Artificial Intelligence in Tourism Businesses: Financial Resilience, Organisational Adaptation and Performance Drivers—A Systematic Literature Review

Abstract

Artificial intelligence (AI) is reshaping tourism businesses by improving decision making, service personalization, operational efficiency, and data-driven management. Beyond these organizational benefits, AI may also strengthen firms’ capacity to cope with market volatility, demand shocks, cost pressures, and other sources of financial fragility. This study provides a systematic literature review and bibliometric analysis of 146 Web of Science articles on AI in tourism published between 2019 and 2023. Following a structured screening process, it identifies the intellectual structure, thematic evolution, and main performance-related drivers associated with AI adoption. The findings show a rapidly expanding field centered on business performance, information technology, big data, robotics, and AI-enabled service innovation. The literature suggests that AI contributes to resilience by enhancing forecasting, resource allocation, customer management, and organizational adaptability under uncertainty. However, explicitly financial perspectives—such as financial vulnerability, resilience, liquidity, solvency, and risk management—remain underdeveloped. This study contributes by reframing AI in tourism as a potential resilience-building capability rather than only a tool for service innovation. Its main limitations are the reliance on Web of Science and a fixed 2019–2023 bibliometric corpus, which future research should extend.

1. Introduction

In the rapidly evolving digital economy, tourism companies must adapt to competitive environments by adopting technologies like AI to enhance responsiveness, agility, and efficiency (Sestino et al., 2023; Yousaf et al., 2021; Kowalkiewicz et al., 2017). AI optimises operational processes, reduces costs, and enables companies to adjust to market demands effectively.

AI plays a critical role in transforming organisational operations and improving economic outcomes. Its capacity to analyse vast datasets and generate accurate predictions allows companies to anticipate market changes and make real-time, informed decisions (Mikalef & Gupta, 2021). By automating routine tasks, AI reallocates resources toward strategic activities, fostering operational efficiency and innovation, which provide sustainable competitive advantages (Sjödin et al., 2021; Elkhwesky et al., 2024).

AI significantly impacts innovation and knowledge management, transforming processes and enhancing human capabilities (Cui et al., 2024; Martens, 2024). This results in faster problem-solving, improved economic performance, and optimised production processes (von Krogh, 2018; Füller et al., 2022; Ersoy & Ehtiyar, 2023; Jarrahi et al., 2023). In tourism, AI’s integration with Big Data and IoT supports personalised experiences and dynamic demand management, increasing customer satisfaction while helping companies adapt to evolving trends (Carlisle et al., 2021; Giotis & Papadionysiou, 2022; Lv et al., 2022; Yoon & Choi, 2023; Yu et al., 2023; Zhu et al., 2023).

Bibliometric analyses in tourism research have emphasised smart destinations and communication technologies (Bastidas-Manzano et al., 2021; Molina-Collado et al., 2022; Yuan et al., 2019), as well as machine learning (Núñez et al., 2024), mobile applications (S. Chen et al., 2020), and augmented reality (Ariza-Colpas et al., 2023). Despite these advances, limited attention has been given to how technology impacts tourism companies’ performance, particularly in managing Big Data, integrating AI, and adopting IoT. These technologies are crucial for transforming processes, analysing trends, and enabling informed decision-making to gain competitive advantages.

In addition to its operational and strategic implications, AI adoption in tourism businesses can also be interpreted through the lens of financial fragility and resilience. Tourism is highly exposed to external shocks, including demand volatility, inflationary pressures, labor shortages, and abrupt market disruptions. In this context, AI-based capabilities may help firms improve forecasting, optimize resource allocation, personalize services, and respond more rapidly to uncertainty, thereby strengthening adaptive capacity and supporting more resilient business performance.

Accordingly, financial resilience is approached in this study as a firm’s capacity to anticipate, absorb, adapt to, and recover from financial and operational disruptions while maintaining business continuity, thereby linking the academic debate on AI-enabled capabilities with practical concerns for tourism managers.

Despite the growing literature on AI in tourism, limited attention has been paid to how these digital capabilities relate to financial resilience, risk management, and the ability of tourism businesses—particularly smaller firms—to absorb shocks and sustain performance over time. Addressing this gap is especially relevant in light of current debates on financial vulnerability and resilience in households, investors, and small businesses. Accordingly, this study reinterprets the tourism AI literature not only as a stream of technological and organizational innovation, but also as an emerging source of evidence on resilience-building under uncertainty.

This study addresses the following questions:

Q1. What is the research interest in AI for innovative management in tourism businesses?

Q2. What are the main topics, knowledge structures, and trends in the literature?

Q3. What economic, organisational, and resilience-related factors shape the adoption and effects of AI in tourism businesses?

Q4. What implications, challenges, and future research directions emerge regarding financial fragility, adaptive capacity, and long-term resilience in tourism businesses?

This paper explores AI not only as an innovation paradigm in tourism management but also as a potential driver of organizational adaptation and financial resilience. Through a systematic literature review and bibliometric analysis, the study identifies the main research streams, thematic trends, and knowledge structures in the field, and discusses their implications for business performance, risk management, and resilience under uncertainty.

2. Theoretical Background

This review is theoretically grounded in three complementary perspectives. First, the resource-based view explains AI as a strategic digital resource that may generate value when combined with data, human skills, organisational routines, and technological infrastructure. Second, the dynamic capabilities perspective helps explain how tourism firms use AI to sense market changes, seize opportunities, and reconfigure resources under uncertainty. Third, socio-technical and technology–organisation–environment perspectives show that AI adoption depends on the interaction between technological readiness, organisational capabilities, employee skills, managerial support, and external pressures. Together, these perspectives frame AI as a driver of operational performance, financial outcomes, risk management, and resilience.

AI is transforming tourist behaviour and company dynamics, impacting profitability through improved marketing strategies and services (Mikalef & Gupta, 2021; Sjödin et al., 2021). Adopting AI requires employee training and process adjustments for greater efficiency. Traditional models of technology acceptance fall short in capturing AI’s advanced features, prompting research on anthropomorphism, emotions, and hedonic motivation (Cui et al., 2024; Martens, 2024). Machine learning enhances demand forecasting, sentiment analysis, and personalised experiences, optimising resource use, reducing costs, and boosting customer loyalty and revenue.

2.1. AI-Related Technology in the Tourism Industry

AI, introduced into business processes in the early 21st century, is based on software that mimics human cognitive functions. Its rapid development has been driven by the Fourth Industrial Revolution, marked by the adoption of intelligent and autonomous systems powered by big data analysis. These systems extract value through advanced data capture and processing, transforming business operations (Tussyadiah, 2020; Borges et al., 2021; Damioli et al., 2021).

In the tourism sector, big data is generated through user interactions, including online content (text, photos), electronic devices (e.g., GPS, Wi-Fi, Bluetooth), and transaction records (e.g., online bookings, customer cards). This data provides valuable insights for optimising services and enhancing customer experiences (J. Li et al., 2018).

AI, as a form of smart technology, encompasses various subfields, such as knowledge representation, machine learning, decision-making, and optimisation (Latah & Toker, 2019). Within organisations, AI performs diverse tasks, ranging from mechanical and analytical to intuitive and empathetic functions, enhancing operational efficiency and customer engagement (Huang & Rust, 2018).

A notable application of AI in tourism is the use of service robots equipped with advanced systems to autonomously deliver services. These robots are categorised as quasi-automated (following pre-programmed routines) or fully automated (adapting decisions based on real-time data). Fully automated robots, often designed with human-like features, are particularly relevant in tourism, performing complex cognitive tasks and enriching human-robot interaction in service delivery (J.-M. Li et al., 2023; Fang et al., 2023; Goel et al., 2022; Lu et al., 2019). Examples include Hilton’s robot “Connie” and the humanoid robots used by Henn na Hotel for customer check-in and assistance (Park, 2020; Ersoy & Ehtiyar, 2023).

The Internet of Things (IoT) further complements AI and robotics by facilitating seamless data exchange among devices (e.g., sensors, smartphones, GPS). IoT generates user-driven data that enhances decision-making, improves tourist experiences, and supports service customisation (Bi & Liu, 2022; Miskiewicz, 2020). By integrating with AI and Big Data, IoT creates smart environments, transforming industry practices and enabling real-time adaptability.

These advancements highlight the co-creation of value through innovative services, driving transformative changes in the tourism sector and shaping new opportunities for sustainable growth (Robina-Ramírez et al., 2023).

2.2. Organisational Performance

Numerous studies have evaluated business performance in tourism companies (Morrison & Teixeira, 2004; Robina-Ramírez et al., 2022). Research has explored networking’s impact on performance (Ramayah et al., 2011; Sánchez-Oro Sánchez et al., 2021), innovation (Martínez-Román et al., 2015; Kitsios & Grigoroudis, 2020), intellectual capital (Costa et al., 2020), technology-based communication (Ashari et al., 2014), financial conditions (M.-H. Chen, 2007), and performance comparisons between company sizes (Reichel & Haber, 2005).

Technological advancements, particularly AI, significantly influence tourism organisations’ performance. AI integration among staff enhances technical skills and knowledge acquisition. Continuous training and collaboration between employees and AI are essential for leveraging digital tools like IoT, Big Data, and robotics to create business value (Barro & Davenport, 2019; Peter et al., 2020; Sharma et al., 2021; Marino-Romero et al., 2024a).

Organisational capabilities also depend on effective communication, formal relationships, and agile decision-making. Creating a digital mindset through matrix or decentralised structures allows adaptability in dynamic markets (Giotis & Papadionysiou, 2022). Management plays a vital role in fostering collaborative networks and open innovation to embed AI-related competencies (Marino-Romero et al., 2024b).

Performance measurement requires both financial and non-financial indicators to assess competitiveness (AlMujaini et al., 2021; Marino-Romero et al., 2022). Emerging technologies like AI positively impact internal factors such as productivity, employee satisfaction, and work environment (Castellacci & Viñas-Bardolet, 2019; Prentice et al., 2020b; Wirtz et al., 2019). Financial metrics, such as sales, costs, and profit margins, also highlight AI’s contribution to business success (Sharma et al., 2021; Henriques & Pereirsa, 2024).

AI’s role extends to external performance during service delivery, where customer experience shapes satisfaction levels. Studies show that perceived quality and value depend on AI’s ease of use and usefulness in service processes. Complex interactions requiring excessive effort may lead to unfavourable emotional responses (Lei et al., 2023; Hadinejad et al., 2021).

Environmental factors, such as market competition and regulatory pressures, also drive AI adoption. These forces compel organisations to adapt, improving market share and innovating services to meet customer demands (Mikalef & Gupta, 2021; Y. Chen et al., 2022).

Optimising performance indicators requires robust infrastructure and skilled personnel to adapt AI to diverse tourism sectors, including hospitality, travel agencies, and airports (Prentice et al., 2020a; Sharma et al., 2021).

2.3. AI, Financial Fragility and Resilience in Tourism Businesses

Tourism businesses operate in highly uncertain environments shaped by seasonality, macroeconomic volatility, changing consumer preferences, and unexpected disruptions. Under these conditions, financial fragility may emerge through unstable revenues, cost pressures, limited liquidity, and constrained adaptive capacity, especially among smaller firms.

In this study, operational performance refers to the efficiency and effectiveness with which tourism businesses transform resources into service, organisational, and economic outcomes. Financial fragility refers to firms’ exposure to revenue instability, liquidity constraints, cost pressures, debt commitments, and limited shock-absorption capacity. Financial resilience is understood as the ability to anticipate, absorb, adapt to, and recover from financial and operational disruptions while maintaining business continuity. Risk management refers to the processes through which firms identify, assess, and mitigate uncertainties affecting operations, revenues, costs, and viability. These concepts are related but distinct: AI adoption may enhance operational performance and risk management, thereby reducing fragility and strengthening financial resilience.

More specifically, AI may affect profitability through revenue optimisation and cost reduction, liquidity through improved demand forecasting and cash-flow planning, and solvency through better risk assessment, resource allocation, and early detection of financial pressures.

From this perspective, AI adoption can be understood as a strategic capability that supports resilience by improving forecasting, automation, and data-driven decision making. However, the literature still focuses mainly on service innovation, customer experience, and operational performance, while financial vulnerability and long-term adaptive capacity remain underexplored.

3. Materials and Methods

This study adopts a systematic literature review (SLR) approach to ensure the reliability and validity of its findings and to examine how the literature on AI in tourism addresses not only organizational performance, but also business adaptation, risk-related challenges, and resilience under uncertainty. The SLR methodology allows for an unbiased identification of relevant scientific articles on technology-based information systems in the tourism sector (Kitchenham et al., 2010; Okoli & Schabram, 2010). This review was conducted in accordance with the PRISMA guidelines for evidence synthesis, and the study selection process is summarized in Figure 1. The review protocol was not prospectively registered in a public repository. The process is structured into four key stages, enabling rigorous, transparent, and reproducible data collection.

The first stage involves planning the review to address the economic, organizational, and resilience-related factors associated with AI application in tourism businesses. This is guided by four research questions (Q1, Q2, Q3, Q4) outlined in the introduction. These questions shape the review process and synthesise key insights into AI’s impact on tourism organisations, which are further discussed in Section 5 and Section 6.

The second stage involves conducting a literature review using the Web of Science (WOS) database, known for its extensive bibliographic resources and quality index (Durán-Sánchez et al., 2019; Abarca et al., 2020). WOS is widely utilised in social sciences, particularly in technology and tourism studies (Núñez et al., 2024; Elkhwesky et al., 2024). The search terms “AI”, “tourism,” and “performance” were refined using Boolean operators to narrow results (date of last search 4 December 2025).

Web of Science was retained as the single bibliographic source to ensure consistency in indexing standards, citation metadata, subject-category filtering, and bibliometric reproducibility. Scopus was not incorporated in the formal corpus in order to avoid database overlap, duplicate records, and inconsistencies in citation and keyword metadata that would have required a separate harmonisation procedure. A completed PRISMA 2020 checklist was used to ensure that all reporting items are properly addressed in the manuscript.

To improve transparency and reproducibility, the complete search strategy is now reported. The search was conducted in the Web of Science Core Collection using topic fields, including title, abstract, author keywords, and Keywords Plus. The core search equation combined three mandatory blocks: artificial intelligence, tourism, and performance. In addition, a fourth block related to financial resilience, fragility, and risk management was used as an analytical coding block to assess the representation of this dimension within the final dataset. This fourth block was not imposed as a mandatory search restriction, because doing so would have excluded relevant studies on AI adoption in tourism that discuss resilience-related mechanisms indirectly, such as demand forecasting, dynamic pricing, cost reduction, operational flexibility, or business continuity.

The search equation used to retrieve the initial corpus was as follows:

TS = ((“artificial intelligence” OR “AI” OR “machine learning” OR “deep learning” OR “robot*” OR “service robot*” OR “big data” OR “smart technolog*” OR “intelligent system*” OR “Internet of Things” OR “IoT”) AND (“tourism” OR “hospitality” OR “hotel*” OR “travel” OR “tourist*” OR “destination*”) AND (“performance” OR “business performance” OR “organisational performance” OR “organizational performance” OR “financial performance” OR “firm performance” OR “productivity” OR “competitiveness” OR “efficiency” OR “innovation” OR “service quality” OR “customer satisfaction”)).

The financial resilience, fragility, and risk-management block was subsequently used during the content analysis stage to classify the final articles according to their explicit or implicit connection with financial vulnerability and resilience. This block included terms such as “financial resilience”, “financial fragility”, “financial vulnerability”, “risk management”, “liquidity”, “solvency”, “profitability”, “cost reduction”, “revenue volatility”, “shock”, “crisis”, “business continuity”, “adaptive capacity”, and “resilience”.

Table 1. Keyword blocks used in the search and content analysis.

Keyword Block	Function in the Study	Search Terms/Coding Terms
Artificial intelligence block	Mandatory search block	“artificial intelligence”, “AI”, “machine learning”, “deep learning”, “robot*”, “service robot*”, “big data”, “smart technolog*”, “intelligent system*”, “Internet of Things”, “IoT”
Tourism block	Mandatory search block	“tourism”, “hospitality”, “hotel*”, “travel”, “tourist*”, “destination*”
Performance block	Mandatory search block	“performance”, “business performance”, “organisational performance”, “organizational performance”, “financial performance”, “firm performance”, “productivity”, “competitiveness”, “efficiency”, “innovation”, “service quality”, “customer satisfaction”
Financial resilience/fragility/risk block	Analytical coding block applied to the final dataset	“financial resilience”, “financial fragility”, “financial vulnerability”, “risk management”, “liquidity”, “solvency”, “profitability”, “cost reduction”, “revenue volatility”, “shock”, “crisis”, “business continuity”, “adaptive capacity”, “resilience”

Source: Author’s own elaboration.

summarises the keyword blocks used in the search strategy and content analysis. The first three blocks were mandatory search components, whereas the financial resilience/fragility/risk block was used during coding to identify explicit or implicit links with financial vulnerability, adaptive capacity, and resilience.

Exclusion criteria were applied to focus on peer-reviewed journal articles in English within the “business economics” domain from 2019 to 2023. This period was selected to reflect the topic’s novelty and rapid research growth. After filtering, 184 articles were validated (Figure 1), providing a focused dataset aligned with the research objectives.

To further improve the transparency and reproducibility of the systematic literature review,

Table 2. Transparency criteria of the systematic literature review.

Element	Description
Database	Web of Science Core Collection
Search date	4 December 2025
Search field	Topic search, including title, abstract, author keywords, and Keywords Plus
Search query	TS = ((“artificial intelligence” OR “AI” OR “machine learning” OR “deep learning” OR “robot*” OR “service robot*” OR “big data” OR “smart technolog*” OR “intelligent system*” OR “Internet of Things” OR “IoT”) AND (“tourism” OR “hospitality” OR “hotel*” OR “travel” OR “tourist*” OR “destination*”) AND (“performance” OR “business performance” OR “organisational performance” OR “organizational performance” OR “financial performance” OR “firm performance” OR “productivity” OR “competitiveness” OR “efficiency” OR “innovation” OR “service quality” OR “customer satisfaction”))
Filters applied	Peer-reviewed journal articles; English language; Business Economics category; publication period 2019–2023
Inclusion criteria	Articles addressing AI-related technologies in tourism or hospitality businesses and their relationship with performance, innovation, management, or organisational outcomes
Exclusion criteria	Non-peer-reviewed documents; non-English documents; records outside the selected period or subject category; articles unrelated to AI, tourism, or business performance after title, abstract, and keyword screening
Screening stages	Initial WOS search; application of database filters; title, abstract, and keyword screening; exclusion of unrelated articles; final validation of 146 articles
Initial records after filtering	184 articles
Final dataset	146 articles
Coder count	Three authors independently participated in the coding process.
Coder agreement	Inter-rater reliability was assessed using Cohen’s Kappa coefficient (κ = 0.78), indicating substantial agreement among coders.
Conflict resolution rule	Disagreements were discussed among the authors until consensus was reached
Software used	Biblioshiny/Bibliometrix (bibliometrix R package, version 4.1.4; RStudio version 2023.06.1) for bibliometric analysis

Source: Author’s own elaboration.

summarises the main methodological decisions applied during the search, screening, coding, and synthesis stages. The table reports the database, search date, search query, filters, inclusion and exclusion criteria, screening process, coder involvement, agreement procedure, and conflict resolution rule.

In the third stage, data coding was applied by analyzing the titles, abstracts, and keywords of all manuscripts. Thirty-eight unrelated articles were excluded, leaving 146 validated for bibliometric analysis. Excluded records typically included papers focused on tourism without an AI-related technology, AI applications outside tourism or hospitality, purely technical studies without business or performance outcomes, and documents unrelated to organisational, financial, or resilience-related implications. To ensure methodological rigour, inter-rater reliability was assessed using Cohen’s Kappa coefficient. The calculation was performed using RStudio (version 2023.06.1) with the vcd statistical package. Bibliometric analyses were conducted using the packages bibliometrix (version 4.1.4), yielding a value of κ = 0.78. According to the criteria established by Landis and Koch (1977), this result indicates a substantial level of agreement. The fourth stage employed bibliometrics through the Biblioshiny tool, evaluating scientific productivity (sources, authors, countries, documents) and mapping conceptual trends (thematic evolution, co-occurrence, and keyword factor analysis), offering insights into knowledge structures (Merigó et al., 2016; Aria & Cuccurullo, 2017).

To increase the transparency and reproducibility of the qualitative synthesis, an explicit content analysis protocol was applied to the final set of 146 articles. The unit of analysis was each individual article included in the final dataset. For each paper, the title, abstract, author keywords, Keywords Plus, and, when necessary, the full text were examined. The coding process was guided by a structured codebook designed to connect the bibliometric results with the research questions of the study. The codebook included four main coding dimensions: (1) AI-related technologies, including artificial intelligence, machine learning, robotics, big data, smart technologies, and IoT; (2) tourism business context, including hospitality firms, hotels, travel agencies, tourism destinations, and other tourism-related organisations; (3) performance-related outcomes, including operational efficiency, service quality, customer satisfaction, innovation, productivity, competitiveness, and financial performance; and (4) financial resilience, fragility, and risk-management implications, including liquidity pressures, revenue volatility, cost reduction, vulnerability to shocks, adaptive capacity, and business continuity.

The coding logic combined deductive and inductive procedures. First, deductive categories were derived from the research questions and from the theoretical background on AI-related technologies, organisational performance, and financial resilience. Second, inductive codes were added when recurrent themes emerged from the reviewed documents. For example, studies focused on service robots and customer interaction were coded under “AI-related technologies” and “service performance”, whereas studies addressing forecasting, demand prediction, dynamic pricing, or resource optimisation were additionally coded as having potential implications for “financial resilience” or “risk management”. Papers referring explicitly to profitability, costs, liquidity, financial performance, vulnerability, or shock absorption were classified as directly related to the financial dimension, while papers discussing adaptability, operational flexibility, or business continuity were classified as indirectly related to resilience.

The synthesis was then conducted by comparing the coded categories with the bibliometric outputs generated through Biblioshiny. This procedure made it possible to distinguish between dominant themes in the field, such as performance, information technology, big data, robotics, and service innovation, and less developed but emerging themes related to financial fragility, resilience, and risk management. To ensure consistency, disagreements in category assignment were discussed among the authors until consensus was reached. The resulting coding framework, including categories, operational definitions, inclusion rules, and illustrative examples, is fully described within the manuscript.

Table 3. Content analysis protocol and coding framework.

Coding Element	Description	Illustrative Examples
Unit of analysis	Individual article included in the final dataset of 146 records.	Each WOS-indexed article retained after screening.
Textual fields analysed	Title, abstract, author keywords, Keywords Plus, and full text when required.	Papers with ambiguous abstracts were checked in full text.
AI-related technologies	Identification of the main technological focus of each paper.	AI, machine learning, robotics, big data, IoT, smart technologies.
Tourism business context	Identification of the tourism setting or organisational unit analysed.	Hotels, hospitality firms, travel agencies, destinations, tourism SMEs.
Performance outcomes	Identification of business, organisational, service, or economic outcomes.	Operational efficiency, customer satisfaction, productivity, innovation, competitiveness, financial performance.
Financial resilience, fragility, and risk-management implications	Identification of direct or indirect links with vulnerability, risk, or shock absorption.	Liquidity pressures, revenue volatility, cost reduction, demand forecasting, dynamic pricing, business continuity, adaptive capacity.
Coding approach	Combination of deductive categories derived from the research questions and inductive codes emerging from the reviewed literature.	Forecasting and resource optimisation were inductively linked to resilience-related implications.
Synthesis rule	Coded categories were compared with bibliometric outputs to interpret dominant and emerging themes.	Co-occurrence and thematic evolution results were interpreted in relation to performance and resilience categories.
Consistency procedure	Disagreements in coding were discussed among the authors until consensus was reached.	Ambiguous papers were reassessed jointly.

Source: Author’s own elaboration.

summarises the content analysis protocol, including the unit of analysis, textual fields examined, coding dimensions, synthesis rule, and consistency procedure.

4. Results

The results are presented by combining bibliometric indicators with the categories obtained from the content analysis protocol described above. This combined approach allows the study to identify not only the most productive sources, authors, and countries, but also the thematic categories through which AI adoption in tourism has been linked to organisational performance, business adaptation, risk management, and financial resilience.

In order to assess the extent to which the financial resilience dimension was represented in the dataset, the 146 articles were classified into three categories. The first category included papers with a direct financial resilience focus, namely studies explicitly referring to financial performance, financial vulnerability, financial fragility, liquidity, solvency, profitability, risk management, or shock absorption. The second category included papers with indirect resilience-related implications, namely studies dealing with demand forecasting, dynamic pricing, cost reduction, operational flexibility, resource optimisation, crisis response, or business continuity, even when they did not explicitly use the term financial resilience. The third category included papers with no explicit or implicit connection to financial resilience, fragility, or risk management.

Table 4. Representation of the financial resilience dimension in the final dataset.

Category	Classification Rule	Examples of Terms or Themes
Direct financial resilience focus	Articles explicitly addressing financial performance, vulnerability, fragility, liquidity, solvency, profitability, risk management, or shock absorption.	financial performance; financial vulnerability; liquidity; profitability; risk management; shock absorption
Indirect resilience-related implications	Articles discussing mechanisms that may support resilience without explicitly framing them as financial resilience.	demand forecasting; dynamic pricing; cost reduction; resource optimisation; operational flexibility; business continuity; crisis response
No explicit resilience connection	Articles focused mainly on technological adoption, customer experience, service quality, or innovation without financial or resilience-related implications.	AI acceptance; customer satisfaction; service robots; smart tourism; user experience
Total	Final dataset retained after screening.	146 articles

Source: Author’s own elaboration.

summarises these categories, the classification rules applied, and illustrative terms or themes used to identify each type of connection with financial resilience.

This classification shows that the financial resilience dimension is only partially represented in the dataset. While several studies refer to financial performance, cost reduction, forecasting, revenue management, or operational flexibility, relatively few papers explicitly address financial fragility, liquidity, solvency, risk exposure, or shock absorption in tourism businesses. This confirms the gap identified in the review: AI-related research in tourism has mainly focused on technological adoption, customer experience, service innovation, information technology, and operational performance, whereas explicitly financial interpretations of resilience remain fragmented and underdeveloped.

The bibliometric study analyses quantitatively the research production on the use of AI in the organisational management of tourism companies. Figure 2 shows the main research results in which 146 articles from 92 different sources were identified, with an average of 28.35 citations per article. In this study, 9364 articles have been referenced with 621 keywords and a participation of 477 authors with an average collaboration of 3.48 co-authors per article and 43.48% of collaborations between authors from different countries. These five years of study show an upward research trend with an average annual growth of 40.04%. Publication bias was assessed by visual inspection of funnel plots and the Egger’s test, with a p-value of <0.05 considered statistically significant. All analyses were carried out using RStudio (version 2023.06.1), employing the ‘metafor’ (version 4.4-0) and ‘meta’ (version 7.0-0) statistical packages.

Next, the results will be analysed according to Section 4 of the scheme shown in Figure 1, synthesising the information to the most important aspects that guarantee scientific evidence. The first analysis is focused on the metrics of scientific production, and provides an overview of the resources, authors, publications, citations and collaborations referred to in the metadata of the articles. The second analysis will focus on conceptual knowledge structures, identifying internal gaps and future research directions on the effects of AI on the discipline of organisational management in the tourism industry.

4.1. Production Based-Measurements

4.1.1. Resources

The journals are going to be organised according to productivity by zones of importance ordering them in a decreasing order through Bradford’s law (Desai et al., 2018).

The shaded area in Figure 3 highlights 12 key journals, with four particularly important ones: International Journal of Contemporary Hospitality Management, Sustainability, Journal of Hospitality Marketing & Management, and International Journal of Hospitality Management. These journals offer innovative insights on AI-related technologies in tourism management. Surprisingly, Sustainability, though not tourism-focused, plays a significant role in knowledge dissemination.

4.1.2. Most Relevant Authors and Most Cited Documents

The most relevant authors within the research area are considered to be those who have published three or more articles in the period considered, with a greater contribution to the progress of science. According to the data in Figure 4 these authors are: Dogan Gursoy from Washington State University USA, Biao Huang from the University of Alberta Canada, Catherine Prentice University of Southern Queensland Australia and Yao-Chin Wang from the University of Florida USA.

As this is an incipient field of research, there is a wide dispersion in the authorship of the articles, since 95% of all authors (453) wrote a single co-authored article (see

Table 5. Author productivity.

Documents Written	No. of Authors	Proportion of Authors
1	453	95.0%
2	20	4.2%
3	3	0.6%
6	1	0.2%

Source: Author’s own elaboration. Note: Lotka’s Law is applied to calculate the frequency of publications.

).

4.1.3. Production and Collaboration by Countries

In order to analyse the most productive countries, their geographical distribution will be shown, based on publications and the map of collaborations between countries.

Table 6. Most productive countries.

Country	Articles	Articles %	SCP	MCP	MCP %
China	53	36.3	34	19	35.8
Usa	13	8.9	5	8	61.5
United Kingdom	8	5.5	4	4	50
India	6	4.1	5	1	16.7
Italy	6	4.1	3	3	50
Korea	6	4.1	5	1	16.7
Germany	5	3.4	3	2	40
Spain	5	3.4	3	2	40
Australia	4	2.7	2	2	50
Saudi Arabia	4	2.7	3	1	25

Source: Author’s own elaboration.

shows the number of articles published by author affiliation and their willingness to cooperate with other countries using two indicators: SCP (showing the number of articles co-authored within the same country) and MCP (reflecting the number of articles co-authored with other countries). Thus, the most productive country by author affiliation is led by China with 53 manuscripts, followed by the USA (13) and the United Kingdom (8). The countries that collaborate most are USA (CCM = 61.5%), China (CCM = 35.8%) and United Kingdom (CCM = 50%). It is shown that the most productive countries are those that associate the most with other countries in a given research field.

Finally, Figure 5 presents the map of collaborations between countries that identifies the network of authors by country, showing the main geographical connections and synergies, which can help scholars to develop international research projects. Figure 5 shows patterns of collaboration among all scientific communities worldwide (minimum number of edges = 5). The most relevant international collaborations due to their recurrence are between the most productive countries: China–United Kingdom, China–USA and USA–United Kingdom. Next, it is worth highlighting the research synergies created in the subject analysed by USA–South Africa and China–Australia.

4.2. Conceptual Structure

Although the dominant themes in the literature are performance, technology, and service innovation, these themes also reveal an emerging concern with how AI-enabled capabilities may strengthen firms’ adaptability and resilience in uncertain and volatile environments.

A map of scientific knowledge will depict dynamic intellectual connections through the conceptual structure (Small, 1997). The Sankey diagram traces AI’s impact on tourism over five years, with theme weighting using a modified inclusion index (Rip & Courtial, 1984). Keyword appearances are analyzed in three stages: 2019–2021 (antecedents) and 2022–2023 (peak) (Figure 6).

Between 2019 and 2021, tourism research focused on terms like system, performance, model, financial performance, and technology (Cai et al., 2019; Loureiro, 2019), aiming to optimise business systems for better efficiency. The industry was particularly interested in financial performance and technology, aiming to maximise economic benefits and enhance customer experiences through innovation (Sugathan & Ranjan, 2019; Azis et al., 2020).

From 2022, tourism evolved towards analysing AI, business models, performance, and technological impact, highlighting the shift to digitalisation and personalisation of experiences. Companies focused on operational efficiency through technology to increase satisfaction for both employees and customers (Azis et al., 2020; Bulchand-Gidumal et al., 2024). AI-based business models became central to improving performance and customer satisfaction.

Key AI technologies in tourism include robotics, IoT, and Big Data, enabled by internet evolution, which supports tourism service distribution and consumption. The industry’s adaptation to new market dynamics underscores its focus on future demands in a competitive, digitalised environment (Widagdo et al., 2024).

Figure 7 presents a co-occurrence network showing keyword relationships in the literature, helping understand interconnections and establish thematic groups (Cruz-Cárdenas et al., 2021).

In recent years, the tourism sector has experienced a significant transformation driven by advanced technologies, a phenomenon often referred to as “technological tourism.” Research highlights “performance” as a critical aspect, encompassing both operational efficiency and the effectiveness of implemented technological solutions (Sujood et al., 2024).

The co-occurrence results indicate that AI is mainly linked to performance, information technology, big data, robotics, and service innovation. Financial resilience is therefore not directly consolidated as an independent theme, but appears indirectly through mechanisms such as demand forecasting, process optimisation, cost reduction, customer management, and operational flexibility. These mechanisms translate the bibliometric evidence into financial implications by linking AI adoption to profitability, liquidity planning, risk management, business continuity, and firms’ capacity to absorb shocks.

Information technology also supports models for analysing and predicting behaviours in tourism. These models, ranging from simulations to advanced analytical structures, help predict consumer behaviour and evaluate the feasibility of innovations (Mao et al., 2021). Their effectiveness relies on user and company acceptance of technology and the type of IT employed.

Recent research underscores how technological systems influence consumption habits by enabling tourists to access information, make reservations, and personalise their trips efficiently. Algorithms predict consumer behaviour and future demand, optimising tourism offerings to align with traveller preferences (Gajdošík, 2019). Factor analysis further identifies research subfields, using correspondence analysis to explore relationships among variables and deepen insights into this evolving field (Abdi & Valentin, 2007).

As shown in Figure 8, the factor analysis highlights key themes influencing trends in tourism research, such as system, performance, impact, information technology, and models (Si-Tou, 2024). These interconnected themes reflect the priorities and challenges faced by tourism companies in adopting AI solutions.

Internal organisation relies on robust technological infrastructure, with “system” and “information technology” emerging as critical elements (Jalilvand & Ghasemi, 2024). AI enables automation, data management, and improved decision-making, ensuring adaptability to evolving sector needs.

Performance, a central focus, measures AI’s ability to achieve business objectives, including operational efficiency, cost reduction, and customer satisfaction (Bulchand-Gidumal et al., 2024). By analysing real-time data, AI enhances responsiveness and drives financial and economic success in the tourism industry.

This suggests that the literature has begun to move beyond purely operational efficiency and toward broader questions of organizational robustness, although explicitly financial resilience remains comparatively underexplored.

5. Discussion

Recent AI innovations have significantly transformed tourism by improving business processes, organisational performance, and value creation. Regarding Q1, the bibliometric analysis shows a clear increase in AI-related tourism research since 2019, with 146 articles, 9364 references, and an annual growth rate of 40.04%, confirming that this is an expanding research field.

Although the field remains emerging and fragmented, several authors have made notable contributions, particularly D. Dogan Gursoy, D. Biao Huang, D. Yao-Chin Wang, and Catherine Prentice. The most influential journals include the International Journal of Contemporary Hospitality Management, Sustainability, Journal of Hospitality Marketing & Management, and International Journal of Hospitality Management, which disseminate research on AI applications in tourism (Huang & Rust, 2018; Latah & Toker, 2019). China, the USA, and the UK lead this literature, reflecting global interest in AI technologies such as service robots, IoT, and big data (J. Li et al., 2018).

The reviewed literature shows that AI improves organisational performance by enhancing productivity, employee satisfaction, profitability, technical skills, continuous learning, and human–AI collaboration (Wirtz et al., 2019; Prentice et al., 2020b; Barro & Davenport, 2019; Peter et al., 2020). Overall, AI supports business model innovation, sustainable growth, and competitiveness in tourism (Mikalef & Gupta, 2021; Y. Chen et al., 2022).

Regarding Q2, the findings confirm that AI has become central to tourism’s digital transformation, particularly since 2022, when research increasingly linked AI to business management, operational efficiency, customer engagement, and service quality (Tussyadiah, 2020; Damioli et al., 2021). AI systems supported by big data allow firms to process user-driven information from online interactions, GPS, Wi-Fi, and transactions, improving service optimisation and customer experience (J. Li et al., 2018). These capabilities are consistent with a resource-based view, where digital infrastructure, data, and human expertise jointly contribute to business value (Dwivedi et al., 2021; Lv et al., 2022).

AI also affects both financial and non-financial performance by supporting decision-making, optimisation, knowledge management, and service transformation (Latah & Toker, 2019). Examples such as Hilton’s “Connie” and Henn na Hotel’s humanoid robots illustrate how AI-based solutions can improve customer interaction, efficiency, and service innovation (Park, 2020; Goel et al., 2022).

Additionally, AI’s role in service delivery intersects with the growing demand for customer-centric strategies. Collaboration between AI and human employees fosters effective service experiences (Lei et al., 2023). IoT-enabled devices generate real-time data to customize services, highlighting the need for a digital mindset and adaptability in decision-making (Giotis & Papadionysiou, 2022) (Figure 9).

Competitive pressures and regulatory frameworks are key drivers of AI integration in tourism companies. These external forces require firms to adapt their operations and services in an increasingly digitised and AI-driven market, so AI must support both internal performance and responsiveness to external demands (Y. Chen et al., 2022). Integrated with Big Data and IoT, AI enables firms to adapt to market dynamics, align services with customer expectations, and create smart environments through real-time data exchange and service customisation (Miskiewicz, 2020; Robina-Ramírez et al., 2023). However, these benefits should be interpreted critically, as AI adoption may also involve implementation and maintenance costs, technological dependence, data-governance risks, and financial pressures, particularly for smaller tourism firms with limited liquidity and weaker absorptive capacity.

Regarding Q3, tourism has been strongly shaped by AI tools associated with the Fourth Industrial Revolution, including machine learning, Big Data, IoT, robotics, and optimisation systems. These technologies improve organisational performance by automating tasks, supporting decision-making, personalising services, anticipating demand, reducing costs, and enhancing customer experiences.

From a financial management perspective, AI should be viewed not only as a tool for service innovation, but also as a resilience-enabling capability. By improving forecasting, revenue management, resource optimisation, and responsiveness to shocks, AI may help tourism businesses reduce fragility, sustain performance, and strengthen financial robustness and adaptive capacity.

Recent post-2023 literature reinforces the relevance of AI for tourism business performance, competitiveness, and resilience. Recent studies show that AI is reshaping hospitality and tourism through data-driven personalisation, predictive customer care, revenue management, operational optimisation, and strategic decision support, while also raising challenges related to implementation costs, digital skills, technological infrastructure, privacy, ethics, and governance (Bulchand-Gidumal et al., 2024; Florido-Benítez & del Alcázar Martínez, 2024; López-Naranjo et al., 2025). Other recent contributions further argue that tourism research should move beyond narrow views of AI as a customer-service or marketing tool and consider its broader organisational, economic, sustainability, and risk-related implications (Hall & Cooper, 2025; Gössling & Mei, 2025).

AI adoption may contribute to financial resilience through four interconnected mechanisms: anticipation, efficiency, adaptation, and risk control. First, AI improves demand forecasting and revenue prediction. Second, it supports automation, cost control, and resource optimisation. Third, it enables faster service redesign, pricing adjustment, and customer personalisation in response to shocks. Fourth, it improves information processing and decision-making under uncertainty. Together, these mechanisms may reduce financial fragility by stabilising revenues, improving liquidity planning, protecting margins, and supporting business continuity, although their effectiveness depends on firms’ digital skills, financial resources, managerial capabilities, and data-governance structures.

The content analysis reported in Table 4 reinforces this interpretation by showing that the financial resilience dimension is present in the dataset, but mainly in an indirect and fragmented way. Most studies do not explicitly frame AI adoption in terms of financial fragility, liquidity, solvency, risk exposure, or shock absorption. Instead, they tend to address mechanisms that may support resilience, such as forecasting, dynamic pricing, cost reduction, operational flexibility, resource optimisation, and business continuity. This finding clarifies the methodological basis of the study’s contribution: the article does not claim that financial resilience is already a consolidated research stream within AI and tourism studies, but rather identifies it as an emerging analytical perspective that can help reinterpret existing evidence on performance, adaptation, and digital transformation.

At the same time, the evidence indicates that the tourism literature still pays limited attention to explicitly financial outcomes such as liquidity pressures, debt exposure, financial vulnerability, and shock absorption capacity. Future research should therefore connect AI adoption more directly with the financial resilience of tourism firms, especially SMEs, in order to clarify when digital transformation strengthens long-term sustainability and when it may instead create new dependencies, costs, or organizational risks.

6. Conclusions

This review shows that AI has become a central driver of transformation in tourism businesses, not only by improving organizational performance but also by reinforcing the conditions for resilience under uncertainty.

The literature indicates that AI-enabled capabilities—such as forecasting, automation, personalization, and data-driven decision making—can strengthen adaptive capacity and support more robust business performance in volatile environments. In this sense, AI should be interpreted as part of a broader resilience-building strategy rather than merely as a tool for technological modernization.

The results and filtering of the bibliometric review offer significant implications and the possibility for future research, which will be addressed in response to question Q₄.

6.1. Theoretical and Practical Implications

Theoretically, this study contributes by moving beyond technical views of AI adoption and reframing AI in tourism as an organisational and resilience-enabling capability. Rather than focusing only on algorithm design or service automation, the review shows how AI-related resources, digital infrastructure, employee skills, and managerial routines interact to shape operational performance, risk management, adaptive capacity, and financial resilience.

The contributions of this study can help tourism industry managers become more competitive by identifying the current challenges of AI, which focus on organisational management and the adoption of new technologies and require constant evaluation of digital strategies and business models. Moreover, it allows public institutions involved in allocating financial support to innovative projects to understand and evaluate the impact that developing AI resources has on key organisational performance indicators.

For managers, the findings highlight that investments in AI may generate value not only through efficiency gains but also through improved preparedness, adaptability, and continuity in the face of disruptions. For policymakers and support institutions, the study suggests that promoting digital capabilities in tourism businesses—especially smaller firms—may contribute to stronger financial resilience, more sustainable competitiveness, and better shock absorption capacity across the sector.

6.2. Limitations and Future Research Directions

Due to the theoretical nature of the study and the selection of data based on journal articles from the WOS platform, our study is not without limitations. Firstly, exploratory studies should be conducted to corroborate the theoretical results of this research, and secondly, our theoretical analysis should be complemented with other SLRs that include new bibliographic databases, such as ScienceDirect, and new documents beyond those considered in this research, such as dissertations, conferences, and books.

In addition, the bibliometric corpus was deliberately restricted to the 2019–2023 period in order to preserve the internal consistency and reproducibility of the search strategy, PRISMA screening process, and bibliometric indicators. Consequently, publications from 2024 onwards were not incorporated into the formal quantitative corpus, although recent studies were used to contextualize emerging developments and should be systematically integrated in future updates of this review.

Future research areas should focus on analysing the technological acceptance by users and companies, the interaction between technological systems and consumption patterns, and the development of analytical tools that help predict behaviours in tourism. Furthermore, research is expected to continue exploring how AI can contribute to improving operational performance, enhancing companies’ ability to anticipate and respond to the needs of an evolving market.

Future research should examine more explicitly the relationship between AI adoption and financial resilience outcomes in tourism businesses, including liquidity management, debt vulnerability, risk exposure, and post-shock recovery capacity. Greater attention should also be paid to tourism SMEs, which may face greater fragility but also derive substantial benefits from data-driven adaptation and digital decision-support systems.