travel planning, including concerns about usage difficulties, perceived risks, and conflicts with cultural traditions. To address these challenges, participants recommended expanding the functionality of AI tools to handle transactions beyond information search, enhanced communication capabilities, improving the technical reliability and adding offline features that work without internet access, improving data privacy policies, and creating a better balance between AI and human interaction (Gill et al., 2022; Shuqair et al., 2024). Based on the above discussions, the following propositions are presented:
Theoretical Implications
This research provides significant theoretical implications. First, this study introduces the Motivation–Adoption–Outcome (MAO) model as a comprehensive framework for understanding how travelers interact with AI tools. The MAO model builds on and expands traditional technology adoption models. In doing so, the MAO model reconceptualizes technology adoption in the context of AI in travel as a dynamic process rather than a static decision, shaped by interactive motivations, adoption behaviors, post-adoption evaluations, and contextual conditions. The MAO model represents the adoption of AI tools in travel as a process that evolves from motivations to adopt AI technology, to the act of adoption, and finally to the outcomes of that adoption.
Secondly, the MAO model also expands traditional technology adoption frameworks by incorporating new motivational aspects that are often overlooked. Existing models emphasize functional aspects such as performance and effort, while the MAO model highlights emotional, trust-based, and experiential motivations that are particularly relevant in tourism contexts. For example, travelers in this study indicated that their motivations were based not only on the efficiency or usability of AI tools, but also on emotions—such as curiosity, comfort, and the desire to avoid stress. Many participants also emphasized trust related to information accuracy, data privacy, and confidence that AI systems would respond in ways that meet their needs. By explicitly theorizing these functional, social, and experiential–emotional drivers as antecedents of AI adoption, the MAO model provides a richer explanation of why travelers adopt AI tools in tourism settings, rather than simply identifying functional motivations. Thus, future research examining AI adoption in tourism should place greater emphasis on travelers’ emotions, trust, and experiential motivations in addition to task performance.
Third, the MAO model extends the Expectation Confirmation Model (ECM) by highlighting the complexity of the relationship between satisfaction and continuance usage intention. While ECM suggests that satisfaction leads to continuance use when expectations are confirmed, this study demonstrates that this relationship can be influenced by both positive and negative contextual factors. Specifically, the MAO model incorporates Innovation Resistance Theory (IRT; Ram, 1987; Ram & Sheth, 1989) to explain the role of perceived challenges as boundary conditions that shape the relationship between adoption, satisfaction, and continuance usage intention. On the negative side, challenges such as impersonal or autonomous recommendations, inaccurate information, excessive verification requirements, technical issues, communication problems, privacy concerns, and the lack of human interaction were identified as factors that may reduce satisfaction and continuance usage intention. These challenges may also moderate the relationship between satisfaction and continuance usage intention, such that the relationship becomes weaker when these perceived barriers are high and stronger when they are low. This discussion of perceived challenges directly responds to recent critiques of pro-innovation bias in AI adoption research in tourism, which argue that resistance arising from perceived challenges is an integral part of understanding AI-related behaviors (Seyfi et al., 2026; Seyfi, Gorji, et al., 2025; Seyfi, Kim, et al., 2025).
On the positive side, improvements in AI capabilities—such as personalization, contextual understanding, comprehensive functionality, improved communication, accessibility, inclusive design, accuracy, and transparency—can increase satisfaction and consequently strengthen continuance usage intention. These improvements may also strengthen the relationship between satisfaction and continuance usage intention, meaning that the relationship becomes stronger when users perceive these improvements to be high and weaker when they are low. This post-adoption dynamic is increasingly acknowledged in recent research. For example, Majid et al. (2025) illustrate the potential of generative conversational AI, particularly chatbots, to sustain pro-environmental behaviors among tourists by delivering timely nudges that encourage responsible travel practices.
Practical Implications
The findings of the study offer several practical implications for AI technology developers of AI and tourism service providers seeking to enhance travelers’ experiences.
For developers of AI technologies: first, they should design AI tools in a way that enhances their ability to recognize diverse cultural backgrounds, preferences, and the contextual settings of users. This means that developers should build algorithms that don’t just provide general recommendations for users but deliver personalized and contextualized suggestions. To do that, the design should integrate features that account for user preferences, behavior, real-time location data, travel season, local customs and cultural tips. This will ensure that AI tools are able to adapt their output to fit specific user requirements instead of offering one-size-fits-all outputs. Additionally, to improve accessibility in various travel settings, AI tools should be designed to operate without an internet connection, especially in locations where the access to the internet is limited.
Second, the findings of the study show that voice and visual features contribute to comfort and ease of use, especially while traveling contending that voice assistance and real-time information allow travelers to engage with services rapidly (without having to type). However, some participants were frustrated that AI tools could only provide them with suggestions and could not then perform tasks such as bookings or transactions, developers should be working toward further integrated, end-to-end functionality that allows users to take action on user suggestions without having to switch tabs or open new apps. Enhancing sensory support systems for travelers, for example, providing real-time visual translation services, and multilingual voice responses, also contributes to faster, more accessible and satisfying experiences for different profiles of travelers.
Third, AI developers need to recognize that travelers are engaging with the AI tool not only for functional use, but also for emotional and experiential value. The findings show that users like AI tools because they are enjoyable, provide a sense of curiosity, and a comforting feeling when they are away from home. Finally, the anxiety shared by participants regarding sharing their own personal data with AI tools emerged as an important consideration, highlighting the need for transparency in data-use policy, stronger privacy precautions, and addressing users’ reluctance to share personal data. AI developers and service providers need to describe how users’ data will be captured, stored, and utilized.
Fourth, it is critical for tourism service providers, like hotels and airlines, to consider how to implement AI tools and robotics in service delivery, not only to enhance productivity, efficiency and profitability for the providers, but in a way that supports the overall service experience. In order to do this, providers must use the technologies in a balanced way within existing systems. Additionally, tourism practitioners should be trained to provide assistance with the AI tools when technical issues arise, or the systems are not functioning. More importantly, when the traveler is unfamiliar with technology, human assistance remains important in order to ensure a smoother and reassuring service experience for the traveler.
Finally, despite the limitations of current tools, many travelers expressed a clear intention to continue using AI tools in future travel, driven by convenience, speed, and previous positive experiences. To build on this momentum, service providers must focus on maintaining simple, fast, and reliable user experiences. Enabling travelers to derive consistent value from AI tools can encourage customer loyalty, positive word-of-mouth, and reduced dependence on traditional services.
Limitations and Directions for Future Research
While this study provides important insights into travelers’ adoption and use of AI tools through the MAO model, several limitations should be acknowledged, offering opportunities for further research. First, the qualitative nature of the study, though rich in depth and contextual understanding, limits the generalizability of the findings to broader populations. Based on the themes identified in this qualitative study, future studies could create a structured survey instrument containing closed-ended items and Likert scales to test and validate the MAO model quantitatively. Moreover, considering that the study was conducted with 50 Omani travelers, the results may not fully reflect the opinions of travelers from other parts of the world or backgrounds. Expanding the scope of the participants would give more comprehensive insights into traveler-AI interactions.
Second, the study primarily focused on the pre-travel and during-travel phases, where AI tools play a significant role in planning and real-time experiences. However, the post-travel phase remains underexplored. Future studies could examine how travelers reflect on their experiences with AI after their trip, and how these reflections influence long-term trust, satisfaction, and continuance usage intention. Third, the current study did not investigate specific demographic differences, such as age, digital literacy, or travel experience, in depth. These factors may influence how travelers perceive and use AI tools. Future research could explore these dimensions to tailor AI design more effectively to diverse user segments. Finally, this study contributes to existing theory by framing AI adoption as a multi-phase process rather than a single-point decision. Therefore, the current study calls for future longitudinal research to track how motivations, usage behaviors, and outcomes evolve from planning to travel, and eventually to post-trip reflection.
Footnotes
Acknowledgements
The authors acknowledge that they used ChatGPT (version 5.2) to assist with grammar checking and proofreading of this manuscript. All content was reviewed by the authors, who take full responsibility for the accuracy and integrity of the content of this publication.’
Ethical Considerations
In accordance with university policy, research involving data collection through surveys and interviews is classified as Level 0 and therefore requires approval from the Head of Department. This approval was obtained prior to the commencement of data collection.
Consent to Participate
During the data collection process, all ethical guidelines were strictly followed. Participants were informed about the purpose of the research, and their informed verbal consent was obtained before any interviews were conducted or recorded.’
Author Contributions
Osman El-Said: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing. Khawla Al-Naamani: Conceptualization; Data curation; Formal analysis; Methodology; Visualization; Writing—original draft.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Open access publication fees were supported by the German University of Technology in Oman (GUtech) through its institutional open access funding scheme.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Data will be available upon request from the corresponding author.*