Integrating Generative AI and IoT for Sustainable Smart Tourism Destinations

Abstract

This paper aims to develop a groundbreaking approach to fostering inclusive smart tourism destinations by integrating generative artificial intelligence (Gen AI) with natural language processing (NLP) and the Internet of Things (IoT) into an intelligent platform that supports tourism decision making and travel planning in smart tourism destinations. The acquisition of this new technology was conducted using Agile methodology through requirements analysis, system architecture analysis and design, implementation, and user evaluation. The results revealed that the synergistic combination of these technologies was organized into three tiers. The system provides information, including place names, images, descriptive text, and an audio option for users to listen to the information, supporting tourists with disabilities. Employing advanced AI algorithms alongside NLP, developed systems capable of generating predictive analytics, personalized recommendations, and conducting real-time, multilingual communication with tourists. This system was implemented and evaluated in Suphan Buri and Ayutthaya, UNESCO World Heritage sites in Thailand, with 416 users participating. The results showed that system satisfaction was influenced by (1) the tourism experience, (2) tourism planning and during-trip factors (attention, interest, and usage), and (3) emotion. The relative Chi-square (χ²/df) of 1.154 indicated that the model was suitable. The Comparative Fit Index (CFI) was 0.990, the Goodness-of-Fit Index (GFI) was 0.965, and the model based on the research hypothesis was consistent with the empirical data. This paper contributions significant advancements in the field of smart tourism by demonstrating the integration of Gen AI, NLP, and the IoT and offering practical solutions and theoretical insights that enhance accessibility, personalization, and environmental sustainability in tourism.

1. Introduction

The tourism industry stands at the brink of a transformative revolution, driven by the rapid advancement and integration of digital technologies. Among these, generative artificial intelligence (Gen AI), natural language processing (NLP), and the Internet of Things (IoT) are proving to be particularly influential [1]. The convergence of these technologies has the potential to redefine the parameters of tourism, making it more accessible, personalized, and environmentally sustainable [1,2]. Gen AI represents a significant technological advancement within the tourism sector, offering new ways to enhance customer experiences and operational efficiency. The applications and impacts of Gen AI in tourism highlights its role in content creation, personalized customer service, and predictive analytics [3].

Gen AI has been increasingly utilized for content creation in the tourism industry, enabling the generation of customized itineraries, promotional materials, and even virtual tourism experiences [3,4]. AI-driven content generation tools can create detailed and personalized travel guides based on user preferences, significantly enhancing the pre-trip planning experience [5]. Furthermore, Gen AI contributes to the development of realistic virtual reality (VR) environments that simulate travel destinations, allowing tourists to experience places virtually before deciding to visit them in person [6].

AI technologies, particularly those using NLP, have transformed customer service interactions. Chatbots and virtual assistants powered by Gen AI are capable of handling inquiries and providing recommendations in real-time, offering a level of personalization that was previously unattainable [7]. These AI-enhanced interfaces not only respond to customer needs around the clock but also learn from interactions to improve their accuracy and helpfulness over time [8].

Gen AI plays a crucial role in predictive analytics within the tourism industry. By analyzing large datasets, AI systems can forecast trends, demand patterns, and customer behaviors with high accuracy [9,10,11]. This is related to the study of Buhalis and Leung [9], who highlight the transformative potential of AI in tourism, particularly in enhancing predictive capabilities [9]. They emphasize that AI can analyze multiple data sources to provide actionable insights, which can be used for strategic planning and improving customer experiences [1,9]. Moreover, the capability allows tourism providers to optimize their offerings and enhance their operational efficiency, leading to improved customer satisfaction and increased profitability [9]. Predictive analytics also aid in risk management, helping companies anticipate and mitigate potential disruptions in travel plans [9,10]. Moreover, the explores how AI-driven customer behavior analysis can lead to personalized tourism experiences [11]. The research indicates that understanding customer preferences through AI can significantly enhance satisfaction and loyalty.

Despite the benefits, the deployment of Gen AI in tourism raises several ethical and practical concerns. Issues of privacy, data security, and the potential loss of jobs to automation are significant challenges that need addressing [11]. Moreover, there is a risk that reliance on AI could lead to homogenized experiences that lack the personal touch traditionally associated with hospitality and tourism services [12].

NLP is transforming the tourism industry by enhancing communication, customer service and personalization. NLP powers chatbots and virtual assistants that handle multiple inquiries simultaneously, offering continuous service and improving from past interactions [7,13,14]. Additionally, NLP personalizes travel recommendations based on user data, improving experiences and boosting company conversion rates [5,15]. It also overcomes language barriers with real-time translation and multilingual support, expanding the market reach [16]. Furthermore, NLP analyzes tourist feedback for sentiment, aiding companies in understanding customer satisfaction and guiding marketing strategies [17,18,19].

Gen AI and NLP are revolutionizing tourism by enabling real-time, context-aware, and multilingual interactions that overcome language barriers, significantly aiding non-native speakers in foreign settings [20]. Integrated with IoT technology, which employs a network of sensors to monitor and adapt to environmental conditions and tourist behaviors, these technologies optimize resource use and enhance visitor experiences [21,22]. Together, they address challenges in the tourism industry by providing tailored services that dynamically cater to individual preferences, benefiting inclusive tourism by supporting tourists with disabilities and ensuring access to culturally sensitive and environmentally conscious information and experiences [23]. This synergy also advances sustainable tourism, promoting resource efficiency and environmental conservation [23,24].

The integration of generative AI and IoT technologies addresses the critical needs and gaps in the current tourism landscape. By providing personalized, efficient, and sustainable solutions, these technologies enhance the overall tourist experience, improve operational efficiency, and promote environmental and social responsibility. The literature underscores the transformative potential of Gen AI and the IoT in creating sustainable smart tourism destinations, paving the way for future innovations and improvements in the industry.

IoT technology in the tourism sector is transforming visitor experiences through enhanced interconnectivity and smart automation. Key studies highlight IoT’s role in monitoring environmental conditions and tourist behaviors, optimizing resource use, and customizing services to improve efficiency and satisfaction [19]. Furthermore, Bernardo et al. [23] studied new tourism services and management attributes, particularly those not commonly included in most conceptual frameworks. These include innovative technical applications and gamification activities, such as virtual reality (VR), to enhance the tourism experience.

Perera et al. [22] discuss IoT’s real-time data-driven capabilities, while Boes et al. [25] focus on its contributions to sustainable tourism practices and resource management. Gretzel et al. [24] emphasize IoT’s potential to create inclusive experiences for all tourists, including those with physical or sensory impairments, ensuring accessibility to culturally sensitive and environmentally conscious tourism. Moreover, a recent study by Sousa et al. [26] demonstrated that the inclusion of immersive virtual experiences for tourism purposes significantly increases tourists’ satisfaction during their stay at destinations. These insights demonstrate IoT’s significant potential to enhance operational efficiency, sustainability, and inclusivity in tourism [27].

1.1. Problem of the Study

The integration of Gen AI and the IoT in the tourism industry holds transformative potential for enhancing the inclusivity and accessibility of smart tourism. Despite significant advancements in these technologies, several key challenges and gaps in knowledge remain that must be addressed to fully leverage their capabilities for creating inclusive tourism experiences. Specifically, the research problem centers on understanding how the synergistic use of Gen AI and the IoT can effectively meet the diverse needs of all tourists, including those with disabilities, and how these technologies can be implemented to ensure ethical, sustainable, and universally accessible tourism practices [28,29].

Firstly, there is a need to identify the specific capabilities and limitations of Gen AI and the IoT in addressing the unique requirements of various tourist demographics [30]. While these technologies promise enhanced personalization and real-time data processing, their application in real-world tourism settings often encounters practical challenges such as data privacy concerns, technological reliability, and the digital divide that may exclude non-tech-savvy tourists [5,30].

Secondly, the integration of these technologies must be examined for its impact on the sustainability of tourism practices. Although IoT devices and Gen AI can optimize resource use and improve operational efficiency, their deployment must be scrutinized for potential environmental impacts, such as increased electronic waste and energy consumption [1,24,31,32].

Lastly, there is an urgent need to explore the ethical implications of deploying Gen AI and the IoT in tourism. Issues such as surveillance, data security, and the potential for bias in AI algorithms pose significant risks that could undermine trust in tourism providers and affect the overall tourist experience [33].

1.2. Contribution

This paper makes significant advancements in the field of smart tourism by demonstrating how the integration of Gen AI, NLP, and the IoT can redefine the tourism industry. The contributions of this research are substantial, offering practical solutions and theoretical insights that enhance accessibility, personalization, and environmental sustainability in tourism.

A major contribution of this research is the successful integration of Gen AI and the IoT to create a comprehensive system that enhances the tourism experience. By employing advanced AI algorithms alongside NLP systems capable of delivering predictive analytics, personalized recommendations, and real-time, multilingual communications, which are critical for modern tourism practices [34]. This integration significantly improves service personalization and operational efficiency, offering a model for future applications in the field [35].

This paper contributes to the focus on the enhancement of accessibility and inclusivity in tourism services. The application of these technologies ensures that tailored services are provided to diverse populations, including individuals with disabilities, thereby making tourism more inclusive [36]. This approach not only broadens the accessibility of tourism activities but also promotes an equitable tourism ecosystem [37]. The paper outlines a scalable model for employing AI, NLP, and the IoT that can be adapted beyond the tourism industry to other sectors aiming to enhance customer satisfaction while balancing environmental and social responsibilities [7]. This cross-industry applicability is a significant contribution, as it provides a roadmap for integrating these technologies in various contexts to achieve similar goals. Overall, this paper enriches the understanding of smart tourism technologies by showcasing their potential to transform the tourism industry into a more adaptive, inclusive, and sustainable sector. The research findings not only contribute to academic knowledge but also offer practical insights that can be implemented by tourism industry practitioners [37,38,39].

1.3. Objective

The objective of this paper is to demonstrate how the integration of Gen AI, NLP, and the IoT can substantially improve tourism planning for individuals with disabilities, thereby fostering greater inclusivity in smart tourism destinations. This research intends to achieve specific objectives: to explore and validate how the synergistic application of AI, NLP, and the IoT can enhance the accessibility and inclusiveness of tourism, particularly for individuals with disabilities. This entails developing and testing systems that offer personalized recommendations, real-time communication, and other services tailored to the needs of diverse tourists, ensuring that the benefits of smart tourism planning are distributed widely and equitably. Additionally, the research seeks to elevate the overall tourist experience through enhanced personalization and real-time engagement.

2. Literature Review

2.1. Generative Artificial Intelligence (Gen AI)

Gen AI refers to AI systems capable of autonomously producing new content like text, images, audio, and video [40,41]. It plays a crucial role in various sectors, with the market projected to grow significantly by 2026 [42]. Gen AI operates by creating open-ended systems through a combination of post-structuralist theories and neo-cybernetic mechanisms, utilizing both bottom–up and top–down approaches for system development [43]. This technology is instrumental in enhancing content creation within the metaverse, revolutionizing search experiences, and driving industry innovation and evolution [44,45].

2.1.1. Gen AI Component

Gen AI encompasses a range of technologies designed to generate new content, from text and images to music and code, based on learned patterns and data. The components of Gen AI are critical to understanding its capabilities and limitations. Neural networks and deep learning the backbone of Gen AI lies in neural networks, especially deep neural networks (DNNs). DNNs are capable of learning complex patterns in large amounts of data, making them ideal for generating new content that resembles the training data [46,47]. LeCun, Bengio, and Hinton [48] highlight the transformative potential of deep learning in AI, emphasizing its role in advancing the capabilities of generative models [48].

• Generative adversarial networks (GANs): Introduced by Goodfellow et al. [46], GANs are perhaps the most popular generative model. GANs consist of two neural networks—the generator and the discriminator—that compete against each other, thus improving the quality of the generated outputs [49]. The generator produces data, while the discriminator evaluates them against the real data [50]. This framework has been pivotal for advancements in realistic image generation [47,51].
• Variational autoencoders (VAEs): Developed by Kingma and Welling [52], VAEs are another key component of generative AI. VAEs are designed to compress data into a lower-dimensional space and then reconstruct them back into the original space, facilitating the generation of new data points. They are particularly known for their effectiveness in generating complex data structures like images [52].
• Transformer models: Recent years have seen a rise in the transformer models introduced by Vaswani et al. [50], which have revolutionized the field of natural language processing. Transformers use mechanisms like self-attention to process sequences of data, making them highly effective for generating coherent and contextually relevant text. They are the foundation of models like GPT (Generative Pre-trained Transformer) and BERT (Bidirectional Encoder Representations from Transformers) [50,53].

2.1.2. Gen AI Application

Gen AI employs sophisticated algorithms to learn patterns and create new content, including text, images, sounds, videos, and code [54]. Notable Gen AI tools such as ChatGPT, Bard, Stable Diffusion, and Dall-E demonstrate the technology’s capability to handle complex prompts and generate outputs that closely mimic human-like qualities [55]. This has sparked significant interest and research into Gen AI applications across various sectors, including healthcare, medicine, education, media, and tourism [55,56].

The idea that Gen AI can play a pivotal role in enhancing consumer satisfaction is closely linked with the acceptance of such technological innovations. In other words, delving into the creation of customer satisfaction and how technology can boost it can ultimately influence a consumer’s ongoing engagement with the technology [57,58]. Tools like ChatGPT and other AI systems can automate customer service, provide round-the-clock support, respond instantly to queries, minimize response times, and supply relevant information, all of which are crucial for improving customer satisfaction. Recent scholarly efforts have focused on exploring how AI tools contribute in this way, including the development of models for human–chatbot interactions, understanding user satisfaction and loyalty towards service chatbots, and studying how the conversational design of chatbots impacts social, behavioral, and emotional outcomes [59,60,61]. Another area of study has been the different types of Gen AI outputs based on textual input data. The diverse array of possible functions showcases the broad and versatile applications of Gen AI, producing outputs such as decisions, audio, images, and text [61,62,63].

2.2. Natural Language Processing (NLP)

NLP utilizes computers to handle and analyze extensive textual data. Situated at the convergence of computer science, artificial intelligence, and linguistics, NLP involves using computational techniques to analyze text or generate language content. This area of study includes speech recognition and text syntax processing, although it primarily focuses on semantic analysis in the social sciences [64]. According to Hirschberg and Manning [65], NLP and computational linguistics are synonymous, though NLP is generally viewed as adopting a more empirical method in everyday contexts.

Figure 1 illustrates the relationships and distinctions within NLP, highlighting three key areas: general NLP tasks, Natural Language Understanding (NLU), and Natural Language Generation (NLG) [66]. NLP has been employed for numerous automated tasks, such as processing, analyzing, and storing large documents; examining customer opinions or call center records; operating chatbots for automated customer service; answering questions about who, what, when, and where; and classifying and organizing text [66,67].

NLP is highly versatile and already embedded in various everyday applications. For instance, machine translation services like Google Translate and DeepL enable machines to accurately convert text from one language to another [68]. Text prediction features, such as auto-correction, assist in quickly resolving typos and grammatical errors [69]. Text summarization condenses a text or collection of texts into a shorter version by highlighting the most important keywords, phrases, or sentences. For handling large collections of texts, topic modeling is a powerful method for extracting and summarizing key concepts and themes [70]. Finally, for texts with subjective content, such as feedback surveys and reviews, sentiment analysis or opinion mining is the most suitable approach [71,72].

NLP for Tourism

NLP has emerged as a transformative technology in the tourism industry, offering various applications that enhance customer experiences, streamline operations, and provide valuable insights. This literature review explores the integration of NLP in tourism, highlighting key areas such as customer service, sentiment analysis, recommendation systems, and content generation [71,72]:

• Customer service and chatbots: One of the prominent applications of NLP in tourism is in customer service, particularly through the use of chatbots. Chatbots leverage NLP to understand and respond to customer inquiries in real time, providing instant support and information [73].
• Sentiment analysis: Sentiment analysis, another critical NLP application, involves analyzing customer reviews and feedback to gauge sentiment and identify trends. By processing large volumes of text data from online reviews, social media posts, and surveys, sentiment analysis tools can help tourism businesses understand customer preferences and identify areas for improvement [74].
• Recommendation systems: NLP-powered recommendation systems are also revolutionizing the tourism sector. These systems analyze user preferences, past behaviors, and contextual information to provide personalized travel suggestions. For instance, systems can recommend destinations, activities, and accommodations tailored to individual tastes [75].
• Content generation: Using NLP technologies such as automated travel guides and blog posts is becoming increasingly popular. These tools can create engaging and informative content by extracting relevant information from various sources and presenting it coherently [76]. Automated content generation helps tourism companies maintain an active online presence and provide valuable information to potential travelers.

The integration of NLP in tourism has significantly improved various aspects of the industry, from customer service and sentiment analysis to recommendation systems and content generation [77]. As NLP technology continues to advance, its applications in tourism are expected to expand further, offering even more sophisticated tools to enhance the travel experience and operational efficiency [78].

2.3. Internet of Things (IoT)

Internet of Things (IoT) is a new innovation that combines five components: hardware, connectivity, software, data, and intelligence. IoT devices can connect and communicate with each other through the Internet network [1] by packing “embedded system devices” into “Things” (or various electronic devices). The components of IoT—sensors, connectivity, data processing, and user interfaces—work together to create an integrated system capable of improving efficiency and enabling new applications across various domains [1]:

• Sensors: These are fundamental components of IoT systems that are responsible for collecting data from the environment. These devices detect physical phenomena such as temperature, humidity, motion, light, and pressure, converting them into data that can be analyzed and utilized [1,79].
• Connectivity: This is crucial for enabling communication between IoT devices and the cloud or other data processing systems. Various communication protocols and technologies are used, including Wi-Fi, Bluetooth, Zigbee, and cellular networks [80]. Effective connectivity ensures that the data collected by sensors are transmitted reliably and efficiently to the appropriate processing units.
• Data processing: Once data are collected and transmitted, they need to be processed and analyzed to derive meaningful insights. Data processing can occur at the edge, closer to the data source, or in centralized cloud servers [81].
• User interfaces (UIs): These are the components through which users interact with IoT systems. UIs can take various forms, such as mobile apps, web dashboards, or voice assistants, allowing users to monitor and control IoT devices [82].

IoT for Tourism

The IoT has rapidly become a significant technological advancement in the tourism industry. By connecting various devices and systems, the IoT enhances the efficiency, personalization, and overall experience of travel. The role of the IoT in tourism focuses on developing smart destinations by enhancing customer experiences and developing efficient operations and data-driven insights [83]. The benefits of applying the IoT in tourism are as follows:

• Smart destinations: The IoT plays a crucial role in developing smart destinations, which use interconnected devices to manage and optimize resources. For example, IoT-enabled sensors can monitor and manage tourist flows, reducing congestion and improving the visitor experience [84]. Smart destinations leverage IoT technology to provide real-time information about attractions, transportation, and local services, making travel more convenient and enjoyable.
• Enhanced customer experiences: The IoT significantly enhances customer experiences by providing personalized services. Wearable devices, such as smartwatches, can offer tourists personalized recommendations based on their preferences and real-time location data [85]. Hotels use IoT devices to automate room settings according to guest preferences, such as lighting, temperature, and entertainment systems, thereby increasing comfort and satisfaction.
• Efficient operations: The IoT improves operational efficiency in the tourism sector by enabling better resource management and reducing costs. For instance, IoT-enabled energy management systems can optimize energy usage in hotels and other tourist facilities, leading to significant cost savings and sustainability benefits [86].
• Data-driven insights: The vast amount of data generated by IoT devices provides valuable insights for tourism businesses. By analyzing these data, companies can understand customer behaviors, preferences, and trends, enabling them to tailor their services and marketing strategies more effectively [23].

The integration of the IoT in tourism offers numerous benefits, including the development of smart destinations, enhanced customer experiences, improved operational efficiency, and valuable data-driven insights. As IoT technology continues to evolve, its applications in the tourism industry are expected to expand, providing even more innovative solutions to enhance the travel experience and optimize operations [1,85,86,87,88].

2.4. Smart Tourism Destinations

Smart tourism is the integration of various kinds of tourism information, such as information on tourist attractions, accommodations, restaurants, and transportation. The tourism industry uses technology and digital communication formats to interact with tourists, such as through smartphone applications. In many areas, other contemporary technologies such as VR, XR, AR, media art, and the metaverse are also being used to help create content to attract tourists [89].

Figure 2 illustrates a comprehensive smart tourism framework that emphasizes the integration of data and various components to enhance the tourist experience. Smart destinations leverage technology to improve the appeal and functionality of tourist destinations, while the smart business ecosystem integrates various businesses within the tourism sector to promote synergy and efficiency. Smart experience focuses on delivering personalized and enriched experiences for tourists through the use of smart technologies.

As smart tourism destinations continue to evolve, they are expected to set new standards for the tourism industry, offering innovative solutions to meet the demands of modern travelers [2,89,90]:

• Integration of IoT: IoT technologies are cornerstones of smart tourism destinations. The IoT enables real-time data collection and monitoring through interconnected devices and sensors. These technologies can monitor tourist flows, optimize transportation systems, and provide personalized services [91].
• Big data analytics: This plays a crucial role in the functionality of smart tourism destinations. By analyzing vast amounts of data from various sources such as social media, mobile apps, and IoT devices [1], tourism managers can gain valuable insights into visitor behaviors and preferences [23,89].
• Smart infrastructure: This encompasses the physical and digital systems that support the operations of a smart tourism destination. This includes smart transportation networks, energy-efficient buildings, and digital information platforms. Smart infrastructure enhances the efficiency and sustainability of tourism operations. For example, smart grids and energy management systems reduce the environmental impact of tourist facilities [92].
• Impact on tourism management: Smart tourism destinations provide significant benefits for tourism management. Real-time data and predictive analytics enable proactive decision making and efficient resource allocation. This results in improved operational efficiency and better visitor management [89,92].

Visitor satisfaction is a critical measure of the success of smart tourist destinations. Personalized services, efficient transportation, and real-time information contribute to a more enjoyable and convenient travel experience. Studies have shown that tourists appreciate the enhanced convenience and personalized experiences provided by smart technologies [89,93]. Consequently, destinations that invest in smart technologies tend to see higher levels of visitor satisfaction and loyalty.

2.5. Research Hypothesis

Gen AI technology continues to evolve and its impact on tourism is expected to grow, offering even more innovative solutions to meet the demands of modern travelers. The integration of Gen AI and the IoT is transforming the tourism industry by enhancing smart destinations, personalized marketing, customer engagement, and operational efficiency [94]. These technologies work together to provide real-time data analysis, personalized experiences, and efficient resource management. As the adoption of these technologies continues to grow, their impact on the tourism industry is expected to expand, offering even more innovative solutions to meet the evolving needs of travelers [94,95]. However, several factors must be considered regarding the integration of Gen AI and the IoT in tourism to test the following research hypotheses:

2.5.1. Attention

The integration of Gen AI and the IoT in tourism has far-reaching implications for enhancing “attention” due to increasing relevance, originality, and accessibility. By leveraging real-time data and AI-driven insights, tourism services can become more personalized, accurate, innovative, and inclusive. As these technologies continue to evolve, their impact on the tourism industry is expected to grow, offering new opportunities and creating greater “attention”, thus enhancing the travel experience and helping to meet the diverse needs of modern travelers [95,96,97]. There are four variables that influence attention in tourists’ planning prior to and during a trip: relevance [91,98], authenticity [89,99,100], originality [100,101], and accessibility [95,102].

H1: 

Attention significantly influences tourists’ planning prior to and during a trip.

2.5.2. Interest

The integration of Gen AI and the IoT holds transformative potential for the tourism industry, particularly in enhancing interest through improved content creation [91,102], inspiration [103], diversity [99,104], presentation [23,101,102], and impact on tourist’s planning [105].

H2: 

Interest significantly influences tourists’ planning prior to and during a trip.

2.5.3. Usage

By offering customized suggestions, Gen AI with the IoT enhances the overall travel experience, making it more enjoyable and relevant for tourists. There are six variables influencing tourists’ planning: personalization [104,106], news [23,106], interaction [11,33,101,104,106,107], awareness [5,101], utility [95,105], and stability [34,84,91].

H3: 

Usage significantly influences tourists’ planning prior to and during a trip.

2.5.4. Tourists’ Planning Prior to Trip

The technologies enhance data collection and analysis, enable personalized experiences, optimize resource management, support real-time decision making, and improve strategic marketing [2,89]. The integration of IoT devices enables extensive data collection from various sources such as sensors, mobile devices, and social media. These data include information about tourist behaviors, preferences, environmental conditions, and transportation patterns [23]. Gen AI processes and analyzes this vast amount of data to uncover patterns, trends, and insights that are critical for informed decision making in tourism planning [104,107].

H6: 

Tourists’ planning prior to and during a trip significantly influences their experiences.

2.5.5. Tourism Experience

Combining Gen AI with the IoT also improves customer engagement by providing interactive and immersive experiences. Moreover, IoT-enabled environments, such as smart hotels and attractions, use AI to adjust settings like lighting, temperature, and multimedia content to match individual preferences, creating a more engaging and comfortable experience for visitors [6,16,99]. Generative AI, when combined with IoT, allows for the creation of highly personalized tourism experiences. IoT devices track real-time data about tourists’ interactions and preferences, which AI then uses to offer tailored recommendations and services [34,84,91,95].

H5: 

Tourists’ planning prior to and during a trip significantly influences satisfaction.

2.5.6. Emotion

As AI and the IoT continue to evolve, their impact on the tourism industry is expected to enhance the emotional aspects of travel by enabling tourists to send feedback [85,87,91,95], create empathy [101,105], build memories [6,16,99,104], and receive personalized recommendations [85,87,95,96,105].

H4: 

Emotion significantly influences tourists’ planning prior to and during a trip.

2.5.7. Satisfaction

Gen AI also enhances customer engagement through interactive and immersive experiences [101,106,107]. These AI-driven tools can simulate human-like conversations, improving the efficiency and quality of customer service. Additionally, AI can create virtual tours and augmented reality (AR) experiences, allowing tourists to explore destinations digitally before their visit [99,104,108]. Gen AI offers numerous opportunities to enhance tourism satisfaction through personalized services, real-time assistance, immersive experiences, operational efficiency, and enhanced safety and security [109].

H7: 

Experience significantly influences satisfaction.

H8: 

Emotion significantly influences satisfaction.

H9: 

Perceived value significantly influences satisfaction.

H10: 

Interest significantly influences satisfaction.

Figure 3 illustrates the hypothesis testing.

3. Research Methodology

3.1. Problem Statement

The problem statement for this research is that there is a need for detailed and integrated innovations that can provide information to support disabled tourists and make the tourism planning system accessible for all travelers planning their journeys, using a case study in Thailand. Tourists often seek detailed, context-specific information to enhance their travel experiences. They inquire about tourist destinations, activities, services, and practical details such as distance and travel time. The tourism industry needs a system that can provide “decision support information” tailored to each tourist’s preferences, ensuring an engaging and informative interaction that leaves a positive impression.

Key Assumptions: The research team made several assumptions about the information tourists seek, their interaction patterns with the system, and their satisfaction with using a Gen AI system [99,104,108]. Based on these assumptions and a review of the relevant literature [99,101,102,103,104,106,107,108,110], the team developed questions to guide system development:

Interest in ancient sites: Tourists want to know about ancient sites in specific areas and what makes them special.

Location and specifics: Tourists inquire about the location, special features, distance, and travel time to these sites.

Activities: Tourists seek information on the available activities at these destinations.

Amenities: Tourists ask about the availability of food, lodging, and souvenirs at the sites.

Hierarchical interaction: The system must support a step-by-step interaction with tourists, starting with a broad overview that matches their interests and gradually providing detailed information to help them make informed travel decisions. This approach ensures that the system addresses the hierarchy of questions typically asked by tourists, facilitating a comprehensive and satisfying travel planning experience.

3.2. Research Design

After selecting Agile development methodology as the primary research methodology, the research team conducted the study using five steps: system requirements analysis, system design, system development, system evaluation, and deployment [111,112,113].

Figure 4 illustrates the Agile development methodology of the proposed system, which adopts an iterative and incremental approach to software development. Here, we provide an explanation of each stage in the cycle [111,112,113]:

System requirements analysis: In this initial phase, the requirements of the system are gathered and analyzed. This involves understanding what the users need, identifying the scope of the project, and defining the functional and non-functional requirements.

System design: Based on the requirements gathered, the system design is created. This involves designing the architecture of the system, defining the system components, and creating design documents. This phase ensures that the development team has a clear understanding of how the system should be built.

System development: During this phase, the actual coding and development of the system take place. Developers create the software components as per the design specifications. This phase is highly iterative, with developers continuously building, testing, and refining the system.

System evaluation: After development, the system undergoes thorough testing and evaluation. This includes unit testing, integration testing, and user acceptance testing. The goal is to identify any defects or issues and ensure that the system meets the defined requirements.

Deployment: once the system has been evaluated and tested, it is deployed to the production environment. This phase involves the actual implementation of the system for use by the end users. Deployment can be carried out in stages or all at once, depending on the project requirements.

Stage of system design: In the system design process, the research team took the users’ needs outlined in step 1 into consideration and designed a system that can provide information that is consistent with tourists’ inquiries, including being able to create multiple answers to queries. The technologies that support this are NLP and Gen AI technologies, enabling the creation of a consistent set of answers using tourists’ context data. The research team designed a system to store tourist information in a database and use RFID devices as user identification cards to pass into the system and connect the user’s context data with the system’s answer set [114].

Figure 5 shows the architecture of a smart tourism system combining both Gen AI and IoT technologies, resulting from the Agile development methodology outlined in the stage of system design. Tourists register and enter their basic information into the system’s database. The system then uses the tourist ID, registration date, and RFID number as a key for logging in tourists, and this RFID number will be embedded in the smart wristband that will be distributed to tourists. When tourists arrive at a place in which tourism businesses provide smart kiosks, RFID numbers are read by the Raspberry Pi board to send system access requests to the tourism server. After the tourism server has verified the tourist’s rights, it will be open for tourists to come in and ask for various kinds of travel information from the system through the speech-to-text module. Inquiry sentences will be sent into the tourism server and converted into “sentences” or “strings”, which are then analyzed for consistency with the digital dialog or interactive sentences that the system has been trained to recognize by identifying basic information and subjects that tourists are interested in, and the information or questions/answers used to create a digital dialog are mostly obtained by scrapping information from websites recommending tourist attractions.

The integration NLP in the system, particularly within the context of the smart kiosk interface, is a critical component. This integration allows the system to effectively handle diverse tourist inquiries and provide accurate, culturally sensitive responses despite the inherent complexity and diversity of natural language:

• Handling diverse inquiries: The NLP module is designed to process a wide range of tourist inquiries, from simple questions about local attractions to more complex queries regarding travel logistics. It employs advanced algorithms to understand the context and intent behind each question, ensuring that relevant and accurate information is provided.
• Ensuring accuracy: To ensure high accuracy, the NLP system utilizes a combination of machine learning models and extensive linguistic databases. These models are trained on large datasets that include various dialects, slang, and idiomatic expressions, enabling the system to comprehend and respond accurately to different forms of natural language input.
• Cultural sensitivity: Cultural sensitivity is paramount in the responses generated by the NLP module. The system incorporates cultural context into its processing, considering local customs, traditions, and etiquette. This is achieved by integrating cultural databases and context-aware algorithms that adjust their responses to be respectful and appropriate for the cultural background of the user.
• Continuous learning and adaptation: The NLP module is designed to continuously learn and adapt from user interactions. Feedback mechanisms are in place to refine and improve the system’s understanding and response generation over time. Adaptive learning ensures that the system remains up to date with evolving language trends and cultural nuances.

By addressing these aspects, the NLP integration within the smart kiosk interface enhances user experiences by providing reliable, contextually appropriate information tailored to the needs of diverse tourists.

System development stage: The development process began with the research team designing the system’s database system, followed by the creation of the hardware part of the system, that is, the smart kiosk, as well as the smart wristband. Natural language processing was used to develop the speech-to-text and text-to-speech parts. The next step was to develop the set of commands to be installed into the tourism server, which consisted of a Web scrapping part and a digital dialog part, as well as the section used to find the similarity of the question. Most of the modules in this section are written in PHP.

Stage of system evaluation: For the process of evaluating the system, the research team chose Agile development methodology, which focuses on user satisfaction. In this stage of the research, the research team evaluated the satisfaction and response to the system of both a group of experts and a group of tourists.

Stage of deployment: In the deployment process, the research team installed smart kiosks at various important points. In the tourism routes of the pilot area, every smart kiosk was connected to the Internet via a Wi-Fi router, and the command set used in the tourism server was uploaded with an FTP client program. This program was stored in a production server, which could be accessed from outside the network.

3.3. Details of System Development

3.3.1. NLP Coding

The primary function of this web page was to facilitate voice input for searches for tourism information in Suphanburi, Thailand.

The HTML structure used for the system started with a basic document type declaration (<!doctype html>). The head section contained the title and embedded CSS styles, while the body section included the main content and JavaScript. CSS styling defined the font style for the entire body using Arial. The button was styled with padding, background color, text color, border properties, cursor style, and rounded corners. The output div was styled with background color, padding, width, margin, and line height, and the classes “hide” and “show” controlled the visibility of the elements.

The HTML content included an image tag to display an image (SuphanLady.jpg), a heading for the application title, a paragraph instructing the user to click the button and speak, a button that triggers the speech recognition function, and an output div to display the results, which are initially hidden.

The JavaScript functionality involved defining the “run speech recognition” function to handle speech recognition. It retrieves information from output and action span elements, creates a new speech recognition object, and defines event handlers for the speech recognition process. “On start event” updates the action element to indicate that the system is listening, “on speech end” updates the action element to indicate that the system has stopped listening and stops the recognition process, and “on result” retrieves the transcript and confidence score from the speech recognition results, redirects the browser to exec_SuphanGPT.php with the transcript as a URL parameter, and reveals the output div. Finally, it starts the speech recognition process.

3.3.2. IoT Development

This is the embedded command line that served as the system’s lightweight web server, written in Micro Python 3. This script is designed to set up a Wi-Fi connection on an ESP32 device, read temperature data from an analog sensor, and serve these data via a simple HTTP server. The webpage displays the temperature in Celsius and the magnetic field strength from the ESP32’s hall sensor. This script sets up a Wi-Fi connection, reads temperature data from an analog sensor, and serves an HTML page displaying the temperature and magnetic field data via a simple HTTP server. The HTML page refreshes every 2 s to show updated data. The steps are as follows:

Import required modules: import necessary modules such as Pin and ADC from the machine library, network for Wi-Fi connectivity, socket for the HTTP server, and esp32 and esp for accessing ESP32-specific functions; disable debugging to reduce log noise (esp.osdebug(None)); collect garbage to free up memory (gc.collect()).

• Initialize Wi-Fi credentials: set the SSID and password for the Wi-Fi network.
• Set up temperature sensor: configure the ADC (Analog-to-Digital Converter) on pin 34 with an attenuation of 11 dB for a wider input voltage range.
• Define Wi-Fi setup function: create the setup_wifi function to connect to the specified Wi-Fi network.
• Set up temperature sensor: configure the ADC (Analog-to-Digital Converter) on pin 34 with an attenuation of 11 dB for a wider input voltage range.
• Define temperature reading function: create the read_temp function to read the analog value from the temperature sensor, convert it to Celsius, and return it as a string.
• Define webpage generation function: create the web_page function to generate and return the HTML string containing the temperature and magnetic field data.
• Main function: create the main function to call the setup_wifi and http_server functions.
• Execute main function: ensure the main function runs when the script is executed.

3.3.3. Integration of Gen AI with IoT

In this part, a set of commands was created that received data from the speech-to-text module and converted it into text, which was then analyzed using generative AI to extract the keywords that the domain of the search term has given the most weight. These keywords were then matched with related information and sent back to the user. To perform this function, a set of commands in exec_SuphanGPT.php were used.

The exec_SuphanGPT.php script processes the value received from the URL parameter (v1) and redirects the user to different pages based on the input. It serves as a simple decision-making engine for navigating specific pages related to tourism information categories. The exec_SuphanGPT.php script processes a URL parameter to determine which tourism-related page to navigate based on predefined decision categories. If the input value matches a category, the user is redirected to the corresponding page; otherwise, an error message is displayed.

The program starts by retrieving a value from the URL parameter v1 and storing it in the variable $val. Several decision variables are then initialized with predefined strings: “Type”, “where”, “Distance”, “Activities”, “Food”, “Time”, “Lodging”, and “Souvenirs”. The program uses a series of if–else-if statements to check if the input value ($val) matches any of these predefined decision variables. If a match is found, the program displays the message “You speak” followed by the matched decision variable and executes a JavaScript snippet to redirect the user to a corresponding PHP file. Specifically, if $val is “Type”, the user is redirected to Type.php. Similar redirects occur for other values: where.php, Distance.php, Activities.php, Food.php, Time.php, Lodging.php, and Souvenirs.php. If the input value does not match any of the predefined decision variables, the program displays a message indicating that the command is not in the system. Finally, the program outputs the appropriate messages and scripts to the web page based on the checks and conditions described.

3.3.4. System Implementation

Integrating Gen AI with the IoT to support smart tourism involves multiple steps, including assembling hardware, writing instructions, and testing the system. Here, we provide a detailed explanation of each step: (1) hardware assembly on the user side, (2) smart kiosk hardware assembly, (3) writing instructions for user-side hardware, (4) writing instructions for smart kiosk, (5) writing commands for the tourism server, and (6) system testing.

3.3.5. Tourist System Utilization and User Evaluation

System utilization: After the expert evaluation phase, the system was implemented and used by tourists in the pilot testing site at Suphanburi Province and Phra Nakhon Si Ayutthaya Province the UNESCO World Heritage Cities, and the system was trialed for 6 months [115]. User evaluation after implementing the integration of Gen AI with IoT, the system was provided to users for evaluation.

Population and sample: The population undertaking the evaluation comprised tourists in Thailand. The population included elderly individuals, tourists with disabilities, and tourists with special requirements, ensuring support for all visitors while adhering to ethical standards for data collection. The non-probability sampling technique of convenience was used. The sample size was more than 400 participants, with a total of 416 based on Cochran [116] with a confidence level of 95% (α = 0.05).

Data collection and analysis: Based on the conceptual framework of the research and the literature review, the researchers designed a questionnaire with three sections. The first section had five closed questions (gender, age, education, occupation, and income) related to the demographics of the sample. The second section of the questionnaire evaluated the use of generative AI integrated with the IoT in tourism and consisted of 27 questions (attention, interest, usage, and emotion). The third section evaluating tourists’ satisfaction had three closed questions (tourism planning before, during trip, tourism experience, and tourism satisfaction). The questions used a five-point Likert scale, with 5 = strongly agree, 4 = agree, 3 = moderate, 2 = disagree, and 1 = strongly disagree. The reliability of the measures was tested using Cronbach’s alpha = 0.974. The data were analyzed using SPSS for descriptive statistics. An exploratory factor analysis (EFA) and confirmatory factory analysis (CFA) were run using LISREL 9.0.

4. Experimental Results

4.1. Development of Integrated Gen AI with IoT System for Tourism

This paper presents the results of research conducted to develop a platform for providing decision-making information to tourists, which was developed by integrating Gen AI with NLP and the IoT. The result is a platform for intelligent tourism that has three tiers of functionality. On the tourist side, which is Tier 1, it uses a small ESP32 family of microcontroller boards in the form of a programmable wristband and RFID tag. This functions like a tourist’s identification card, enabling them to log in and ask for the information needed for making various decisions. When contact is made from the user tier or Tier 1 to the smart kiosk or Tier 2 (which runs the HTTP Server #1 program), the system will display various messages to users in the form of a web page.

Figure 6 shows the user interface on the smart kiosk side. After logging in with the smart wristband, users or tourists will be able to make inquiries for information by speaking into the smart kiosk microphone instead of typing. The smart kiosk has a speech-to-text module to convert speech into “text” or “Strings” and send it to the tourism server (which runs the HTTP Server #2). On the tourism server side (Tier 3), when a conversation or sentence from the smart kiosk arrives, the system will compare the text in the tourist’s sentence with the digital dialog, which has been generated since the tourist registered in the system’s database. The most relevant information will then be fed as a conversation back to the tourist, utilizing the text-to-speech module to interact with tourists in the form of speech through the smart kiosk speakers.

4.2. System Evaluation

4.2.1. Expert Evaluation

After the system was implemented, a comprehensive evaluation of the system was undertaken across various categories, including usability, stability, accuracy, and completeness by 20 experts and specialists in Gen AI and IoT. To evaluate the system, the research team used the User Acceptance Test, including all five aspects of evaluation.

The gender distribution is evenly split, with 10 female experts and 10 male experts. Most experts fall within the 35 to 40 age range. The majority of the experts hold a Master’s degree (12 experts), while the remaining 8 experts have a Doctoral degree. This distribution indicates that the majority of the experts have advanced educational qualifications, suggesting a high level of expertise and knowledge in their fields. Figure 7 illustrates the occupation distribution of the 20 experts who evaluated the information system. Each occupation is represented by exactly one expert, indicating a diverse range of professional backgrounds. The occupations included in the evaluation are IT Manager, Software Engineer, Cloud Engineer, System Architect, Business Analyst, Research Scientist, IT Trainer, IT Auditor, Technical Writer, IT Support Specialist, UX/UI Designer, Project Manager, Software Developer, IT Consultant, Systems Analyst, Cybersecurity Specialist, Database Administrator, Network Administrator, Data Analyst, and AI Specialist. This diversity ensures that the evaluation incorporates perspectives from various areas of expertise within the IT and related fields, leading to a comprehensive assessment of the information system.

Figure 8 illustrates the comparison of the system by education level. The analysis of satisfaction ratings by education level reveals some interesting insights: Experts with a Master’s degree tend to have slightly higher and more consistent satisfaction ratings compared to those with a Doctoral degree. This consistency might be due to a more homogeneous group of Master’s degree holders in terms of their expectations and experiences with the information system. The wider range of satisfaction ratings among Doctoral degree holders suggests more diverse opinions within this group. This diversity could be attributed to the varied research and practical experiences of Doctoral degree holders, leading to different expectations and evaluations of the system’s performance. Both groups generally rate the system positively, with median satisfaction ratings around or above 4.0. This overall positive feedback indicates that the information system meets the expectations of highly educated experts. These findings align with previous research suggesting that individuals with different educational backgrounds may have varying perspectives on technology adoption and satisfaction.

Table 1. Expert evaluation of the system.

Indicator	Mean	Standard Deviation
Ease of use	4.00	0.85
Ease of access	4.00	0.77
System stability	4.00	0.65
Network stability	4.00	0.87
Runtime errors	3.50	0.55
System accuracy	4.50	0.89
System completeness	4.00	0.98
Overall rating	4.00	0.76

illustrates the system was found to be generally easy to use and access, with scoring a mean of 4.00 with moderate variability in user experiences (standard deviations of 0.85 and 0.77, respectively). System and network stability also scored 4.00, indicating reliability, though runtime errors have a slightly lower mean of 3.50, suggesting occasional issues (standard deviations of 0.65, 0.87, and 0.55). System accuracy is the highest-rated indicator with a mean of 4.50, showing high accuracy but moderate variability (standard deviation of 0.89). The system’s completeness is rated 4.00, indicating it meets requirements well, although perceptions vary more widely (standard deviation of 0.98). Overall, the system performs well across all indicators, with an overall mean rating of 4.00 and a moderate variability in user experiences (standard deviation of 0.76).

4.2.2. User Evaluation

User evaluation results for the system, obtained through a questionnaire, indicate a robust causal relationship model analysis for using integrated Gen AI and the IoT to promote tourism. This model comprises 7 latent variables and 29 observable variables, with relationships ranging from 0.339 to 0.923, as detailed in Appendix A

The demographics of the 416 samples found that the majority of tourists were female (338 people or 81.45%), with 77 males (18.55%); most of them were aged between 35 and 44 years (35.42%), earned between 35,001 and 40,000 THB/month, and were government employees (29.40%).

Figure 9 illustrates the causal relationship model analysis for using integrated Gen AI and the IoT to promote tourism and reveals that ATTENTION, INTEREST, USAGE, and EMOTION have a direct positive influence on tourists’ planning prior to and during a trip, with effect sizes of 0.079, 0.221, 0.249, and 0.609, respectively. Additionally, ATTENTION, INTEREST, USAGE, and EMOTION have an indirect positive influence on EXPERIENCE through the variable tourists’ planning prior to and during a trip, with effect sizes of 0.072, 0.201, 0.226, and 0.554, respectively. Tourists’ planning prior to and during a trip has a direct positive influence on EXPERIENCE, with an effect size of 0.909, and an indirect positive influence on satisfaction through the variable EXPERIENCE, with an effect size of 0.623. EMOTION and EXPERIENCE have a direct positive influence on satisfaction, with effect sizes of 0.144 and 0.685, respectively. Tourists’ planning prior to and during a trip does not have an influence on satisfaction, but the interaction between EMOTION and tourists’ planning prior to and during a trip has a direct positive influence on satisfaction, with an effect size of 0.296.

When considering the components of ATTENTION, the aspect with the highest weight is originality, with tourists perceiving that AI provides direct, novel information. Authenticity follows, where tourists see AI as capable of presenting in-depth and detailed insights into tourism activities. Relevance is the next significant component, where tourists believe AI supports tourism activities. These three aspects explain 75.70%, 72.10%, and 63.80% of the variance in ATTENTION, respectively.

For INTEREST, the aspect with the highest weight is diversity, with tourists feeling that AI can generate and present a variety of useful information for planning trips. Presentation follows, where tourists believe AI presents information in an easily understandable format. Content is next, with tourists seeing AI as capable of delivering substantial information. These three aspects explain 70.70%, 67.50%, and 64.70% of the variance in INTEREST, respectively.

Regarding USAGE, awareness has the highest weight, with tourists believing AI can provide detailed, user-specific information. Stability follows, where tourists view AI as a stable system, as well as interaction, where tourists see AI as capable of engaging well with them. These three aspects explain 78.70%, 69.90%, and 68.80% of the variance in USAGE, respectively.

For EMOTION, the aspect with the highest weight is memories, with tourists believing AI helps preserve good memories. Recommendation follows, where tourists see AI as capable of giving good suggestions, as well as empathy, where tourists feel AI understands user emotions. These three aspects explain 73.20%, 64.70%, and 48.50% of the variance in EMOTION, respectively.

In the context of tourists’ planning prior to and during a trip, the aspect with the highest weight is reliance on AI during the trip. This is followed by reliance on AI for pre-trip planning and providing feedback to the AI system post-trip. These three components explain 84.30%, 79.90%, and 65.30% of the variance in tourists’ planning prior to and during a trip, respectively.

For experience, the highest weight aspect is tourists recommending AI for a better travel experience to others. Following this, tourists plan to use AI for future trips, and lastly, tourists intend to use AI support for other purposes in the future. These three aspects explain 80.20%, 71.20%, and 50.90% of the variance in experience, respectively.

Regarding satisfaction, the aspect with the highest weight is overall satisfaction with the accuracy of the information provided by AI. Following this is overall satisfaction with the timeliness of the information, and lastly, overall satisfaction with the security of the information. These three aspects explain 83.30%, 80.60%, and 70.00% of the variance in satisfaction, respectively.

The interaction between experience and tourists’ planning prior to and during a trip has a direct positive influence on satisfaction, with an effect size of 0.421. The details are shown in

Table 2. Direct effects and indirect effects.

Latent Variables	Tourists’ Planning Prior to and During Trip			Experience			Satisfaction
	TE	DE	IE	TE	DE	IE	TE	DE	IE
USAGE	0.249 ***	0.249 ***	_	0.226 ***	0	0.226 ***	_	_	_
INTEREST	0.221 ***	0.221 ***	_	0.201 ***	0	0.201 ***	_	_	_
ATTENTION	0.079 **	0.079 **	_	0.072 **	0	0.072 **	_	_	_
EMOTION	0.609 ***	0.609 ***	_	0.554 ***	0	0.554 ***	0.144 **	0.144 **	-
Tourists’ planning prior to and during trip	_	_	_	0.909 ***	0.909 ***	_	0.623 ***	_	0.623 ***
Experience	_	_	_	_	_	_	0.685 ***	0.685 ***	_
INT1	_	_	_	_	_	_	0.421 ***	0.421 ***	_
INT2	_	_	_	_	_	_	0.296 **	0.296 **	_

** p-value < 0.01; *** p-value < 0.001.

.

This table presents the results of a structural equation modeling (SEM) analysis, which examines the relationships between latent variables and their effects on tourists’ planning prior to and during a trip, experience, and satisfaction.

For USAGE, the total effect (TE) and direct effect (DE) on tourists’ planning prior to and during a trip 0.249 ***, while for experience, the TE is 0.226 ***, with no direct effect but an indirect effect (IE) of 0.226 ***.

INTEREST shows similar patterns, with a TE and DE of 0.221 *** for tourists’ planning prior to and during a trip, and for experience, a TE of 0.201 *** with an IE of 0.201 ***.

ATTENTION impacts tourists’ planning prior to and during a trip with a TE and DE of 0.079 ** and experience with a TE of 0.072 ** and an IE of 0.072 **.

EMOTION shows significant effects across all stages: a TE and DE of 0.609 *** on tourists’ planning prior to and during a trip, a TE of 0.554 *** on experience (with an IE of 0.554 ***), and on satisfaction, a TE and DE of 0.144 **.

Tourists’ planning prior to and during a trip has a substantial effect on experience, with a TE and DE of 0.909 ***, and on satisfaction, with a TE of 0.623 *** and an IE of 0.623 ***.

Experience directly affects satisfaction, with a TE and DE of 0.685 ***. INT1 and INT2 also show direct effects on satisfaction, with INT1 having a TE and DE of 0.421 ***, and INT2 a TE and DE of 0.296 **.

In summary, USAGE, INTEREST, ATTENTION, and EMOTION directly influence tourists’ planning prior to and during a trip and indirectly influence experience. Tourists’ planning prior to and during a trip directly affects experience and indirectly affects satisfaction, while experience directly impacts satisfaction. EMOTION directly influences all three stages, and INT1 and INT2 directly affect satisfaction. This analysis demonstrates how these initial factors influence the overall tourism experience and satisfaction.

ATTENTION, INTEREST, USAGE, and EMOTION collectively explain 48.70% of the variance in tourists’ planning prior to and during a trip, which, in turn, explains 82.70% of the variance in experience. Additionally, experience, EMOTION, the interaction between tourists’ planning prior to and during a trip (INT1), and the interaction between tourists’ planning prior to and during a trip and EMOTION (INT2) collectively explain 93.90% of the variance in satisfaction. Specifically, INT1 explains 10.05% and INT2 explains 7.80% of the variance in satisfaction. The detailed results are shown in

Table 3. Model fit indices for the research model.

Latent Variables	Observed Variables	Standardized Regression Weights ($\mathit{\beta }$)	se	t	${\mathit{R}}^{\mathbf{2}}$
ATTENTION	Att1	0.8010	-	-	0.6380
	Att2	0.8490	0.0570	19.0110 ***	0.7210
	Att3	0.8740	0.0550	19.5270 ***	0.7570
	Att4	0.7890	0.0610	17.4730 ***	0.6290
INTEREST	In1	0.7950	-	-	0.6470
	In2	0.7900	0.0610	18.3020 ***	0.6350
	In3	0.8630	0.0500	18.7990 ***	0.7070
	In4	0.8250	0.0500	17.8080 ***	0.6750
USAGE	Usa1	0.7370	-	-	0.5530
	Usa2	0.7100	0.0680	14.6900 ***	0.5160
	Usa3	0.8340	0.0750	17.3510 ***	0.6880
	Usa4	0.8820	0.0680	18.4960 ***	0.7870
	Usa5	0.8020	0.0720	16.4810 ***	0.6370
	Usa6	0.8410	0.0720	17.1960 ***	0.6990
EMOTION	Emo1	0.6580	-	-	0.4510
	Emo2	0.6940	0.1070	12.3570 ***	0.4850
	Emo3	0.8720	0.0910	14.4420 ***	0.7320
	Emo4	0.7880	0.0880	13.8900 ***	0.6470
Tourists’ planning prior to and during trip	TPD1	0.9280	-	-	0.7990
	TPD2	0.9480	0.0360	28.6560 ***	0.8430
	TPD3	0.8590	0.0460	22.0040 ***	0.6530
Experience	Exp1	0.8860	-	-	0.7120
	Exp2	0.9220	0.0460	23.7150 ***	0.8020
	Exp3	0.7750	0.0550	16.6470 ***	0.5090
Satisfaction	Sat1	0.9010	-	-	0.8330
	Sat2	0.8860	0.0320	29.3300 ***	0.8060
	Sat3	0.8180	0.0450	24.7170 ***	0.7000
Experience tourists’ planning prior to and during trip	INT1	0.421	0.003	18.431 ***	0.1030
EMOTION tourists’ planning prior to and during trip	INT2	0.296	0.003	3.427 **	0.0780

** p-value < 0.01; *** p-value < 0.001.

.

Table 4. Conditional effects of the predictor at values of the moderator(s).

Moderator	BootMean (HIGH TPD)	BootSE (LOW TPD)	BootLLCI	BootULCI
Constant	−0.057	0.5164	−1.1141	0.8632
Experience	0.474	0.109	0.2673	0.6938

presents the conditional effects of experience on satisfaction at different levels of tourists’ planning prior to and during a trip (TPD), acting as a moderator. When tourism planning before, during a trip (TPD) is high, there is a positive and statistically significant relationship between experience and satisfaction, with a mean effect of 0.474. The confidence interval for experience (0.2673 to 0.6938) suggests a strong and reliable positive effect of experience on satisfaction when TPD is high.

Figure 10 illustrates the relationship between the interaction of EXPERIENCE with tourists’ planning prior to and during a trip (TPD) and satisfaction, using BootMean and BootSE estimation methods, each with a 95% confidence interval. The BootMean estimation is represented by the condition −0.057 + 0.474 EXPERIENCE. The BootSE estimation is represented by the condition 0.5164 + 0.1090 EXPERIENCE. The shaded areas around each line indicate the 95% confidence intervals for these estimations. This graph visually demonstrates the impact of varying levels of EXPERIENCE on satisfaction, according to both the BootMean and BootSE methods.

Table 5. Conditional effects of the predictor at values of the moderator(s).

Moderator	BootMean (High TPD)	BootSE (Low TPD)	BootLLCI	BootULCI
Constant	−0.127	0.4414	−1.0197	0.7437
EMOTION	0.2707	0.0756	−0.062	0.6101

illustrates that the effect of EMOTION on satisfaction, when TPD is high, is also not significant, as its confidence interval includes zero (from −0.062 to 0.6101). This suggests that the relationship between EMOTION and satisfaction is not reliably positive when moderated by high levels of TPD.

Figure 9 illustrates the relationship between the interaction of EMOTION with tourists’ planning prior to and during a trip (TPD) and satisfaction, using BootMean and BootSE estimation methods, each with a 95% confidence interval. The BootMean estimation is represented by the condition −0.127 + 0.2707·EMOTION−0.127 + 0.2707·EMOTION. The BootSE estimation is represented by the condition 0.4414 + 0.0756·EMOTION0.4414 + 0.0756·EMOTION. The shaded areas around each line indicate the 95% confidence intervals for these estimations.

Table 6. Model fit indices evaluation.

Fit Index	Criterion	Value	Evaluation
χ2/df	Less than 2.00	1.154	Pass
p-value	Greater than 0.05	0.071	Pass
GFI (Goodness-of-Fit Index)	Greater than 0.95	0.965	Pass
AGFI (Adjusted Goodness-of-Fit Index)	Greater than 0.95	0.961	Pass
CFI (Comparative Fit Index)	Greater than 0.95	0.990	Pass
NFI (Normed Fit Index)	Greater than 0.95	0.977	Pass
RMR (Root Mean Square Residual)	Less than 0.05	0.032	Pass
RMSEA (Root Mean Square Error of Approximation)	Less than 0.05	0.040	Pass

illustrate the results of the model fit indices indicate that χ² = 221.532, df = 192, χ²/df = 1.154 (less than 2), p-value = 0.071, Goodness-of-Fit Index (GFI) = 0.965, Adjusted Goodness-of-Fit Index (AGFI) = 0.961, Comparative Fit Index (CFI) = 0.990, Normed Fit Index (NFI) = 0.977 (greater than 0.95), Root Mean Square Residual (RMR) = 0.032, and Root Mean Square Error of Approximation (RMSEA) = 0.040 (less than 0.05). All indices indicate a good fit.

Table 7. Hypothesis testing.

Hypothesis	bb	Summary
H1	0.079 **	Significant at the 0.01 level
H2	0.221 ***	Significant at the 0.001 level
H3	0.249 ***	Significant at the 0.001 level
H4	0.609 ***	Significant at the 0.001 level
H5	0.05	Not significant
H6	0.909 ***	Significant at the 0.001 level
H7	0.685 ***	Significant at the 0.001 level
H8	0.144 **	Significant at the 0.01 level
H9	0.296 **	Significant at the 0.01 level
H10	0.421 ***	Significant at the 0.001 level

** significant at the 0.01 level; *** significant at the 0.001 level.

presents the results of hypothesis testing. The result found the following: H1: Attention was significant at the 0.01 level, indicating a moderately strong relationship with tourists’ planning prior to and during a trip. H2: Interest significantly, at the 0.001 level, influenced tourists’ planning prior to and during a trip. H3: Usage significantly, at the 0.001 level, influenced tourists’ planning prior to and during a trip. H4: Emotion significantly, at the 0.001 level, influenced tourists’ planning prior to and during a trip. H5: Not significant, indicating no strong evidence to support this hypothesis. H6: Tourists’ planning prior to and during a trip significantly, at the 0.001 level, influenced experiences. H7: Attention significantly, at the 0.001, influenced satisfaction. H8: Emotion significantly, at the 0.01 level, influenced satisfaction, indicating strong relationships. H9: Usage significantly, at the 0.01 level, influenced satisfaction, indicating strong relationships. H10: Interest is highly significant at the 0.001 level of influence on satisfaction, indicating very strong relationships.

In summary, the user evaluation of the integration of Gen AI, NL, and IoT systems, employing SEM modeling, examined the relationships between latent variables and their effects on tourists’ planning, experience, and satisfaction. Key findings include usage significantly impacted planning (TE and DE of 0.249 ***) and experience (TE of 0.226 ***). Interest affected planning (TE and DE of 0.221 ***) and experience (TE of 0.201 ***). Attention influenced planning (TE and DE of 0.079 **) and experience (TE of 0.072 **). Emotion had significant effects on planning (TE and DE of 0.609 ***), experience (TE of 0.554 ***), and satisfaction (TE and DE of 0.144 **). Planning had substantial effects on experience (TE and DE of 0.909 ***) and satisfaction (TE and IE of 0.623 ***). Experience directly affected satisfaction (TE and DE of 0.685 ***). These findings emphasize the significant roles of usage, interest, attention, and emotion in enhancing tourists’ planning, experience, and satisfaction, highlighting their importance in the design and implementation of tourism services [117,118].

5. Discussion

5.1. General Discussion

The tourism industry is undergoing a transformative revolution driven by the rapid advancement and integration of digital technologies. AI, NLP, and the IoT are at the forefront of this change, significantly influencing the sector by making tourism more accessible, personalized, and environmentally sustainable [119,120]. This paper makes substantial contributions to the field of smart tourism by demonstrating how the integration of generative AI, NLP, and the IoT can support tourists’ decision making and planning. The research offers both practical solutions and theoretical insights that enhance the accessibility, personalization, and environmental sustainability of tourism [117,119,120,121,122].

Innovative technological integration: A major contribution of this research is the demonstration of the successful integration of Gen AI and the IoT to create a comprehensive system that enhances the tourism experience. By employing advanced AI algorithms alongside NLP, developed and demonstrated systems capable of delivering predictive analytics, personalized recommendations, and real-time, multilingual communications, which are critical for modern tourism practices [121]. This integration significantly improves service personalization and operational efficiency, providing a model for future applications in the field [122].

Enhanced accessibility and inclusivity: The system developed can enhance accessibility and inclusivity in tourism services, especially for tourists with disabilities. The application of these technologies ensures that tailored services are provided to diverse populations, including individuals with disabilities, making tourism more inclusive [117]. This approach not only broadens the accessibility of tourism activities but also promotes an equitable tourism ecosystem. Especially, the use of this system can be implemented for blind tourists, allowing them to use voice input to develop personalized planning at their destinations [117].

Sustainability and environmental impact: The use of the IoT for real-time data collection on resource usage and environmental impacts can contribute significantly to sustainable tourism practices. For instance, in its test implantation, the system provided information about water reserves at UNESCO World Heritage sites in Phra Nakhon Si Ayutthaya and, thus, helped raise environmental conservation awareness among the tourists involved in the study. Additionally, this paper highlights how integrated technologies can aid in tourism resource management and the preservation of natural habitats, aligning with sustainable development goals in the tourism industry [120].

Cross-industry applicability: The paper outlines a scalable model for employing AI, NLP, and the IoT that can be adapted beyond the tourism industry to other sectors aiming to enhance customer satisfaction while balancing environmental and social responsibilities [120]. This cross-industry applicability, such as in retail commerce, logistics, and smart transportation, is significant. It provides a roadmap for integrating these technologies in various contexts to achieve similar goals [117,120].

Overall, this paper enriches the understanding of smart tourism technologies by showcasing their potential to transform tourism into a more adaptive, inclusive, and sustainable sector. The research findings contribute not only to academic knowledge but also offer practical insights that can be implemented by industry practitioners worldwide [117,121]

5.2. Theorical Implications

This study presents the results of research aimed at developing a platform that provides decision-making information to tourists by integrating Gen AI with NLP and the IoT. The outcome is an intelligent tourism platform featuring a three-tier functionality structure. The result of separating the system architecture into three areas results in a system that is highly flexible and easy to understand. Further, making changes in each tier has very little impact on other tiers.

An integration of AI, NLP, and the IoT represents a significant advancement in smart tourism. By employing advanced AI algorithms alongside NLP developed a system capable of generating predictive analytics, personalized recommendations, and conducting real-time, multilingual communication with tourists [119,122]. This capability is essential for modern tourism practices, as it allows for a high degree of personalization and engagement with tourists, catering to their unique needs and preferences. The system’s ability to convert speech to text and back to speech ensures a user-friendly experience, enabling tourists to obtain information efficiently and effectively [119]. The use of RFID tags and programmable wristbands facilitates easy access and interaction, thereby enhancing the overall user experience [119,120].

The results of the user evaluation revealed the key factors affecting tourists’ planning, tourists’ trip experiences, and the overall satisfaction of tourists using this system. These findings highlight the critical role of emotional and cognitive factors in shaping tourist behaviors and satisfaction. The result shows that attention, interest, usage, and emotion significantly impact both the planning and experience phases of a trip. The study findings on the influence of emotional and cognitive aspects on tourist behaviors and satisfaction underscore the importance of incorporating these elements into tourism planning and service design [123].

By highlighting the roles of cutting-edge technologies in redefining tourism practices, this research proposes a model that balances consumer satisfaction with environmental stewardship and social inclusivity. This theorical model suggests its applicability in other industries seeking similar goals, demonstrating the cross-industry potential of integrating AI, NLP, and the IoT to achieve sustainable and inclusive outcomes [124,125].

The paper details the technical implementation, design considerations, and challenges encountered in creating a more equitable and intelligent tourism ecosystem. It also discusses the ethical dimensions related to data privacy and security, emphasizing the importance of responsible technology deployment. This consideration is crucial as the use of AI and the IoT in tourism involves handling large amounts of personal data, necessitating stringent measures to protect user privacy and ensure data security [121].

The developed system provides stakeholders with effective tools for interacting with and providing information to tourists. The feedback from conversations and question-and-answer sessions can be utilized to plan and manage various aspects of the tourism industry, including budget, personnel, locations, and services. The evaluation of user satisfaction with the system aligns with findings from previous studies of user satisfaction in the tourism industry [119,120]. This alignment indicates that the system meets the needs and expectations of users, thereby enhancing the overall management and operational efficiency within the tourism sector.

The integration of generative AI, NLP, and IoT technologies in this research offers substantial benefits to stakeholders in the tourism industry, IoT developers, and AI researchers. The study findings demonstrate how advanced technologies can enhance service delivery, operational efficiency, and user satisfaction, thus contributing to the ongoing development of smart tourism and other technology-driven sectors [126,127].

Finally, for future research, the researchers plan to improve the smart wristband’s ability to accommodate elderly travelers or those who are visually impaired. There are also plans to develop NLP and Gen AI modules that work effectively with Thai sounds and sentences.

5.3. Tourism Business Implications

The tourism industry is undergoing a transformative revolution driven by the rapid advancement and integration of digital technologies. AI, NLP, and the IoT are at the forefront of this change, significantly influencing the sector by making tourism more accessible, personalized, and environmentally sustainable [118,128]. This paper makes substantial contributions to the field of smart tourism by demonstrating how the integration of generative AI, NLP, and the IoT can support tourists’ decision making and planning. The research offers both practical solutions and theoretical insights that enhance the accessibility, personalization, and environmental sustainability of tourism [129,130]. There are several key points for implications for tourism business sustainability:

Improves service personalization and operational efficiency: A major contribution of this research is the demonstration of the successful integration of Gen AI and the IoT to create a comprehensive system that enhances tourism business efficiency. By employing advanced AI algorithms alongside NLP developed systems capable of delivering predictive analytics, personalized recommendations, and real-time, multilingual communications, which are critical for modern tourism practices [131].

Enhanced accessibility and inclusivity: The developed system can enhance accessibility and inclusivity in tourism services, especially for tourists with disabilities. The application of these technologies ensures that tailored services are provided to diverse populations, making tourism more inclusive [132]. This approach not only broadens the accessibility of tourism activities but also promotes an equitable tourism ecosystem. For instance, the system can be implemented for blind tourists, allowing them to use voice input to develop personalized planning at their destinations [133].

Enhances tourist satisfaction and social responsibility: The paper outlines a scalable model for employing AI, NLP, and the IoT that can be adapted beyond the tourism industry to other sectors aiming to enhance customer satisfaction while balancing environmental and social responsibilities [129,130]. This cross-industry applicability, such as in retail commerce, logistics, and smart transportation, is significant [131].

Overall, this paper enriches the understanding of smart tourism technologies by showcasing their potential to transform tourism into a more adaptive, inclusive, and sustainable sector. The research findings contribute not only to academic knowledge but also offer practical insights that can be implemented by industry practitioners worldwide [132,133].

6. Conclusions

The tourism industry is undergoing a transformative revolution driven by the rapid advancement and integration of digital technologies. AI, NLP, and the IoT are at the forefront of this change, significantly influencing the sector by making tourism more accessible, personalized, and environmentally sustainable [32,95,105].

The research problem focuses on understanding how the synergistic use of Gen AI and the IoT can effectively meet the diverse needs of all tourists, including those with disabilities, and how these technologies can be implemented to ensure ethical, sustainable, and universally accessible tourism practices [28,29]. The identified research gap reveals a lack of innovation in integrating digital technologies such as AI, NLP, and the IoT within the tourism industry [30,31,32,33,34,35]. To address this gap, the objective of this paper is to demonstrate how the integration of Gen AI, NLP, and the IoT can significantly enhance tourism planning for people with disabilities and increase inclusivity in smart tourism destinations. Moreover, this paper aims to propose a framework that not only enhances the accessibility and personalization of tourism experiences but also addresses the socio-technical challenges associated with these advanced technologies, ensuring tourism is inclusive for all individuals [134,135].

The system was developed by applying Agile methodology for employing advanced AI algorithms alongside NLP, and systems capable of generating predictive analytics, personalized recommendations, and conducting real-time, multilingual communication with tourists. The IoT’s real-time data collection capabilities are utilized to monitor resource usage, environmental impacts, and visitor behaviors across various destinations.

This paper makes substantial contributions to the field of smart tourism by demonstrating how the integration of generative AI, NLP, and the IoT supports the decision making of tourists. The research offers both practical solutions and theoretical insights that enhance the accessibility, personalization, and environmental sustainability of tourism [31,38,85,125,126].

The findings reveal that this integrated technological approach not only improves resource management but also significantly enhances the tourist experience by providing tailored services that cater to the needs of diverse populations, including individuals with disabilities. Moreover, the implementation of these technologies aids in the preservation of natural habitats and promotes cultural sensitivity, contributing to the sustainability and ethical responsibility of the tourism sector. The paper outlines the technical implementation, design considerations, and challenges encountered in creating a more equitable and intelligent tourism ecosystem. It also discusses the ethical dimensions related to data privacy and security, emphasizing the importance of responsible technology deployment [126].

By highlighting the role of cutting-edge technologies in redefining tourism practices, this study proposes a model that balances consumer satisfaction with environmental stewardship and social inclusivity. This model suggests its applicability in other industries seeking similar goals, demonstrating the cross-industry potential of integrating AI, NLP, and the IoT to achieve sustainable and inclusive outcomes [31,38,85].

Additionally, this research study has a limitation, as all content is still within the context of Thailand. Furthermore, the smart wristband’s functionality is currently limited in serving blind tourists. Therefore, for future research, the researchers plan to improve the smart wristband’s ability to accommodate elderly travelers and those who are visually impaired. There are also plans to develop NLP and Gen AI modules that work effectively with Thai sounds and sentences.

Overall, the research demonstrates the potential of combining AI, NLP, and IoT technologies to enhance the tourism experience by providing an intelligent and interactive platform that meets the needs of modern tourists [127]. The findings support the feasibility and effectiveness of such a system, paving the way for future advancements in smart tourism technologies [118,128,129,130,131,132,133,136,137].