Adoption of AI by micro and small health enterprises: effects of entrepreneurial orientation on the TOE model Open Access

Technology shapes corporate structures, influencing their size and determining their potential actions (Giustiziero et al., 2023). Understanding how technologies are adopted enables a better grasp of the diffusion process, improves management of the implementation process, and highlights the figures and dynamics that a company should consider when adopting new technologies. The most disruptive technology to date is Artificial Intelligence – AI. AI’s ability to learn, connect, and adapt (Huang and Rust, 2021) is redefining business logic across all domains, from production (Calabrese et al., 2023; Felsberger et al., 2022) to processes (Jarrahi et al., 2023; Frank et al., 2019), enhancing competitive advantage (Akter et al., 2023) and influencing innovation strategies (Tekic and Fuller, 2023; Mariani et al., 2023; Yu et al., 2024). According to a McKinsey Report (2022), AI prevalence has doubled since 2017. The report highlights a shift from developing AI products to focusing on their application in service operations, aimed at optimizing and enhancing value creation through cost reduction and revenue increase (McKinsey Report, 2022). The impact of AI spans all sectors but in different ways. For example, Brea and Ford (2023) identified that AI-related cognitive technology does not lead to the same levels of innovation across sectors: scientific and technology-based innovation firms benefit the most, while those related to manufacturing and services, such as pharmacies, are still less able to exploit the potential of such technologies and their adoption fully (Mariani et al., 2023). The increasing adoption of AI-based systems raises important questions regarding the main levers that lead a business not only to adopt AI but also to reflect on the models available to date in terms of technology adoption for businesses. Despite the growing interest and use of AI solutions, there are significant differences between firms: micro and small enterprises lag furthest behind in adopting AI technologies, and when they adopt them, they mostly adopt external models and tools (IBM Global AI Adoption Index, 2022). This could be due to the different sizes and resources available (Pedota et al., 2023). The World Bank reports that SMEs make up about 90% of the world’s enterprises. By 2021, there were 332.99 million SMEs worldwide (Statista, 2022), and missing out on AI adoption could mean a serious loss of opportunities to expand value creation and a threat to their competitive advantage. SME is the backbone of many economies, contributing significantly to employment and country growth. However, SMEs encounter significant challenges when it comes to adopting new technologies. For example, only 7% of small enterprises and 15% of medium-sized firms have launched AI-related projects, in contrast to 59% of large corporations (UniSolutions, 2025). Given these hurdles, it is crucial to develop specific models to facilitate technology adoption among SMEs, ensuring their sustainability and support (Schwaeke et al., 2024).

Their agility and innovative potential often drive economic growth and technological advancement. Regardless SME also faces unique challenges, including limited resources and expertise, which can hinder the adoption of advanced technologies like AI (Upadhyay et al., 2023). Understanding how small businesses adopt AI can provide insights into the barriers and facilitators of technology integration in this crucial firm’s typology. It can help identify micro and small businesses’ specific needs and concerns, such as compatibility, the firm’s relative advantage, organizational readiness, and the role played by their individual entrepreneur, which might differ from larger and bigger corporations. This knowledge can guide policymakers and technology developers in creating targeted support programs and more accessible AI solutions tailored to the small business context. The study aims to enrich theoretical and managerial knowledge using mixed methods and contribute to the broader understanding of technology diffusion. It can reveal patterns and behavior that are not apparent in studies focused solely on big enterprises, offering a more comprehensive picture of technological adoption across different firms’ sizes and what is the influence of the entrepreneur’s individual propensity in support of this adoption. In particular, it highlights one of the most frequently adopted models in terms of technology acceptance by firms: the Technology-Organization-Environment (TOE) Model by DePietro et al. (1990), adapted to explain the rate of AI acceptance in small and micro firms. This theoretical framework was chosen because it has been widely demonstrated in the literature to be valid for analyzing the determinants of AI adoption in organizations (Idris, 2015; Yang et al., 2015; Chatterjee et al., 2021), particularly in SMEs. Unlike other theoretical models (e.g. the Technology Acceptance Model, Theory of Planned Behavior, Unified Theory of Acceptance and Use of Technology), which focus on individual attitudes toward technology, TOE provides a broader perspective by emphasizing the organizational and environmental context. Consequently, the need to examine AI adoption at the firm level rather than at the employee level led the authors to adopt this model as the theoretical lens for conducting the study and as confirmed by the preliminary qualitative study conducted.

Pharmacies were chosen as the focal point of this study because they are predominantly micro, small, and occasionally medium-sized businesses, typically managed and owned by pharmacist-entrepreneurs (D’Souza and Scahill, 2020) and have a strong retail-oriented nature. They are significantly impacted by the AI transformation sweeping through the healthcare sector (Leone et al., 2021), from pharmaceutical companies (Rathi et al., 2024; Sharma et al., 2023) to hospitals and other healthcare infrastructures (Pham et al., 2024; Ali Mohamad et al., 2023). The study has its relevance because is one of the first that seeks to shed light on the role played by the entrepreneurial orientation of the entrepreneur in adopting new technologies, a very relevant dimension in micro-small and medium-sized enterprises. Chebo and Wubatie (2021) found that technologically oriented entrepreneurs are more likely to adopt innovation and technology. Furthermore, entrepreneurial orientation ensures better performance in small enterprises (Fatima and Bilal, 2020). However, little is known about entrepreneurial orientation’s role in micro and small businesses, especially in technology and innovation decisions (Ritala et al., 2021; Peltier et al., 2012). Upadhyay et al. (2023) have called for more studies in this regard, given the relevance it may have in the adoption of AI solutions. In the end, the study expands our knowledge of which dimensions can be considered part of the conceptualization of individual entrepreneurial orientation (Howard and Floyd, 2024; Santos et al., 2020). In light of the growing interest in and diffusion of AI, it becomes crucial to understand what, at the enterprise level, are the main determinants of adoption for small and micro businesses, which remain an underexplored area. A mixed-methods approach was adopted to address the research objective and answer the question: “What are the most influential dimensions for the adoption of AI in small and micro firms?” Initially, a qualitative analysis through semi-structured interviews with 11 pharmacists was conducted to identify the most suitable adoption model, supported by a literature review. The founded model was then tested quantitatively to determine the dimensions influencing AI adoption in micro and small businesses, such as pharmacies, through a regression analysis with a questionnaire of 217 pharmacists. The study is one of the first to investigate which size may be most influential in the adoption of AI, focusing in particular on the healthcare sector and its small and medium-sized enterprises. Furthermore, it is one of the first studies to investigate the role of the entrepreneurial mindset in addition to size in the characteristics of a company. The paper is structured as follows: the next section is devoted to the literature review and theoretical framework, the third section to the methodology, and the fourth to the presentation of the results of the qualitative study, the developed model, and the subsequent quantitative study. The fifth section discusses theoretical and managerial implications, and the sixth addresses limitations and future research directions.

One of the main interests of this study is to explore AI adoption patterns. Venkatesh et al. (2022) highlighted the need for new models to understand AI acceptance across different sectors and to investigate the outcomes AI adoption may bring about. Traditional models of technology adoption have focused extensively on employee and consumer perspectives, relying on foundational frameworks such as the Technology Acceptance Model first and second version (Davis and Venkatesh, 1996; Davis, 1989; Venkatesh and Davis, 2000), the Theory of Planned Behavior (TPB) (Ajzen, 1991), and the Unified Theory of Acceptance and Use of Technology (UTAUT) (Venkatesh et al., 2003).

The exponential growth of technology in all areas of human life has highlighted the need to understand consumer attitudes and perceptions. The first theories of social-psychological derivation are the Theory of Reasoned Action (TRA) and, consequently, the Theory of Planned Behavior (TPB). The TRA, developed by Ajzen and Fishbein (1988), postulates that behavioral intention is the main predictor of behavior, without considering those behaviors over which individuals have “incomplete” control. To overcome this limitation, the authors introduced the concept of “perceived behavioral control”, which led to the TPB. This model posits that individuals’ behavior results from an intention to engage in that behavior, which is influenced by attitudes toward the behavior, subjective norms, and perceptions of the ability to engage in that behavior. The main limitation of this model is that it excludes unconscious motivations that may influence a given behavior, assuming that every choice is based on rationality and systematicity. Davis and Venkatesh (1996), therefore, developed the Technology Acceptance Model by adapting previous models, the development of which has been extensively discussed in the literature (Marangunić and Granić, 2015; Legris et al., 2003).

Particularly, the Technology Acceptance Model – TAM in its first version focuses on only two variables able to explain the acceptance and use of technology: (1) perceived usefulness, the degree to which an individual expects that using a particular system or technology will help them perform better or be more effective at their job and (2) perceived ease of use, related to the less effort in use a technology that led to adopt it. The validation of perceived usefulness as the main determinant of behavior led Venkatesh and Davis (2000) to formulate an extension of the TAM, known as TAM2, which identifies the variables influencing this usefulness adding other “theoretical constructs spanning social influence processes (subjective norm, voluntariness, and image) and cognitive instrumental processes (job relevance, output quality, result demonstrability, and perceived ease of use)” (Venkatesh and Davis, 2000, p. 187).These models have been adapted and integrated with more corporate dimensions to better understand technology adoption in organizational contexts. For instance, the Small Business Technology Acceptance Model (SBTAM) (Peltier et al., 2012) examines how environmental factors, as well as owner and organizational characteristics and their perceptions toward technology, influence technology adoption. Similarly, the Firm-TAM (Doe et al., 2019, 2022) assesses how factors related to the firm, technological characteristics, personal factors, societal factors, risk propensity, and firm size significantly influence organizational technology adoption. These studies have shown that micro-firms are particularly influenced by societal factors, such as stakeholder roles and technology.

Another model used to explain the adoption of new technologies is the UTAUT, developed by Venkatesh et al. in 2003. This model integrates previous theoretical frameworks by identifying four main constructs as predictors of behavior: performance expectancy, effort expectancy, social influence, and facilitating conditions. Additionally, the model recognizes demographic factors (i.e. age and gender) as moderators of the effect of these constructs on usage behavior. Cobelli et al. (2023) used the Unified Theory of Acceptance and Use of Technology (UTAUT) model in combination with the market orientation paradigm to demonstrate how the latter dimension impacts performance, effort expectancy, and social influence, thereby affecting the adoption of telemedicine by pharmacies. We then have UTAUT2 (Unified Theory of Acceptance and Use of Technology 2), an extension of the original UTAUT model, developed to analyze technology adoption by individuals more in-depth. Compared to the original model, which primarily focused on the organizational context, UTAUT2 includes new variables that account for individual behaviors and motivations, making it more suitable for studying technology adoption in the consumer context. It incorporates constructs such as experience, hedonic value, costs, and voluntariness. For a more comprehensive analysis, see Venkatesh et al. (2012). In general, various models and frameworks exist to study adoption. For example, the Trust-disposition Model (ATM) was used by Shareef et al. (2023) to examine the adoption of autonomous home assistance systems, focusing on building trust through factors like “Personal Ability and Control” (PAC) and “Empathic Cooperation and Social Interaction” (ECSI). Another example is the Expected Trust Model for Machine Autonomy (ETM4MA), investigated by Shareef et al. (2023), this model highlights how trust and social interaction influence the acceptance of automated assistance systems, particularly among the elderly. Models have also been built based on the Uses and Gratification Theory, Signaling Theory, and Prospect Theory. These theoretical models have mainly been employed to understand AI adoption based on users’ needs and gratifications, as shown by Maroufkhani et al. (2022) in the context of voice assistants. However, there has been limited research on acceptance models for micro and small firms (Asiaei and Ab Rahim, 2019; Upadhyay et al., 2023). Omrani et al. (2022) distinguished between internal and external factors influencing enterprise adoption of technology. Internal factors include organizational skills, internal regulations, innovation rate, digitization, and innovation issues, while external factors encompass governmental dimensions, partnership opportunities, market infrastructure, and other environmental elements. They identified the most influential internal dimensions and emphasized the need for developing and integrating new models for companies.

One of the most widely adopted models for business technology adoption is the Technology-Organization-Environment (TOE) model (Mikalef et al., 2022), initially conceptualized by DePietro et al. (1990) based on the Contingency Theory of Organizations. The TOE model posits that technology adoption in a firm is driven by three constructs: Technology, Organization, and Environment. The Technology construct includes characteristics of the technological innovation such as compatibility, complexity, resources possessed by the firm, relative advantage, trialability, observability, trust, and security (AlKhater et al., 2018; Chiu et al., 2017; Gangwar et al., 2015; Senyo et al., 2016). The Organization construct encompasses the degree of organizational readiness and innovativeness, financial and organizational resources, managerial support, and employee readiness (Chiu et al., 2017; Gökalp et al., 2020; Nam et al., 2021). The Environmental construct includes dimensions such as competitive pressure, consumer influence, business partnerships, regulatory systems, and government support (Chiu et al., 2017; Abed, 2020; Sharma et al., 2023). While the TOE model aligns with the Contingency Theory of Organization and systematically addresses the constructs likely to influence technology adoption, it has faced criticism for its requirement of specialized identification of individual dimensions within each construct (Musawa and Wahab, 2012; Wang et al., 2010). Baker (2012) highlighted the need for further investigation and reflection on the TOE model. The model has been integrated with other frameworks, such as the Technology Acceptance Model (TAM), to explore specific industry applications. For example, Katebi et al. (2022) integrated TOE with TAM in the construction industry to understand the relationship between the individual size of engineers and the company’s setup in adopting Precast Concrete Components.

Therefore, the TOE model has been widely used to identify the determinants of AI adoption in organizations, taking into account not only technological and organizational aspects but also external factors that may influence usage behavior (Idris, 2015; Yang et al., 2015; Chatterjee et al., 2021). It has also been used in various sectors, including public enterprises, to assess the role of each dimension in enhancing dynamic capabilities (Mikalef et al., 2022) and in manufacturing firms to identify factors influencing AI technology adoption (Chatterjee et al., 2021). The application of the technology–organization–environment (TOE) model in the context of AI adoption is well-documented and validated across various studies. For instance, Chatterjee et al. (2021) explored how organizational resources influence AI implementation, highlighting the importance of infrastructure and knowledge. Horani et al. (2023) focused on the competitive pressures that drive SMEs to adopt AI technologies, emphasizing the role of market dynamics. Kinkel et al. (2022) examined the regulatory environment and its impact on technology adoption in manufacturing showing as organizational dimension are leading one, while Merhi and Harfouche (2023) investigated the cultural factors that facilitate or hinder AI integration within organizations. Sharma et al. (2023) analyzed the ecosystem surrounding SMEs, stressing the need for collaboration and partnerships to successfully implement AI solutions. In their systematic literature review, Schwaeke et al. (2024) provide a comprehensive understanding of technology adoption by firms, analyzing these factors at the three distinct levels of the TOE model. Their work identifies critical trends and gaps in the literature, offering a nuanced perspective that can empower researchers and practitioners to enhance AI adoption strategies effectively. However, its use in the healthcare sector remains limited. Ahmadi et al. (2017) applied it to assess the adoption of data computerization systems in a public hospital, Yang et al. (2022) used it to study the adoption of elderly care service resources, and Kong et al. (2021) applied it to the Business-to-Business sector of the healthcare industry.

In the healthcare sector, the potential of AI can contribute to reducing many contemporary criticalities (Santosh and Gaur, 2022; Leone et al., 2021), from aging to oncological diseases, both in their detection as well as in the definition of the treatment process (Dragoni et al., 2023; Tolkach et al., 2023; Uthamacumaran et al., 2023). New AI solutions are accelerating the spread of new business models (Garbuio and Lin, 2018; Kulkov, 2021) and the development and trial processes of new drugs (Giacomoni, 2022; Patel and Shah, 2022). By introducing new work processes and reducing the burden on traditional healthcare facilities, AI can transform them into smart healthcare organizations (Ahmad et al., 2022; Bini, 2018; Farahani et al., 2018). Among the diverse stakeholders in the healthcare ecosystem, pharmacies play a significant role. These establishments supply and distribute medications, making them pivotal within the healthcare sector, particularly in retail pharmacy settings (Nørgaard and Sporrong, 2019). Despite their critical role, pharmacies remain relatively under-investigated within the healthcare ecosystem. The COVID-19 pandemic underscored the essential role of pharmacies, enabling disease monitoring (Fittler et al., 2022), providing frontline solutions, and facilitating widespread vaccination efforts (Patterson and Holdford, 2019).

The evolution of the pharmaceutical market has prompted pharmacies to adapt, offering new services that align with consumer-oriented markets, such as delivery services and e-commerce (Martins and Queiros, 2015; Papanagnou and Matthews Amune, 2018; Liu et al., 2020; Wu and Dong, 2023). Despite these adaptations, pharmacies predominantly operate as small and micro-sized enterprises (Rough et al., 2021; Policarpo et al., 2019), where entrepreneurial decision-making significantly influences their growth and operational strategies (Jambulingam et al., 2005; Lombardi et al., 2021). The non-ancillary nature and role demonstrated by pharmacies in the healthcare ecosystem, coupled with the increasing prevalence of digital solutions in this sector and the relevance to value co-creation practices, can lead to better health outcomes (Schiavone et al., 2021). Trenfield et al. (2022) highlight how technologies such as AI, Blockchain, interconnected healthcare systems, and virtual, augmented, and mixed realities are poised to transform the pharmaceutical industry comprehensively. AI enhances the efficiency of drug development, aids in patient monitoring, predicts prognosis, and supports surgical procedures. It enables more effective discovery and development of new drug trials, facilitates patient recruitment, and improves surgery outcomes. In the interconnected world of healthcare, several technologies are particularly relevant for pharmacies. These include telemedicine, telehealth, e-prescriptions, e-dispensing, electronic health records, virtual consultations, home delivery of medications, remote clinical data utilization for pharmaceutical monitoring, and geolocation services for enhanced patient tracking and optimized delivery solutions (Aungst et al., 2021). These services and processes are ripe for AI-driven advancements. Despite off, the adoption of AI solutions in the healthcare ecosystem and pharmaceutical industry faces various challenges, including inadequate technological and financial resources and low levels of digital literacy among stakeholders (Viegas et al., 2019; Samson et al., 2022). Furthermore, there remains a notable gap in studies focusing on AI adoption within organizations, especially micro and small firms in the healthcare sector.

Studies investigating the impact of entrepreneurial orientation (EO) on an organization’s adoption of new technologies, as highlighted by Kumar et al. (2021), are limited. The literature on entrepreneurial orientation predominantly focuses on the organizational level, emphasizing its role in guiding technology strategies in micro and small firms (Peltier et al., 2012). The role of CEOs (Fretschner et al., 2022) and entrepreneurs (Upadhyay et al., 2023) is crucial in technology adoption and strategy formulation within firms. Although EO is often synonymous with individual entrepreneurial orientation (IEO) (Kollmann et al., 2007), which can influence organizational behavior (Covin and Wales, 2019), academic literature (Adeniyi et al., 2024; Santos et al., 2020; Kraus et al., 2019) underscores the importance of studying EO at the individual level rather than merely its organizational application. Individual entrepreneurial innovation is positively influenced by a combination of exogenous and endogenous factors (Kariv et al., 2025). Commonly studied elements of IEO include three factors: risk-taking tendency, innovativeness, and proactiveness (Wales et al., 2013). Adeniyi et al. (2024) demonstrate that IEO, particularly innovation and proactiveness, impact entrepreneurial readiness, defined as the ability to respond to entrepreneurial activities (Darmasetiawan, 2019). IEO also influences performance, affecting both competitive strategy (Fatima and Bilal, 2020) and internationalization strategies (Forcadell and Úbeda, 2022). Peltier et al. (2009, 2012) identified that factors such as the owner’s age and education can impact technology adoption. Similarly, the perceived benefits and cost reductions associated with a particular technology can also influence the owner to adopt it (Goslar, 1987). Based on their findings, the authors proposed the SBAM model, which highlights that technology adoption by an owner is influenced by product class knowledge, attitude toward change, decision-making demographics, relative advantage, switching costs, market uncertainty, and environmental hostility. However, they did not investigate the intrinsic nature of an entrepreneur. However, there is still a paucity of studies on the impact of IEO on innovation and technological adoption despite its potential to affect digital strategy (Ritala et al., 2021; Peltier et al., 2012). Despite the limited number of studies focusing on individual entrepreneurial orientation (IEO), its significant role in the adoption and transformation within enterprises has been demonstrated. For instance, Alnoor et al. (2024) highlighted its impact on a company’s digital strategy. Similarly, Gupta et al. (2016), drawing from the Technology Acceptance Model (TAM), emphasized the role of individual entrepreneurial orientation in facilitating technology adoption under mandated conditions. Furthermore, Gumbi and Twinomurinzi (2025) underscored its importance in fostering and sustaining an innovation-friendly culture that facilitates technology adoption. However, these studies have largely remained confined to examining the traditional three dimensions of entrepreneurial orientation. In contrast, this study adopts the IEO conceptualization proposed by Santos et al. (2020). The authors advocate for the inclusion of additional dimensions that drive entrepreneurs, such as passion and perseverance (Gerschewski et al., 2016), comprising these five elements for an individual-level assessment. Howard and Floyd (2024) have identified certain critical elements within this scale and recommended further analyses to enhance its conceptualization.

The profession of pharmacist has ancient, centuries-old roots. In Antiquity and the Middle Ages, the functions of the pharmacist were often performed by apothecaries or herbalists, figures who combined knowledge of botany and chemistry to prepare natural remedies. In times such as the Renaissance, apothecaries acquired a more defined role within communities, gradually separating themselves from physicians and distinguishing themselves as experts in the preparation and sale of medicines. In the 17th and 18th centuries, with the birth of the first official pharmacopoeias, the pharmacist’s role became more regulated. During the 19th century, the profession continued to evolve with industrialization and large-scale production of drugs, marking the transition from artisanal preparations to standardized medicines (Levey, 1973; Parascandola, 1983; Anderson, 2005). Pharmacies thus became focal points for public health, and pharmacists gained increasing recognition as health professionals. Over the course of time, the figure of the pharmacist and the role of pharmacies have evolved significantly, moving from being simple distributors of medicines to true community health hubs. The figure of the pharmacist is, therefore, closely linked to an ethical mission that is also enshrined in the professional oath, which commits the pharmacist to protect public health and to act with science and conscience while respecting professional dignity (Fino et al., 2022). Originally, pharmacists focused mainly on the preparation and distribution of medicines, but over time, especially since the second half of the 20th century, their role has expanded to include health advice and preventive activities. This transformation has been accelerated in recent decades thanks to servitization and digitalization, which have introduced new tools and services, such as teleconsultation, vaccinations and rapid diagnostics. Thus, pharmacies have begun to take on a dual function: on the one hand, they continue to be commercial players in the retail market; on the other, they fulfill an important social mission within the healthcare system. Thanks to their widespread presence in the territory, they have become essential points of reference for the community, facilitating access to health services in a continuous and immediate way (Esmalipour et al., 2021; Toma and Crisan, 2022). In parallel, the intensification of competition and the expansion of digital technologies have confronted pharmacists with new ethical dilemmas, including the management of often opaque online channels and the need to balance commercial and social objectives. This challenge requires pharmacists to develop both digital and health skills in order to be able to offer quality services that are environmentally and socially sustainable (Singleton et al., 2014).

The study employs a mixed-methods approach (Teddlie and Yu, 2007; Mukhopadhyay and Jain, 2024; Grošelj et al., 2020), structured into two distinct phases. The first phase involved qualitative research through semi-structured interviews conducted with eleven Italian pharmacies. Semi-structured interviews were chosen to explore the primary factors influencing pharmacists’ decisions regarding AI adoption (Awa et al., 2010; Davis et al., 1989) and arrive at the choice of the best model to test and the formulation of hypotheses. Drawing from both the interview findings and theoretical insights, a conceptual model based on the TOE framework was developed. A quantitative analysis was undertaken to empirically test the model and hypotheses, employing linear regression analysis (Ye et al., 2019) to explore the influential variables across the TOE Model and Entrepreneurial Orientation dimensions. Detailed findings from this analysis are presented in the subsequent dedicated sections.

The sample selection employed convenience sampling (Campbell et al., 2020), dividing participants into two groups: current adopters of AI solutions in their pharmacies (n = 5) and potential adopters (n = 6) of Italian AI software for pharmacy (RoiStar). The interviews were conducted between November 2022 and January 2023 online in order to accommodate the different schedules of the pharmacists, to ensure their comfort during the interview process and to allow reaching even the most remote pharmacies in remote areas, Table A1 in  Appendix 1. This approach was designed to ensure that participants were comfortable and could provide thoughtful, in-depth responses. On average, each interview lasted approximately 42 min. Italian pharmacists were selected due to Italy’s extensive network of pharmacies, predominantly owned by individual pharmacists, facilitating a clearer measurement of their entrepreneurial orientation and its impact on business decisions (Federfarma, 2022). The interview guide consists of open-ended questions to pharmacists, focusing on the dimensions they perceived as most influential in shaping AI adoption within their businesses on the basis of the cited literature on adoption. The interview framework was guided by several dimensions: technological (both at the individual and enterprise levels), organizational (structured by individual and enterprise-wide), and environmental/external factors (considered at both organizational and individual levels). To mitigate potential interviewer bias and reduce cognitive and reflexivity biases on the part of the authors (Berger, 2015), each interview was conducted by two interviewers. This approach ensured diversity in questioning techniques and interpretation, enhancing the reliability of the data. Additionally, strict adherence to a predefined interview guide was maintained to ensure consistency across all interviews. The interviews conducted online were recorded and then transcribed; the resulting verbatims from the recording formed the basis of analysis for the encoding process.

The interviews’ coding utilized a deductive approach (Azungah, 2018; Braun and Clarke, 2012), starting from established technology acceptance models to identify the most applicable for AI adoption in pharmacies. The coding process employed thematic analysis (Braun et al., 2022), allowing for a synthesis of inductive insights alongside deductive reasoning, essential for exploring new dimensions pertinent to the phenomenon under study. The analysis involved repeatedly reviewing each interview, marking key phrases and sections that were relevant to the main research goal, and arriving step by step at greater abstraction capable of grasping second order and first-order concepts that can define the dimensions of the model. After the coding procedure, the authors shared themes and findings, commenting and comparing them with the other researchers to remove and try to reduce the influence of pre-existing knowledge and other authors’ personal unconscious influences (Sobh and Perry, 2005). Based on the qualitative interviews, the study identified the Technology-Organization-Environment (TOE) model as the most suitable framework for understanding AI adoption among pharmacies (Katebi et al., 2022). The qualitative analysis provided insights into the specific dimensions within each construct of the TOE model that are most relevant for AI adoption and lead to the formulation of the hypothesis. Within the Technology dimension, guided by Rogers’s dimensions (1995) of technology adoption—namely relative advantage, compatibility, complexity, trialability, and observability—only certain dimensions emerged as significant for the adoption of AI and now will be exposed.

Specifically, technological complexity was highlighted as a critical factor. Technological complexity refers to the perception that a technology is difficult to comprehend and use (Rogers, 1995).

Compatibility measures consistency with other technologies, values, needs, and past experiences of the potential adopters and emerged as another relevant and potential influent factor.

Other relevant dimension is firm’s relative advantage which is the benefit, and the potential advantage of innovation and new technology is also related to its costs (Cobos et al., 2016; Oliveira et al., 2014; Pizam et al., 2022) and as mentioned by the interviewers:

On this basis, we develop the following hypothesis:

Regarding the construct of Organization, the respondents reported that the following dimensions, such as Organizational Readiness, pertain to the availability of resources, such as human resources in adopting new technology and the capacity of a firm to embrace the innovation and related processes (Misuraca and Viscusi, 2020), the interviews in fact position that:

But it also refers to the Manager’s Support, which aligns with the support, commitment, and belief of top managers in the process of adopting new technology. This concept was proposed by Pizam et al. (2022), and as also highlighted by Chen et al. (2022), it helps to reduce conflicts and create a better environment for adoption, as emerged:

The IT Resources possessed by the enterprise, in accordance with Chittipaka et al. (2023) and Gokalp et al. (2020), represent the existing firm’s technical infrastructure and capabilities:

The following hypothesis related to the Organizational dimension emerged:

While on the Environment Construct, social influence emerged as a relevant theme. Usually, the model also measures the impact of regulations and legislation. To date, the choice of whether or not to adopt AI systems does not appear to be in the interest of the legislator, and the final decision remains to the entrepreneur and the organization; even during the interviews, government support was never mentioned or even as a possible threat to adoption. Due to this, the relevant themes that emerged for the Environment dimension are the influence of Business Partners, that is, the pressure from business partners as providers, suppliers and other business relationships and collaborations (Chittipaka et al., 2023), and as testified by the interviews:

The perceived Competitive Pressure, albeit the firm’s pressure perceived by its competitors (Halpin, 2010) and so the voluntariness to achieve a competitive advantage (Shi and Yan, 2016), coincides with our interviews:

Thus, the third hypothesis related to Firm’s Environment is the following:

In addition to the traditional dimensions of the TOE model, it was extended. In fact, the following was inserted in the role of entrepreneurial orientation as a potential strategic dimension in the final decision to adopt a new technology. This is due to the role and attitude of the entrepreneur who, as emerged during the interviews, is usually the one directly called upon to make the final decision regarding the adoption of new technology in micro and small firms and guide not only by innovation, risk attitude and proactiveness but also from passion and perseverance in business, so the following hypothesis was defined:

Perceived risk, as emerged from interviews and supported by the present academic literature, was also included. Perceived risk has been considered over time at different levels, Kateby et al. (2022) and Cao et al. (2018) insert it as a component of the technology dimension; since it emerged relevant in two interviews with potential adopters, we decided to include but to not consider as a part of the different constructs and this is in line with Stewart (2022).

In the final part of the qualitative interviews, participants were asked to outline their expectations regarding the impact of AI systems on supply chain risk mitigation. To identify the variable, the scale developed by Nayal et al. (2023), which assesses the potential positive influence of AI on supply chain risk mitigation is used. The dimensions identified during the qualitative investigations and from the literature review were adopted to develop the adoption model, Figure 1.

As can be seen from the model, there is the business and TOE model dimension broken down into its key dimensions of technology, organization and environment and the concepts that emerged that substantiate them; then there is the individual entrepreneurial mindset dimension and the integrated dimensions as proposed in Gerschewski et al. (2016) and finally the risk dimension that emerged as a construct that could negatively influence adoption.

In the next paragraphs, the measurement of the model that guided the quantitative study will be presented.

Based on the constructs and the different dimensions posed as hypotheses for each of them, specific measurement scales were identified from the literature and taking the behavioral intention to adopt AI systems as the dependent variable, see Table A2 in  Appendix 2. A Likert scale from 1 (totally disagree) to 7 points (totally agree) was used to measure the different dimensions. The regression model shown in Equation (1) and Equation (2) was developed to test the conceptual model and hypotheses proposed:

Pharmacies have a unique nature of being typically single owner guided, a survey was conducted on a large sample of pharmacies to validate the reliability of the model and propose an initial reflection on firms’ adoption models. The sample for this research was identified with the collaboration of the Order of Italian Pharmacists of Rome, which plays a crucial role in safeguarding public interest by ensuring that individuals providing pharmaceutical services possess the necessary qualifications. This collaboration not only enhances the credibility of the research but also ensures that the participants are representative of the professional community, adhering to the standards set by the Order. The questionnaire was constructed using a Google Form and distributed by the Order of Pharmacists through their newsletter. This approach involved sending two emails to their members between June and August 2023, allowing for a broad reach within the pharmacist community. Importantly, the anonymity of respondents was strictly maintained throughout the process, ensuring that individual responses could not be linked back to participants. This adherence to confidentiality is in line with the ethical standards outlined by the Order’s Code of Ethics, which guarantees the protection of clients and colleagues and attention to remain active in the business.

Out of 500 invited members, 217 respondents completed the survey, yielding a response rate of 44%. This sample size aligns with previous studies in the field, such as Asiaei and Ab Rahim (2019), who analyzed AI adoption among 209 small and medium-sized enterprises, Stewart (2022), who surveyed 208 bank employees to integrate the TOE and TAM frameworks, and Katebi et al. (2022), who validated their model with a sample of 188 respondents. While the response rate introduces the possibility of non-response bias, it is also plausible that those who participated are among the most engaged and knowledgeable professionals in the sector. Their perspectives may, therefore, be particularly valuable in assessing the readiness and barriers to AI adoption, as they likely represent early adopters or decision-makers most inclined to consider technological innovation as in became the agent of change. The average age of respondents was 52 years, with 53% being women, 41.6% men, and 1.6% preferring not to disclose their gender. Regarding educational qualifications, 89% held a single-cycle degree, consistent with their age, while the remaining respondents had a three-year degree or other specialist training. The respondents had been employed in the sector for an average of 26 years. Since AI implementation frequently requires advanced digital literacy and analytical skills, the perspectives of this cohort are particularly relevant in assessing the feasibility of AI integration in pharmacy operations. The study also focuses on pharmacies, an industry with distinct regulatory and operational constraints that may limit the generalizability of findings to other sectors. Most companies have not yet implemented AI systems (78%). The most common type of pharmacy was small and micro-sized (micro: 46.2%, small: 22.6%, medium-large: 27.1%), with a wide range of histories and types. The youngest pharmacy was 3 years old, while the oldest had been in operation for over 150 years. The average age of the pharmacies was 38 years, with a mode of 50 years and a median of 30 years. However, the pharmacy sector serves as an ideal case study for understanding AI adoption in highly regulated environments, where technological solutions must navigate strict compliance requirements. Nevertheless, smaller enterprises typically face greater financial and infrastructural challenges in adopting AI, making their inclusion crucial in identifying barriers that must be addressed to facilitate widespread adoption. Additionally, nearly half of the respondents’ pharmacies provided services beyond traditional pharmacy operations, which might suggest a bias toward more innovative businesses. While this could influence perceptions of AI adoption, it also means that a substantial portion of the sample comprises enterprises already open to innovation, reinforcing the relevance of the study in capturing businesses most likely to lead AI adoption trends. In light of these considerations, while certain biases are inherent in the study’s design, they do not undermine the validity of its findings. Rather, they offer a nuanced perspective on AI adoption, highlighting both the challenges and opportunities present within the sector. The study provides critical insights into the factors shaping technology adoption, particularly within small and micro-sized enterprises, and contributes to a deeper understanding of the mechanisms influencing AI integration in regulated industries.

To assess the validity of the scale, a Cronbach’s Alpha analysis was conducted, yielding values below 0.700, which reassured the scale’s adequacy. Multiple regression analysis was performed using SPSS v.28 software. Two linear regressions were conducted (Aloulou, 2019; Liu et al., 2025). The first regression analyzed the main constructs of the Technology-Organization-Environment (TOE) framework and Individual Entrepreneurial Orientation (IEO). The regression model, executed with the “insert” option in SPSS, confirmed the overall model’s validity (R² = 0.754 and adjusted R² = 0.748; F-test = 129.950). The Durbin-Watson value of 1.937 indicated no problematic autocorrelation. The most significant predictors of behavioral intention to adopt were the Technology level (β = 0.430, 99% confidence interval) and the Environment (β = 0.303, 99% confidence interval). Risk was significant but negative (p = 0.041, β = −0.088), followed by EOI (β = 0.142, 95% confidence interval) at a 90% confidence interval. The Organization construct was not significant (p = 0.212, β = 0.068). The second regression considered the different dimensions of each main TOE construct using the “insert” option in SPSS. This regression model also confirmed its validity (R² = 0.768 and adjusted R² = 0.757; F-test = 68.701). The Durbin-Watson value of 1.995 indicated no problematic autocorrelation (see Tables 1 and 2). The most significant predictors of behavioral intention to adopt, at a 99% confidence interval, were Compatibility (β = 0.251) and Competitive Pressure (β = 0.212). At a 95% confidence interval, Individual Entrepreneurial Orientation (β = 0.147) was significant, and at a 90% confidence interval, complexity (β = 0.110) and Firms’ IT resources (β = 0.117) were significant. Other variables resulted as not significant (see Table 3).

The study aimed to identify the dimensions most significantly influencing the adoption of AI solutions among small and medium-sized enterprises (SMEs) in the healthcare sector, specifically pharmacies. The results of the qualitative analysis indicated that the Technology-Organization-Environment (TOE) framework is the most suitable model to be tested, with the relevant dimensions integrated accordingly. The dimensions related to Technology and the Environment play a significant role, confirming hypotheses H1 and H3. However, hypothesis H2 is not supported, as the role of the Organization was not found to be really significant. The analysis of aggregate measures revealed that compatibility and competitive pressure—particularly the fear of falling behind competitors—are the most significant factors driving the adoption of AI. Although the dimensions related to the Organization were not significant, the technological resources of the company and the perceived ease of implementing and training employees on new AI solutions should not be underestimated. These findings partially align with Tornatzky and Fleischer (1990), who identified relative advantage, compatibility, and complexity as the most consistent factors related to innovation adoption. The strategic adoption of AI systems is driven by competitive pressure and the enterprise’s digital capabilities, both hard and soft, enabling the creation of an integrated system that enhances value. However, this investment decision is not always viable for micro and small enterprises. Dewar and Dutton (1986) noted that larger organizations are more prone to adopt technologies due to their resources. Micro and small enterprises can overcome this obstacle by investing in their digital capabilities and infrastructure. In addition to the TOE model measures, the dimension of entrepreneurial orientation at the personal level was integrated, considering the pharmacist’s potential role in guiding technology adoption in the pharmacy. This variable has been relatively underexplored in technology adoption models, especially regarding the dimensions of passion and perseverance (Gerschewski et al., 2016; Santos et al., 2020). The sample data confirmed hypothesis H4 as significant and positive in influencing the adoption of AI. Consistent with findings for family firms by Upadhyay et al. (2023), small firms also need to act entrepreneurially and accept uncertain market outcomes (George and Marino, 2011; Parveen et al., 2015). Hypothesis H5, related to risk, is not supported. The measure did not appear significant, possibly due to a lack of familiarity with AI solutions, given their limited implementation. This finding aligns with the qualitative analyses, which indicated few mentions of risk and a greater focus on implementation benefits and challenges, depending on leadership. Interviewees were also asked about the expected outcomes from adopting such technologies, addressing the research gap identified by Venkatesh et al. (2022). The questionnaire results indicated that the expected outcomes from AI adoption in micro and small enterprises fall into three main dimensions: changes in business strategy and business model, improved predictive ability in terms of orders and customer offerings and third: enhanced agility, in fact AI systems are expected to improve decision-making capabilities, flexibility, learning, knowledge management, supply chain integration (Nayal et al., 2023), and the reliability of pharmacy operations.

On a theoretical level, this study is among the first to shed light on adoption models in enterprises, specifically examining the application of the TOE Model in micro and small enterprises. These businesses are often underestimated, despite remaining one of the most prevalent forms of entrepreneurship today. The study evidence as perceived risk for firms appears as not significant; AI, with its computing power and “reasoning” capabilities, is anticipated to transform the predictive capabilities of enterprises (Burstrom et al., 2021), enhance organizational agility in decision-making and time management, and improve operations, knowledge management, and flexibility, this could not be perceived as risk but as potential. AI impact extends beyond automation and augmentation (Krakowski et al., 2023) to the realms of strategy and business models, facilitating more customized and sustainable offerings (Sjodin et al., 2023).

We extend the existing theory as this is one of the first studies to consider the individual’s entrepreneurial mindset as a significant factor in the decision-making process of technology adoption within a firm. As demonstrated by the TOE framework, multiple factors influence technology adoption; however, the role of individual entrepreneurial orientation has been largely overlooked. By integrating this perspective, we provide a novel approach that emphasizes how individual-level cognitive traits can influence organizational behavior, thereby enriching the stream of literature on firms’ adoption models and contributing to a more comprehensive understanding of technology adoption dynamics in SMEs.

Moreover, at the theoretical level the study expands our knowledge of what outcomes are expected to result from the implementation of AI solutions. The outcomes expected from the adoption of a new technology as pointed out earlier and by Venkatesh et al. (2022) himself have so far not been thoroughly investigated while often what motivates adoption is precisely the expectations and promises inherent in that technological solution. In light of this evidence, it follows that the theory adopted then strongly influences the decisions of managers and those who develop and sell technology solutions, they, thanks to theory, know which dimensions in general most influence adoption, this study has highlighted which according to the sample investigated most can be strategic levers to wield to ensure greater and easier adoption of AI solutions.

At the managerial level, the study highlights the need for creators and developers of AI solutions to focus on simplicity and compatibility. Both developers and entrepreneurs must evaluate whether they possess the technological and organizational resources necessary to support and integrate these innovations within their infrastructure before adoption. Developers should leverage the predictive capabilities of such algorithms, utilizing generative AI activities to stay competitive and guard against new market entrants. AI technologies can streamline various operational processes, such as inventory management and value chain optimization, especially in logistic functions, such as customer interaction, making pharmacies more efficient and cost-effective. By using predictive algorithms, pharmacies can forecast demand more accurately, reduce stockouts, and ensure the availability of critical medications and more customized offers. This leads to better inventory control, reduced waste, and improved economic performance, which are crucial for gaining a competitive edge. The perceived benefits of AI adoption will outweigh any initial hesitancy and costs associated with acquiring new skills. AI’s ability to analyze vast amounts of data in real-time allows pharmacies to track emerging trends, respond quickly to market changes, and even, thus, potentially support the health and social policies of a state and adapt to customer needs more effectively. This data-driven approach ensures that pharmacies remain agile and innovative, further contributing to their long-term competitive advantage. The focus on entrepreneurial orientation highlights the need to leverage multiple dimensions that cannot be reduced to the three commonly investigated ones. Instead, it is essential to also consider passion and perseverance, which drive an entrepreneur to continue leading and believing in their business. These factors, in turn, increase the likelihood of adopting new solutions within the enterprise. Consequently, developers and managers must take these dimensions into account and implement business strategies and communication approaches that effectively resonate with these motivational levels.

Entrepreneurs and pharmacists, on the other hand, must have the knowledge, skills, and entrepreneurial spirit to capitalize on these opportunities. Given that AI systems are capable of “reasoning”, many operations could be delegated to them. Small and Micro enterprises, such as pharmacies, which are technologically advanced, will find it challenging to adopt AI solutions if the benefits, compatibility, ease of use, and required knowledge are not evident. In this regard, support from policymakers could be significant. For instance, better stock management facilitated by AI could reduce pharmaceutical wastage and consequently decrease state expenditure; thus, policies promoting such benefits could positively impact the entire supply chain. Moreover, policymakers can play a crucial role in harmonizing regulations across different regions, ensuring that AI-driven service provision models are consistently applied and effectively integrated into pharmaceutical practice. By creating regulatory frameworks that facilitate AI adoption, policymakers could enhance service quality, optimize healthcare costs, and foster a more resilient pharmaceutical sector. Standardized policies could also encourage investments in AI-driven pharmacy management systems, ensuring that even smaller businesses can leverage these technologies without facing disproportionate regulatory or financial burdens. Pharmacists, and more generally, leaders of micro and small enterprises, must strengthen their entrepreneurial mindset to remain alert to innovations and ready to implement them when appropriate. The organizational level of companies does not seem to significantly influence AI adoption, nor does the associated risk, which may be attributed to the current low diffusion and utilization of these technologies and the lack of resources required to build an efficient and functional AI system. Supported by an appropriate regulatory support and strategic investments, AI could become a transformative force in pharmacy service provision, leading to more efficient operations, enhanced patient care, and a more sustainable healthcare ecosystem.

The present paper is the first to provide insight into the progression of AI adoption in the pharmacy sector, specifically highlighting its uptake by micro and small firms. Its primary contribution to the existing theory lies in evaluating, for the first time, the impact of individual entrepreneurial orientation as an adjunct to the generally more organizationally focused Technology-Organization-Environment (TOE) framework. Consequently, the symbiosis between the enterprise and the entrepreneur emerges as a compelling theme: innovation within the enterprise, particularly in micro and small enterprises such as pharmacies, remains driven by entrepreneurial impetus.

This underscores the cardinal role of the entrepreneur in fostering innovation and the organizational necessity to align with and reflect the entrepreneur’s identity. Future research should analyze the influence of entrepreneurial orientation as a moderator of technology adoption within the TOE model and better capture the potential impact of various dimensions, particularly on AI outcomes, such as sustainability dimensions, organizational behavior, strategic management dimensions, and performance measures in general. It would also be interesting to investigate whether technology acceptance differs between urban and rural pharmacies. While our study did not reveal such differences, there may still be varying incentives and motivations for innovation in these contexts. Future research could further explore how adoption varies across regions or follows other patterns. Future research should expand into different countries and sectors, incorporating longitudinal studies to track changes over time. The application of AI technologies varies significantly depending on the specific business function and medical specialization being addressed—whether in administration, radiology, oncology, or other areas. Each of these domains entails unique challenges and acceptance models, thereby requiring tailored approaches. Consequently, further research is necessary to appropriately account for these differences in future studies by analyzing the new key determinants and impacts of AI adoption and diffusion in other firms of the healthcare industry, such as clinics and similar SME healthcare facilities, as analyzing differences between medical practices.

Artificial Intelligence is restructuring our environments, relationships, processes and activities, including those related to management. Like other technologies and solutions, its adoption is not so obvious; obstacles and barriers to proper implementation may delay its diffusion or give rise to incorrect or costly implementation practices. Adoption of new technology by small and medium-sized enterprises is also not as easy as it might be for larger companies with greater spending, investment and risk management capabilities. Small and medium-sized enterprises, therefore, risk always being one step behind in the implementation of technological solutions capable of building and strengthening their competitive advantage as much as essentially guaranteeing their sustainability, which, in the case of pharmacies, offering a public utility service and fulfilling an important social function, therefore remains a matter of great concern. The study conducted has not only made it possible to gain an in-depth understanding of which dimensions of the TOE model are most likely to drive pharmacy companies to adopt such technologies, but has empirically tested that the most significant and relevant determinants are at the technological level: compatibility with other management tools so as to offer complementary, comprehensive management without bottlenecks and redundancies, all elements, on the other hand, capable of stressing and delaying the operations of any business; at level of resources possessed by the business, not so much monetary as technological, but also system and in terms of human resources are another strategic dimension to facilitate adoption. The perceived risk in adoption is confirmed to be negatively reducing willingness and potential speed of acceptance. In the case of pharmacies, organizational readiness is irrelevant; on the other hand, Italian pharmacies are experiencing a boost in terms of digitalization, which is already reflected in the good digital readiness of these enterprises. This, combined with their smaller and more flexible size, their role as drug distributors and, therefore, customer-oriented organizations, allows them to adapt more easily to adopt new technological systems, as has already been the case with robotic systems and telemedicine services. The role of business partners is not relevant, highlighting how this choice is often motivated more by internal reasons of strengthening the company than by mere relational stimulus, which is also confirmed by the significant weight that competitive pressure revealed to have. The study also tested the role of individual entrepreneurial orientation in the adoption of a company’s technology. In the case of pharmacies and, by extension, in small and even more micro-enterprises, the final decision on whether to adopt technology is often determined by the entrepreneur or owner. This study is the first to confirm the significant role that the good father’s final word plays in deciding on a technological investment.

We would like to express our sincere gratitude to Dr. Eugenio Mealli and his ROISTAR team for their invaluable support throughout the analytical process. Their pioneering work in developing one of the first AI-driven solutions for pharmacies provided essential insights to the success of this study.

Funding: This research was made possible through the funding provided by Sapienza Università di Roma under the project RM12117A8AAFF305, titled “Ateneo 2021 – The pharmaceutical e-commerce development through retail digitalization and new management models”. The authors would like to express their gratitude for the support extended by this initiative. The project was assigned the Unique Project Code (CUP) (No: B89I22000090001).