Abstract
The global business landscape is currently experiencing a transformative shift propelled by technological advancements, particularly Artificial Intelligence, which is reshaping operations and enhancing organizational capabilities. This paper investigates the adoption intention of AI within the Small and Medium Enterprises in the United Arab Emirates, a country recognized for its embrace of technological innovation. Addressing a key research gap, the study emphasizes the limited exploration of AI adoption determinants within SMEs in emerging economies. Built on the application of the technology–organization–environment framework and integrating the Technology Acceptance Model, the research identifies factors influencing AI adoption intention in UAE hospitality SMEs. A quantitative survey of 315 respondents was conducted, and data were analyzed using structural equation modeling (SEM) in SmartPLS. The findings reveal significant relationships between perceived competitive advantage, perceived usefulness, perceived top management support, perceived employee capability, perceived competitive pressure, perceived government regulations and AI adoption intention, affirming the validity of the proposed conceptual model. The results further show that organizational (top management support) and environmental factors (competitive pressure, government regulation) exert stronger influence than technological perceptions, highlighting the role of leadership and institutional pressures in shaping adoption. The study contributes theoretically by contextualizing TOE–TAM integration and practically by offering actionable insights for SME leaders, policymakers, and stakeholders to drive AI adoption.
Keywords
Introduction
The business landscape is undergoing a transformative shift, driven by rapid technological advances that redefine operations and augment organizational capabilities (Bergdahl et al., 2023; Sharma, Singh, Islam, & Dhir, 2022). Among these, Artificial Intelligence (AI) has emerged as a disruptive force that not only automates tasks but also enhances efficiency, innovation, and strategic decision-making (Horani et al., 2023; Islam et al., 2021). AI broadly refers to “the ability of machines to perform tasks that normally require human intelligence, such as perception, reasoning, learning, and decision-making” (Dwivedi et al., 2021). AI is now widely recognized as a driver of competitive advantage, enabling businesses to optimize processes, personalize customer experiences, and extract actionable insights from large datasets (Islam, 2026; Sharma, Singh, Gaur, & Afaq, 2022; Rasool et al., 2020). AI has been shown to enhance firm competitiveness, operational efficiency, and innovation by enabling process automation, predictive analytics, and personalized customer engagement (Dwivedi et al., 2021; O. Nazir et al., 2026). For SMEs, adopting AI can translate into improved market positioning and agility, particularly in highly competitive and fast-changing industries (Gupta et al., 2022; Neumann et al., 2023).
Across the world, the United Arab Emirates (UAE) has positioned itself as a hub of AI-led innovation, underlined by the National Strategy for Artificial Intelligence 2031. Small and medium enterprises (SMEs), constituting more than 90% of businesses worldwide and over 94% of firms in the UAE, are central to this transformation (Bloomberg Report, 2023; Forcadell et al., 2021). Their growth is crucial for employment generation and economic diversification, making AI adoption particularly significant for their competitiveness (J. Wang et al., 2022). In the UAE, SMEs contribute over 94% of registered firms and more than 60% of the national GDP, positioning them as a cornerstone of economic diversification and innovation (Dubai SME Report, 2023). Unlike SMEs in many developed contexts, UAE SMEs operate in a unique ecosystem shaped by government-driven digital transformation agendas, mandatory disclosure regulations, and the dynamics of a highly multicultural workforce (Gupta et al., 2022; Poma et al., 2020). These characteristics create both opportunities and challenges for AI adoption, underscoring the importance of investigating this context in depth (Alkhwaldi, 2024; Salhab et al., 2023).
Despite the promise of AI, SMEs face challenges in adoption, and evidence from emerging economies such as the UAE remains limited. This is particularly important because the UAE’s socio-economic environment is characterized by a diverse expatriate workforce, high digital penetration, and a strong policy-driven emphasis on innovation and sustainability (Poma et al., 2020). At the same time, regulatory frameworks such as Federal Law No. 11 of 2024 on mandatory climate-related disclosures and government-backed funding initiatives for SMEs create a unique institutional environment that both enables and pressures firms to adopt emerging technologies like AI (Gupta et al., 2022; Neumann et al., 2023).
Previous research has examined AI adoption in domains such as healthcare, tourism, and retail (Bonetti et al., 2023; Malik et al., 2021; Pillai & Sivathanu, 2020). However, most studies focus on large organizations or consumer perspectives (Cao et al., 2021; J. S. Chen et al., 2021), leaving the managerial viewpoint within SMEs underexplored (Abaddi, 2025; Baabdullah et al., 2021; J. Wang et al., 2022). Moreover, literature on AI adoption has been dominated by developed-country contexts, with relatively little evidence from emerging economies (Abdulmuhsin et al., 2025; Van et al., 2024). In particular, little is known about how UAE SMEs balance government-driven AI initiatives, socio-economic diversity, and competitive market dynamics when making adoption decisions, factors that may shape adoption differently compared to Western contexts (Cimino et al., 2025; Forcadell et al., 2021). This creates a clear gap in understanding the determinants influencing SMEs’ AI adoption in contexts such as the UAE.
Against this backdrop, the present study addresses the following research questions: (1) What are the key technological, organizational, and environmental factors influencing AI adoption intention among SMEs in the UAE? (2) How does integrating perceived usefulness from TAM into the TOE framework enrich our understanding of adoption drivers in this context?
By addressing the above discussed gaps, this study makes both theoretical and practical contributions. Theoretically, it extends the TOE framework to the context of SMEs in an emerging market. It also integrates insights from the TAM to capture the role of perceived usefulness, thereby offering a more holistic explanation of adoption intentions. Practically, it offers insights for SME leaders, policymakers, and industry stakeholders to design strategies that encourage effective AI adoption, thereby enhancing competitiveness and contributing to national economic growth.
Literature Review
In response to the crucial knowledge gaps pertaining to the drivers of AI adoption intention in SMEs, this study employs the TOE model alongside integrating TAM as the theoretical background for the proposed conceptual model. This section discusses the theoretical background and introduces hypotheses that revolve around the key constructs of this study as shown in Figure 1.
Conceptual model:
Theoretical Background
Tornatzky et al. (1990) first introduced the TOE model, which has since been widely used to explain organizational adoption of innovations. Building on early innovation-adoption research. TOE framework highlights that adoption is shaped by the dynamic interaction of technological, organizational, and environmental conditions. The model provides a comprehensive picture of the main factors that determine the adoption of new systems from an organizational perspective (Hasani et al., 2017; Sharma, Singh, Islam, & Dhir, 2022). It has been extensively applied across different contexts such as smart cities (Ullah et al., 2021), green innovation (Zhang et al., 2020), social commerce (Abed, 2020), and big data analytics (El-Haddadeh, 2020).
According to the TOE model, factors predicting the successful implementation and adoption of emerging systems fall into three key dimensions: technology, organization, and environment (Tornatzky et al., 1990). The technological dimension refers to internal and external technologies (hardware, software, and digital infrastructure) that influence adoption, including perceptions of relative advantage, compatibility, and complexity (Nagy et al., 2025). In this study, perceived competitive advantage and perceived usefulness capture the technological dimension (see Figure 1).
The organizational dimension includes a firm’s resources, size, structure, leadership, and human capabilities. In this study, it incorporates perceived top management support and perceived employee capability. Strong leadership commitment and workforce readiness are consistently shown as essential for overcoming challenges associated with integrating AI into organizational processes (Baabdullah et al., 2021).
The environmental dimension reflects the external pressures and supports that affect adoption, including industry structure, regulatory systems, and competitive forces (Martins et al., 2019). In this study, the environmental dimension is captured by perceived competitive pressure and perceived government regulations (S. Sun et al., 2020). These factors highlight that SMEs’ adoption decisions are not made in isolation but are shaped by their external ecosystem, including competitors, customers, suppliers, and policy frameworks.
The TOE framework is especially suitable for examining AI adoption in SMEs because it integrates both internal organizational readiness and external environmental pressures. While the technology acceptance model (TAM) primarily focuses on individual-level perceptions of technology, such as perceived usefulness and ease of use, the TOE framework accounts for broader structural constraints and contextual pressures that organizations face (Abed, 2020; Martins et al., 2019). By combining these perspectives, this study leverages the strengths of both models, capturing organizational and environmental influences through TOE while incorporating TAM’s emphasis on perceived usefulness as a key technological determinant. This makes the integrated approach particularly relevant in the UAE context, where government-driven AI initiatives and competitive market dynamics strongly influence SMEs’ technological adoption decisions.
Hypothesis Development
This section develops hypothesis about the key drivers of AI adoption intention by SMEs in the UAE.
Perceived competitive advantage is the extent to which an organization recognizes a technology to provide superior benefits (Sharma, Singh, Islam, & Dhir, 2022). Firms often adopt innovations because of the strategic benefits they deliver, such as efficiency gains, cost savings, and new market opportunities (Chatterjee et al., 2022). Prior studies have consistently shown that when SMEs believe emerging technologies such as cloud computing or social CRM deliver competitive differentiation, they are more inclined to adopt them (Roberts & Candi, 2024). In the context of AI, its ability to generate innovation and strengthen market position has been highlighted as a key driver of adoption (J. S. Chen et al., 2021). Thus, SMEs that perceive AI as a source of competitive advantage are more likely to intend adoption. Therefore, it is hypothesized as:
Perceived usefulness is defined in the TAM as the extent to which a person believes that using a technology will enhance performance (Davis, 1989). This construct has been widely validated across multiple settings, with perceived usefulness emerging as a strong determinant of adoption intention compared to other perceptions such as ease of use (Latreche et al., 2024). Within SMEs, perceived usefulness captures the belief that AI can improve decision-making, streamline operations, and enhance competitiveness (Baabdullah et al., 2021; Jalil et al., 2025). Integrating TAM with the TOE framework, perceived usefulness aligns with the technological dimension, reinforcing the argument that SMEs are motivated to adopt AI when they expect it to generate tangible business value. Therefore, the study hypothesizes as below:
Perceived top management support refers to the extent to which senior leadership actively backs the adoption of innovative technology (Wahab et al., 2025). This support may include authority, strategic direction, and the allocation of resources necessary for adoption (X. Wang & Dass, 2017). Prior research shows that top management commitment strongly influences organizational motivation to adopt new technologies, including cloud computing, mobile applications, and software-as-a-service (Oliveira et al., 2020). In SMEs, where decision-making is often centralized, management support becomes particularly critical (Johanson & Oliveira, 2024). Scholars consistently emphasize that leaders who prioritize technology adoption create a culture that reduces resistance and facilitates integration (Cao et al., 2021; Khanfar et al., 2025). This has also been demonstrated in the context of AI adoption, where leadership commitment is directly linked to organizational readiness (Chatterjee et al., 2022). Accordingly, when SME leaders actively champion AI adoption through clear vision and resource allocation, adoption intentions are likely to increase. Therefore, the following hypothesis is framed:
Perceived employee capability reflects the degree to which employees possess the skills, knowledge, and experience required to integrate new technologies into business processes (Sharma, Singh, Islam, & Dhir, 2022). Skilled personnel are widely recognized as a foundation for successful technology adoption, yet SMEs often face resource constraints and shortages of qualified staff compared to larger firms (Biea et al., 2024). This skill gap often necessitates additional costs for external expertise. The literature consistently highlights employee capability as a key determinant in adopting emerging technologies, such as mobile marketing and digital innovation (Eze et al., 2021). In the context of AI, employees’ technical proficiency enhances firms’ ability to successfully deploy applications, adapt to change, and improve decision-making processes (Chiu et al., 2021). SMEs with a workforce that demonstrates strong digital capabilities are therefore more likely to embrace AI adoption, resulting in improved performance and customer engagement (Arroyabe et al., 2024; Jalil et al., 2025). Building on the above arguments, the following hypothesis is proposed.
Perceived competitive pressure is the amount of pressure an organization faces from its competitors within the same industry (Sharma, Singh, Islam, & Dhir, 2022). In dynamic markets, early adopters often gain a first-mover advantage, compelling other firms to adopt in order to remain competitive (Cao et al., 2021). Competitive pressure has been shown to drive technology adoption across various domains, including enterprise resource planning (Xu et al., 2017) and Enterprise 2.0 (Jia et al., 2017). For SMEs, this pressure is particularly salient, as survival often depends on emulating industry leaders and responding swiftly to market shifts (Baabdullah et al., 2021). In the context of AI, competitive intensity has been empirically linked to greater adoption intentions, as businesses perceive lagging behind as a strategic risk (J. Chen & Tajdini, 2024; Singh et al., 2021). Thus, when SMEs experience stronger competitive pressure, they are more likely to intend adopting AI to sustain market relevance. Therefore, the following hypothesis is proposed:
Perceived government regulations refer to the subjective understanding business organizations hold regarding the existing or anticipated rules, policies, and legal frameworks imposed by the government (Gupta et al., 2022). In the AI context, governments that provide incentives, ethical frameworks, and innovation-friendly policies create confidence among SMEs to adopt responsibly (Pan et al., 2022). For example, the UAE’s National Strategy for Artificial Intelligence 2031 provides a roadmap that reassures businesses of long-term institutional support. Prior research highlights that government assistance is particularly vital in emerging economies, where institutional support can offset uncertainty and encourage SMEs to experiment with transformative technologies (Jalali, 2025). Thus, the following hypothesis is proposed:
Methodology
Research Design
This study adopts a cross-sectional survey research design, which is appropriate for investigating the determinants of AI adoption intention among SMEs in the UAE. A cross-sectional design allows the collection of data from a large sample of SME managers and executives at a single point in time, thereby facilitating the examination of hypothesized relationships between technological, organizational, and environmental factors. Such designs are widely recognized as effective in organizational and innovation adoption studies, particularly when the aim is to test theoretical models rather than establish causal relationships (Bryman, 2012). This approach is also particularly suitable for exploratory research in emerging markets, where SMEs operate under resource constraints and rapidly changing business conditions (Baruch & Holtom, 2008; J. Wang et al., 2022). Moreover, the cross-sectional survey design complements the use of PLS-SEM, as it provides standardized data that enable robust statistical testing of complex conceptual models (Hair et al., 2024).
Sample and Data Collection
Given that the study encompasses SMEs in the UAE, a cross-sectional survey design is deemed appropriate. For this study, the target population comprises SMEs operating within the hospitality sector in the UAE. The hospitality sector was selected because it is a cornerstone of the UAE’s diversification and tourism-driven economy, contributing nearly 12% to GDP and employing a significant proportion of the SME workforce (UAE SMEs, 2022). Hospitality SMEs are also at the forefront of digital transformation, with increasing reliance on AI-enabled solutions such as chatbots, smart booking systems, and personalized customer engagement tools (Bonetti et al., 2023; Gupta et al., 2022). This makes the sector a particularly suitable and timely context for exploring the determinants of AI adoption. The survey utilized Survey Monkey, a free survey administration tool as an integrated solution for conducting quantitative research to facilitate the data collection process. Invitations were extended to all respondents via email, containing an explanation of the research and a hyperlink to the survey website. Participants were encouraged to remain anonymous, and no identifiers were included. Of the 1,000 invitations sent, 500 respondents initially agreed to participate. However, during the data collection phase, 185 respondents with drew their earlier commitment, resulting in a final respondent count of 315. This equates to an effective response rate of 31.5%.
To ensure adequacy, multiple established criteria were considered in determining the sample size. Tabachnick and Fidell (2006) recommend a minimum of 300 responses for factor analysis, while Hair et al. (2012) suggest an item-response ratio of 1:10. With 26 items across constructs, this translates to a minimum requirement of 260 responses. Similarly, Green’s (1991) formula (N≥50 + 8m) recommends a minimum of 258 responses for 26 predictors. Consistent with prior studies (Paruthi et al., 2023; Shahid et al., 2022), the final sample size of 315 is therefore justified as appropriate and representative, meeting and exceeding the recommended thresholds for robust statistical analysis.
The majority of the respondents were male, constituting 85.0% of the sample, while females make up the remaining 15.0%. Notably, 85% of the respondents were male, which reflects the male-dominated managerial composition of SMEs in the UAE, particularly within the hospitality and technology-oriented sectors. The largest age group among the respondents falls within the 31 to 40 range, comprising 44.8% of the participants. The 21 to 30 age group accounts for 15.2%, the 41 to 50 age group for 29.8%, and those aged 50 and above for 10.2%. The most common ranks among respondents are executives (34.9%) and managers (25.1%), followed closely by senior managers (27.0%) and top management (13.0%). The profiles demonstrated that the sample mirrors the demographic background of a typical SMEs management in UAE. Table 1 illustrates the respondent profiles.
Demographic Profiles (N = 315).
In terms of firm size, the majority of businesses represented have employee counts ranging from 6 to 50 (74.0%). Businesses with fewer than 5 employees constitute 10.2%, while those with 51 to 200 employees make up 14.0%, and those with over 200 employees represent 1.9%. On the merit of sales, the sales distribution shows that the majority of businesses fall within the AED4 to 50 million range (67.9%). Smaller percentages are distributed among businesses with sales below AED3 million (12.1%), AED51 to 250 million (16.8%), and above AED250 million (3.2%). The majority of respondents have been in business for 11 to 15 years (60.0%), followed by those with 5 to 10 years of experience (13.7%) and those with less than 5 years (15.2%). Businesses with 15 to 20 years of operation constitute 10.2%, and those with more than 20 years represent 1.0%. The distribution across different emirates indicates that Abu Dhabi and Dubai have the highest representation, with 32.1% and 27.9%, respectively. Other emirates, including Ajman, Sharjah, Fujairah, Ras Al-Khaimah, and Um al Quwain, have varying but notable percentages. In general, the sample in this study represents the population characteristics of the SMEs in UAE.
Measures
Using a seven-point Likert scale ranging from 1 = “strongly disagree” to 7 = “strongly agree,” all utilized measures in this study were adopted from prior research. Some minor modifications were made to the scale items to suit the SME context of UAE. Prior to sharing the questionnaire link, we assessed the questionnaire’s clarity by pilot testing it on a sample of 20 respondents comprising of 3 doctoral students, 7 master’s students, and 10 bachelor’s students who had a prior experience of working in the UAE SMEs. In addition, the questionnaire was assessed by two Marketing professors. Based on their collective inputs, some minor edits were made to the questionnaire to improve its readability and accuracy.
Perceived competitive advantage was measured by adapting four items from Lian et al. (2014) and Sharma, Singh, Islam, and Dhir (2022), with a sample item reading as “AI would enable our SME to communicate our products/services in a better way.” Perceived usefulness was measured by borrowing four items from Chatterjee et al. (2021) with a sample item reading as “adopting AI will make our SME more efficient.” Perceived top management support was measured by adapting four items from Lian et al. (2014) and Sharma, Singh, Islam, and Dhir (2022). A sample article includes “Top Management would provide the necessary support for the adoption of AI.” Perceived employee capability was measured by adapting four items from Sharma, Singh, Islam, and Dhir (2022), with a sample item as “This SME’s employees would be capable of learning new AI-related technology easily.” Three items to measure perceived competitive pressure were taken from Ghobakhloo and Chin (2019) and Sharma, Singh, Islam, and Dhir (2022), with a sample item reading as “our SME is under pressure from competitors to adopt AI.” Perceived government regulations were measured by adapting four items from Gupta et al. (2022), with a sample item as “the government introduces relevant policies to boost AI in SMEs for its development.” Lastly, AI adoption intention by SMEs was measured by adapting three items from Maduku et al. (2016), with a sample item reading as “our SME intends to use AI.” These scales were chosen because they have been widely validated in prior studies on technology and innovation adoption, ensuring both reliability and comparability with existing literature. Their adaptation to the UAE SME context enhances contextual relevance, while still maintaining alignment with global standards in AI adoption research.
Data Analysis and Results
This study applied established procedures to evaluate the reliability and validity of the measures. The examination of construct reliability and convergent validity in Table 2 employs three criteria for reliability and a validity measure. The three metrics for construct reliability, namely Dijkstra-Henseler’s rho (ρA), Jöreskog’s rho (ρc), and Cronbach’s alpha (α), are utilized. It is worth noting that Dijkstra-Henseler’s rho (ρ_A) generally falls between Cronbach’s alpha and composite reliability ρ_C (Hair et al., 2024), with Dijkstra and Henseler (2015) proposing the exact (or consistent) reliability coefficient ρ_A. Table 2 reveals that all reliability values surpass the .70 cutoff threshold (although they lean toward the higher end), affirming the reliability of the constructs (Hair & Alamer, 2022). Besides, convergent validity, interpreted through the average variance extracted (AVE), denotes the average reliability of indicators. AVE values exceeding 0.50 are indicative of satisfactory convergent validity (Hair & Alamer, 2022). As shown in Table 2, the AVE values of all the constructs of this study exceed 0.50, indicating satisfactory convergent validity.
Construct Reliability and Convergent Validity.
The item loadings for the SEM, shown in Table 3 have a minimum value of 0.838. As Hair et al. (2024) recommends a cutoff value of 0.70, the loadings meet the cutoff value and are all significant at 5% level.
Item Loadings.
Once individual constructs undergo testing for reliability and validity, the assessment of discriminant validity becomes crucial. Fornell and Larcker (1981) propose a criterion ensuring that measurement variables effectively capture the intended construct and possess uniqueness to that specific construct. This criterion, termed the discriminant validity test, is exemplified in Table 4 below. The Fornell-Larcker criterion involves comparing the diagonals (representing the square root of construct AVE) with the respective rows and columns (indicating the absolute value of bivariate correlations of the construct with all the other constructs). Discriminant validity is affirmed when the diagonals surpass their corresponding rows and columns. As shown in Table 4, all diagonal values exceed those in the respective rows and columns, conclusively establishing discriminant validity (Hair & Alamer, 2022).
Discriminant Validity.
Note. SD = standard deviation. Diagonal bold figures represent the square roots of the AVEs.
Correlations values are significant at p < .001.
This study also examined the potential for common method bias (CMB) utilizing Harman’s single-factor test as proposed by Podsakoff et al. (2003). Correspondingly, all items were subjected to Principal Axis Factoring without rotation. The results revealed that a single factor emerged as explaining only 36.89% of total variance, which being less than 50% ruled out the apprehension of CMB in our data (Podsakoff et al., 2003).
In addition, variance inflation factor (VIF) scores were calculated for all constructs, and all values were well below the conservative threshold of 3.3, further indicating that multicollinearity and common method variance were not a concern. To further mitigate the risk of bias, procedural remedies were also applied during data collection, including respondent anonymity, and the use of validated scales. Together, these procedural and statistical approaches provide assurance that CMB is unlikely to compromise the validity of our findings (Podsakoff et al., 2003).
Structural Model Results
The analysis of the structural model confirms that all six hypothesized relationships are positive and significant. Perceived top management support (β = .48), competitive pressure (β = .47), and government regulations (β = .43) emerged as the strongest drivers of AI adoption intention, while perceived usefulness (β = .39), competitive advantage (β = .38), and employee capability (β = .30), also exerted significant positive effects. Collectively, these results demonstrate that both organizational leadership and environmental pressures play a more dominant role compared to purely technological perceptions in shaping SMEs’ AI adoption intentions.
To scrutinize the proposed model, this study employed the “Partial Least Squares Structure Equation Modelling” (PLS-SEM) using SmartPLS 4. The significance of relationships was examined through bootstrapping methods (Hair et al., 2024). The measurement model exhibited a good fit to data on various fit indices (e.g., χ2 = 488.973, df = 189, χ2/df = 2.587, CFI = 0.949, NFI = 0.918, RFI = 0.96, IFI = 0.962, TLI = 0.946, and RMSEA = 0.06). Table 5 presents the results, organized in descending order based on the original sample (PLS-SEM path coefficients). The significance of each coefficient is tested at the 5% level. The path coefficients, ranging from 0.294 to 0.479, signify the strength and direction of the relationships between various constructs in the model.
Path Coefficients Results.
Although Table 5 also reports the bootstrapped sample means of the path coefficients, these values are not discussed further as PLS-SEM emphasizes the interpretation of original coefficients, t-statistics, and p-values for hypothesis testing (Hair et al., 2024). The SEM results shown in Table 5 reveal compelling evidence of all the proposed factors positively affecting AI adoption intention, thereby supporting all the six proposed hypothesis. The p values are all below the conventional significance threshold of.05, indicating robust support for the proposed relationships. The consistently strong support for each path underscores the reliability and stability of the relationships observed. The robustness of the findings suggests that the identified constructs are not only statistically significant but also substantively meaningful contributors to the prediction of AI adoption.
Discussion
The aim of this study was to identify the factors that could potentially drive the intention to the adoption of AI by SMEs in the UAE. Drawing upon data from 315 respondents across different managerial levels, the analysis confirmed that all six hypothesized determinants significantly influence adoption intention, thereby validating the applicability of the TOE framework (Tornatzky et al., 1990) in this context. Moreover, the findings provide complementary support to the TAM, particularly in highlighting the role of perceived usefulness in shaping adoption intentions. By integrating TOE and TAM, this study not only captures structural and contextual conditions but also accounts for individual-level technology perceptions, offering a more nuanced explanation of AI adoption in SMEs.
From the technological dimension, both perceived competitive advantage and perceived usefulness positively affected AI adoption intention. This aligns with TAM’s central argument that technology adoption depends heavily on beliefs about its usefulness. SMEs in this study viewed AI as a tool to improve efficiency, productivity, and customer service, which resonates with prior findings in AI adoption research (Baabdullah et al., 2021; Chatterjee et al., 2021). The role of perceived competitive advantage also reinforces arguments that firms adopt innovations when they believe these technologies provide differentiation and improved market positioning (Aftab et al., 2024; S. Sun et al., 2020). However, when compared with other determinants, technological factors had a smaller influence, indicating that usefulness and competitive advantage alone are insufficient without organizational readiness and environmental pressures (Dwivedi et al., 2021; Mikalef & Gupta, 2021). This implies that in the UAE context, technology perceptions alone are secondary to institutional signals and managerial readiness, reflecting the country’s strong policy-driven innovation ecosystem. Together, these results indicate that SMEs in the UAE are strategically motivated by both internal performance gains and external market benefits.
Within the organizational dimension, top management support emerged as a strong driver of adoption intention. This is consistent with studies that emphasize leadership commitment as pivotal in allocating resources, legitimizing change, and overcoming organizational resistance (Khan et al., 2025; Oliveira et al., 2020). In SMEs, where ownership and decision-making are often centralized, the influence of top management becomes even more critical (Baabdullah et al., 2021). Similarly, employee capability was found to significantly enhance adoption intentions. Prior research has highlighted that a skilled workforce mitigates barriers to adoption and facilitates integration of advanced technologies (Eze et al., 2021). Yet, its effect was comparatively weaker than top management support, highlighting that employee competence, although necessary, cannot independently drive AI adoption unless championed and resourced by leadership (Srivastava et al., 2022). This suggests that organizational readiness in SMEs is highly leadership centric. In the UAE’s multicultural business environment, this finding underscores how leadership endorsement not only provides resources but also fosters alignment across a diverse workforce, thereby reducing resistance to AI initiatives (Poma et al., 2020).
From the environmental perspective, both competitive pressure and government regulations were significant. The positive effect of competitive pressure echoes findings where firms adopt technologies to avoid lagging behind peers and maintain strategic parity (Jia et al., 2017; Jibril et al., 2024). This effect may be particularly salient in the UAE’s hospitality sector, where technological adoption is rapidly diffusing as firms compete for digitally savvy customers. Similarly, government regulations played an enabling role. Supportive policies such as the UAE’s National Strategy for Artificial Intelligence 2031 likely reduced uncertainty and encouraged SMEs to consider AI adoption (Gupta et al., 2022; Pan et al., 2022). This finding corroborates prior studies suggesting that regulatory clarity and incentives significantly encourage SMEs in emerging economies to embrace transformative technologies (J. Wang et al., 2022). Notably, the relative strength of competitive pressure and government regulation as predictors indicates SMEs in the UAE are particularly responsive to external signals, where market rivalry and institutional support jointly act as catalysts for adoption (Neumann et al., 2023; Sawang et al., 2024). The UAE’s regulatory environment, including mandatory climate-related disclosures and government-backed SME financing programs, further illustrates how policy-driven frameworks directly shape adoption decisions, making the institutional context decisive.
Taken together, these findings show that the TOE framework provides a comprehensive lens to understand AI adoption by SMEs in the UAE, while TAM enriches the analysis by explaining the central role of perceived usefulness. More importantly, the dominance of organizational (top management support) and environmental (competitive pressure, government regulation) factors over technological perceptions highlights that in emerging-market contexts, adoption decisions are less about intrinsic evaluations of technology and more about leadership endorsement and institutional pressures (Cao et al., 2021; J. Wang et al., 2022). This UAE-specific insight demonstrates that SMEs operate in a policy-driven, highly competitive ecosystem, where government initiatives and managerial leadership outweigh the intrinsic appeal of AI. Such findings contextualize adoption behavior in emerging economies, offering guidance for policymakers and industry leaders to create enabling conditions that accelerate SME digital transformation.
Implications
These results of this study carry significant theoretical and practical implications, as discussed in the following section.
Theoretical Implications
The identified paths highlight specific latent constructs that play a crucial role in shaping the AI adoption intention of SMEs in the UAE. The insights offered by this study can guide researchers and policymakers in focusing their efforts on enhancing or leveraging these key constructs to optimize the development and adoption of AI technologies.
First, by focusing on the determinants driving SMEs to adopt AI, the current study contributes to a more detailed understanding of AI adoption within the SME sector. This fills a crucial void in the literature, which has predominantly focused on larger organizations (Baabdullah et al., 2021; Jalil et al., 2025).
Second, while the majority of earlier research on AI has been dedicated to understanding customer perspectives, this study shifts the focus to the managerial standpoint within SMEs, thereby extending the body of knowledge on organizational adoption. Prior research has predominantly examined adoption from consumer-facing or large-firm perspectives (Bonetti et al., 2023; Cao et al., 2021), while limited work has captured how SME decision-makers frame and influence adoption choices (Arroyabe et al., 2024; J. Wang et al., 2022).
Third, research on the adoption of AI has predominantly been carried out in advanced nations, with a notable emphasis on countries like the United States, with little empirical research conducted in other geographic contexts such as the UAE. By addressing this gap, this study offers a context-specific perspective on how institutional environments and national innovation strategies influence adoption in emerging economies (Forcadell et al., 2021; Neumann et al., 2023; Wei & Pardo, 2022).
The application of the TOE framework as the theoretical foundation for AI adoption by UAE SMEs reaffirms its robustness in explaining organizational adoption of emerging technologies (Abed, 2020; M. A. Nazir et al., 2025). This study also advances theory by integrating TOE with TAM, showing that perceived usefulness significantly influences AI adoption intention. The results demonstrate how individual-level beliefs (TAM) (Davis, 1989) interact with organizational and environmental factors, enriching adoption research by highlighting that usefulness-driven intentions are most likely to translate into adoption when managerial support and favorable environmental conditions are present (Oliveira et al., 2020; Pan et al., 2022).
Practical Implications
This study offers actionable implications for SME managers in the UAE by translating findings into concrete strategies for AI adoption. The results highlight not only which factors matter most but also how they can be operationalized in practice.
First, adoption drivers must be linked to day-to-day practices. Top management support should translate into budgets, pilot projects, and visible endorsement to reduce resistance (Oliveira et al., 2020). Employee capability, though significant, requires reinforcement through structured training and modular learning to ensure effective use of AI (Eze et al., 2021; Ghobakhloo & Ching, 2019). Environmental factors such as competitive pressure and regulations call for benchmarking against peers and compliance with evolving policies, including Federal Law No. 11 of 2024 (Gupta et al., 2022; Pan et al., 2022).
Second, SMEs should adopt a pilot approach, beginning with high-impact projects like chatbots, demand forecasting, or recommendation systems. These “quick wins” demonstrate value, reduce uncertainty, and build organizational confidence, especially in resource-constrained SMEs (Chatterjee et al., 2021).
Third, SMEs should actively engage with government training initiatives, AI sandboxes, and funding schemes. By aligning with policy incentives and demonstrating compliance with evolving data governance rules, SMEs not only reduce risk but also gain legitimacy in the eyes of customers and business partners (Gupta et al., 2022).
Fourth, workforce upskilling through partnerships with local universities and training platforms is essential. For example, hospitality SMEs could train staff in AI-enabled booking systems and digital marketing analytics, ensuring that AI tools are effectively used in daily operations (Eze et al., 2021).
Fifth, SMEs should leverage networks in free zones or associations to share resources and reduce adoption costs (Xu et al., 2017). Managers can form alliances within free zones (e.g., Jebel Ali Free Zone Authority, Dubai Internet City) or sectoral associations to share best practices, pool resources, and jointly access AI service providers.
Sixth, framing AI as a customer experience enhancer rather than a back-end tool can strengthen competitiveness (Baabdullah et al., 2021). Lastly, top management should establish a visible culture of AI leadership. Beyond financial support, SME owner-managers must act as AI champions, communicating the long-term vision of AI adoption, celebrating small successes, and reducing employee resistance to change.
Limitations and Future Research Directions
While this study makes meaningful contributions, several limitations must be acknowledged. First, the research employed a cross-sectional survey design, capturing managerial perceptions of AI adoption intention at a single point in time. Although this provides valuable insights, it restricts the ability to establish causality or track adoption over time. Future research could adopt longitudinal designs to examine how determinants of AI adoption change over time as SMEs progress from intention to actual implementation.
Second, the study relied on self-reported data from SME managers and executives, which may introduce common method and social desirability biases. Despite using validated scales and safeguards, these limitations cannot be fully ruled out. Future work could address this by integrating multi-source data (e.g., employee perspectives, customer insights, or archival records) to triangulate findings and reduce bias (Podsakoff et al., 2003).
Third, the sample was restricted to UAE hospitality SMEs, which, while highly relevant to the economy, limits generalizability across other sectors or regions. The UAE’s institutional, regulatory, and cultural environment may also shape adoption differently from other markets. Future research should broaden the scope to include cross-industry and cross-country studies for deeper insights into contextual contingencies (Wei & Pardo, 2022).
Fourth, another limitation concerns the demographic profile of respondents, as 85% were male managers. While this reflects the male-dominated managerial landscape of UAE SMEs, it may constrain inclusiveness, particularly regarding gendered differences in adoption perceptions. Future studies should aim for more balanced gender representation to provide a fuller understanding of adoption drivers.
Fifth, while this study drew on the TOE framework and TAM, it focused on a limited set of six determinants. Other influential factors, such as cost structures, organizational culture, technological readiness, customer acceptance, and perceived risks, were not included. Future research could incorporate these constructs to develop a more holistic model of AI adoption. Moreover, exploring interactions among TOE dimensions and TAM constructs would advance integration and capture adoption complexity.
Finally, SMEs are highly sensitive to external shocks such as crises, technological disruptions, or sudden regulatory shifts. This study did not account for the impact of such macro-level forces. Future research could explore how events like COVID-19, geopolitical tensions, or sustainability regulations interact with firm-level determinants to influence adoption decisions.
Conclusion
This study examined the determinants of AI adoption intention among SMEs in the UAE, drawing on the TOE framework and TAM. The findings reveal that all six proposed factors significantly shape adoption intention, with top management support, competitive pressure, and government regulations exerting the strongest effects, while perceived usefulness, competitive advantage, and employee capability also play meaningful roles. These results highlight that organizational leadership and environmental influences are critical, complementing technological perceptions in driving AI adoption. Theoretically, the study extends the TOE framework by integrating TAM, and practically, it provides SME managers and policymakers in the UAE with actionable insights to foster AI adoption.
Despite these contributions, several limitations should be noted. The reliance on a cross-sectional survey restricts causal inference, the focus on UAE hospitality SMEs limits generalizability, and the use of self-reported measures may introduce bias. These constraints provide opportunities for future work to refine and extend the findings.
Future research should employ longitudinal designs to capture changes in adoption behavior over time, expand the scope to diverse industries and cross-country comparisons, and incorporate additional constructs such as cost, risk, and organizational culture to develop a more holistic understanding of AI adoption. Further studies could also explore how external shocks, regulatory shifts, or sustainability pressures influence SMEs’ technology adoption trajectories.
In summary, this study provides empirical evidence that SMEs’ AI adoption in the UAE is shaped by technological, organizational, and environmental factors, with leadership commitment and external pressures playing dominant roles. By acknowledging its limitations and offering clear directions for future research, the study contributes to advancing both theoretical discourse and the practical roadmap for AI adoption among SMEs in emerging markets.
Footnotes
Acknowledgements
The first author would like to acknowledge Prince Sultan University, Saudi Arabia for their support. This research was supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project Number (PNURSP2025R542), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project Number (PNURSP2025R542), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Data will be made available on personal request.