Artificial Intelligence in Small and Medium-Sized
Enterprises: A Systematic Analysis of
Adoption Barriers and Success Factors in
German SMEs Using the
TOE-DOI Framework
A systematic analysis of adoption barriers and success factors in German SMEs using the TOE-DOI framework
It is hardly possible to overestimate the importance of small and medium-sized enterprises to the German economy. Accounting for over 99 percent of all companies and approximately 60 percent of employees subject to social security contributions, the Mittelstand significantly contributes to employment, innovation, and economic value creation (Schwäke, Peters, Kanbach, Kraus & Jones, 2024; Müller, Buliga & Voigt, 2021). The European Commission defines SMEs as enterprises with fewer than 250 employees and an annual turnover not exceeding 50 million euros or a balance sheet total of maximum 43 million euros (European Commission, 2003). In Germany, approximately 3.5 million enterprises fall into the SME category according to this definition, together generating about 55 percent of total economic output (Destatis, 2023).
In the face of increasing globalization, changing customer expectations, and technological disruptions, these companies are under growing pressure to implement digital technologies to secure their competitive position (Ulrich, Frank & Kratt, 2021; Vial, 2019). The Covid-19 pandemic further intensified this pressure, highlighting the necessity of digital business models and automated processes in an unprecedented manner (Priyono, Moin & Putri, 2020). Companies that had already invested in digital infrastructure before the pandemic demonstrated their resilience, while others faced significant operational challenges (Soto-Acosta, 2020).
Artificial Intelligence has emerged as a particularly promising technology in this context. The term encompasses a broad spectrum of technologies, from rule-based expert systems to machine learning and deep neural networks capable of recognizing complex patterns in data and making predictions (Russell & Norvig, 2021). Scientific studies demonstrate that AI applications can unlock significant productivity gains, process optimizations, and innovation potentials for enterprises of all sizes (Enholm, Papagiannidis, Mikalef & Krogstie, 2022; Davenport & Ronanki, 2018). McKinsey Global Institute estimates that AI technologies could generate an additional global economic output of 13 trillion US dollars by 2030 (Bughin et al., 2018).
Baabdullah (2024) notes that smaller organizations can benefit from AI-supported decision processes to a similar extent as multinational corporations, provided implementation is carried out adequately. Typical application areas in SMEs include automation of customer service inquiries through chatbots, optimization of inventory through demand forecasting, quality control using image recognition algorithms, and personalized marketing measures based on customer behavior data (Chalmers, MacKenzie, Carter & Quail, 2020). Brynjolfsson and McAfee (2017) argue that AI has the potential to trigger fundamental changes in value creation, comparable to the upheavals caused by the steam engine or electrification.
However, empirical reality paints a more nuanced picture. Despite proclaimed potential, the actual adoption of AI technologies in SMEs remains sluggish and is characterized by high failure rates (Hansen & Bøgh, 2021; Ransbotham, Kiron, Gerbert & Reeves, 2017). A Gartner study predicted that 85 percent of all AI projects would fail to achieve their intended objectives (Gartner, 2019). Particularly for the German Mittelstand, Ulrich et al. (2021) attest to a pronounced reluctance in AI adoption, attributable to delayed policy support measures and lower digital maturity compared internationally. A quantitative survey of 283 German SMEs in 2020 revealed that numerous companies have not yet fully recognized the transformative potential of AI and predominantly rely on proven, rule-based systems (Bley, Leyh & Schäffer, 2016).
This discrepancy between theoretical potential and practical implementation raises fundamental questions: What specific obstacles slow AI adoption in SMEs? What organizational, technological, and contextual factors distinguish successful from failed implementations? And how can companies, technology providers, and policymakers collectively contribute to overcoming the identified barriers?
Against this background, the systematic investigation of inhibiting and promoting factors of AI adoption in SMEs is gaining increasing scientific and practical relevance. Existing studies suggest a complex bundle of barriers, ranging from insufficient data infrastructure and skills shortage to cultural resistance and unclear profitability prospects (Ghobakhloo, Iranmanesh, Vilkas, Grybauskas & Amran, 2022; Sánchez, Calderón & Herrera, 2025; Alsheibani, Cheung & Messom, 2018). The skills shortage in data science and AI development represents a particularly severe problem: According to Bitkom estimates, Germany currently lacks over 137,000 IT specialists, with demand for AI specialists growing particularly strongly (Bitkom, 2024).
At the same time, initial success analyses identify potential enablers: engaged leaders with a digital vision, an agile error culture, external consulting partnerships, and the use of low-code and no-code technologies that enable AI applications without deep programming knowledge (Liwanag, Ebardo & Cheng, 2025; Wamba-Taguimdje, Wamba, Fosso Wamba & Tchatchouang Nguegang, 2020). These democratizing technologies have the potential to significantly lower entry barriers for SMEs by reducing the need for specialized programming knowledge and shortening development cycles (Sahay, Indamutsa, Di Ruscio & Pierantonio, 2020).
For the structured analysis of the diverse influencing factors, this study draws on two established theoretical frameworks: the TOE model (Technology-Organization-Environment) according to Tornatzky and Fleischer (1990) and selected constructs from the Diffusion-of-Innovations theory according to Rogers (2003). The TOE framework enables a systematic consideration of technological, organizational, and environmental determinants of technology adoption (Chatterjee, Rana, Dwivedi & Baabdullah, 2021; Baker, 2012). Additionally, DOI theory considers perception-related attributes such as relative advantage, complexity, compatibility, trialability, and observability, which are of considerable importance particularly in owner-managed SMEs with centralized decision structures (Horani et al., 2023; Frambach & Schillewaert, 2002).
This paper addresses the following research questions: (1) What main factors inhibit or promote the adoption of AI in German SMEs? (2) How do technological, organizational, and environmental factors interact within systemic dynamics? (3) What role do newer technological approaches such as low-code and no-code platforms play in overcoming adoption hurdles? The scientific contribution lies in consolidating the current state of knowledge, its theoretically grounded structuring, and the derivation of implications for research, business practice, and economic policy decision-making.
The TOE framework developed by Tornatzky and Fleischer (1990) has established itself as the dominant theoretical construct for explaining organizational technology adoption (Oliveira & Martins, 2011; Zhu, Kraemer & Xu, 2006). Compared to individual-centered models such as the Technology Acceptance Model (Davis, 1989) or the Theory of Planned Behavior (Ajzen, 1991), the TOE framework explicitly focuses on organizational decision processes and their contextual embedding. It posits that the decision to adopt a technology is determined by three interdependent contextual dimensions: the technological, organizational, and environmental context (Chatterjee et al., 2021; Baker, 2012).
The technological context encompasses both internally available and market-available technology. Relevant variables include perceived usefulness, technical complexity, compatibility with existing systems, and availability of necessary technical infrastructure (Ghobakhloo et al., 2022; Awa, Ukoha & Emecheta, 2016). In the context of AI technologies, additional factors such as data availability, model accuracy, explainability of algorithmic decisions, and integration into existing IT landscapes play a central role (Mikalef & Gupta, 2021). The rapid advancement of AI technologies, particularly in the area of generative AI and large language models, adds an additional dimension of dynamism and uncertainty to this context (Dwivedi et al., 2023).
The organizational context refers to internal characteristics of the enterprise, including company size, management structure, available resources, innovation culture, and organizational readiness for change (Arroyabe, Arranz, de Arroyabe & de Arroyabe, 2024; Ramdani, Kawalek & Lorenzo, 2009). For SMEs, specific characteristics are relevant: flat hierarchies enabling rapid decision processes, but also strong dependence on individual key personnel, limited slack resources for experimental projects, and often informal organizational structures that can mean both agility and vulnerability (Welsh & White, 1981; Nooteboom, 1994). Absorptive capacity, understood as a company's ability to recognize, assimilate, and commercially exploit external knowledge, represents a critical organizational factor (Cohen & Levinthal, 1990).
The environmental context addresses external factors such as competitive intensity, industry dynamics, regulatory frameworks, availability of external support, and pressure from customers and suppliers (Horani et al., 2023; Zhu & Kraemer, 2005). In the current debate on AI regulation, this context gains particular relevance: The EU AI Act, which came into force in August 2024, establishes a risk-based regulatory framework that defines extensive compliance requirements particularly for high-risk AI applications (Veale & Borgesius, 2021). For SMEs, this means greater legal certainty on one hand, but also potential compliance costs that could disproportionately burden smaller companies.
The TOE framework has been frequently applied and empirically validated in the SME context (Sánchez et al., 2025; Thong, 1999; Premkumar & Roberts, 1999). A meta-analysis by Jeyaraj, Rottman and Lacity (2006) identified top management support, external pressure, and IS department professionalism as the most consistent predictors of organizational IT adoption. A significant limitation, however, is that individual perceptions and cognitive biases of decision-makers are not explicitly considered, although these are of considerable importance in SMEs with often centralized decision structures (Schwäke et al., 2024).
The diffusion theory founded by Rogers (2003) focuses on the characteristics of an innovation as determinants of its adoption rate (Moore & Benbasat, 1991). In contrast to the TOE framework, which primarily considers contextual factors, DOI theory directs attention to the perceived properties of the innovation itself and their influence on individual and organizational adoption decisions. Rogers identifies five central innovation attributes: the relative advantage over existing solutions, the perceived complexity of use, compatibility with existing values and practices, trialability before full implementation, and observability of results for potential adopters (Baabdullah, Alalwan, Slade, Raman & Khatatneh, 2021; Agarwal & Prasad, 1998).
Relative advantage describes the extent to which an innovation is perceived as better than the solution it replaces. This can be economic in nature (cost savings, efficiency gains) but can also encompass social prestige aspects or convenience benefits. Empirical studies consistently show that perceived relative advantage is one of the strongest predictors of technology adoption (Tornatzky & Klein, 1982). In the context of AI technologies, this advantage frequently manifests in the form of automation gains, improved decision quality, or opening up new business areas through data-driven products and services.
Perceived complexity refers to the degree of difficulty associated with understanding and using an innovation. Complexity typically acts as an adoption barrier, as it increases the required learning effort and perceived risks (Thompson, Higgins & Howell, 1991). In the context of AI technologies, these attributes prove particularly relevant. The complexity of modern machine learning algorithms, especially deep learning-based approaches, and their lack of explainability constitute significant adoption barriers (Füller, Hutter, Wahl, Bilgram & Tekic, 2022; Arrieta et al., 2020). The so-called black-box problem, where even developers cannot trace the specific decision paths of a trained model, undermines potential users' trust and complicates fulfillment of regulatory transparency requirements (Gunning et al., 2019).
Compatibility describes the extent to which an innovation aligns with existing values, past experiences, and current needs of potential adopters. High compatibility reduces uncertainty and facilitates integration of the innovation into existing routines and processes (Karahanna, Straub & Chervany, 1999). For SMEs with established IT landscapes and business processes, integrating AI solutions frequently poses a considerable challenge, particularly when legacy systems offer no standardized interfaces or data is stored in incompatible formats (Hansen & Bøgh, 2021).
Trialability and observability relate to the possibility of testing an innovation before final adoption, and to the visibility of results to others. Pilot projects and proof-of-concept implementations therefore play an important role in reducing perceived risks and generating organizational learning (Ries, 2011). Conversely, a clearly communicated relative advantage, for example in the form of concrete efficiency gains or cost savings, can significantly increase adoption readiness (Chatterjee et al., 2021; Venkatesh, Morris, Davis & Davis, 2003).
Recent research advocates for an integration of TOE and DOI elements to jointly consider structural and perception-related influencing factors (Sánchez et al., 2025; Hsu, Kraemer & Dunkle, 2006). The combination of both perspectives addresses a significant limitation of isolated models: While the TOE framework captures contextual conditions under which adoption takes place, DOI theory supplements the subjective perception and evaluation of the innovation by decision-makers. This dual perspective appears particularly suitable for the SME context, as adoption decisions are often made by individuals – typically the owner or managing director – whose individual perceptions and preferences exert considerable influence on strategic technology decisions (Thong, 1999).
Horani et al. (2023) empirically demonstrate that an integrated TOE-DOI model has greater explanatory power than isolated application of individual theories. In their study on AI adoption in Jordanian companies, the combined model achieved a variance explanation of 67 percent compared to 48 percent for pure TOE application. This integrative perspective enables a holistic analysis that considers both contextual conditions and individual perceptions of decision-makers (Gangwar, Date & Ramaswamy, 2015). For the present analysis, both theoretical lenses are therefore employed to develop the most comprehensive understanding possible of adoption dynamics.
The investigation was designed as a systematic literature review following Tranfield, Denyer and Smart (2003). Unlike narrative reviews, systematic literature reviews are characterized by an explicit, reproducible methodology that ensures transparency regarding search strategy, selection criteria, and analysis procedures (Kitchenham, 2004). This methodological approach minimizes bias in literature selection and enables structured identification of research gaps as well as synthesis of heterogeneous findings into coherent insights (Schwäke et al., 2024; Webster & Watson, 2002). The systematic approach follows the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), which represent an established standard for conducting and reporting systematic reviews (Moher, Liberati, Tetzlaff & Altman, 2009).
The data basis comprised a structured search in the scientific databases Scopus, Web of Science, and IEEE Xplore for peer-reviewed journal articles and conference papers from the publication period 2020 to 2025 (Levy & Ellis, 2006). The choice of this period is justified by the rapid development of AI technologies and their increasing practical relevance for SMEs in recent years. The search strategy combined concepts using Boolean operators: ("artificial intelligence" OR "AI" OR "machine learning") AND ("SME" OR "small and medium" OR "Mittelstand") AND ("adoption" OR "implementation" OR "diffusion") AND ("Germany" OR "German" OR "Europe"). Additionally, relevant studies from reference lists of identified works were screened using the snowball method to capture relevant publications possibly not found through database search (Greenhalgh & Peacock, 2005).
Inclusion criteria comprised: focus on AI adoption or use in enterprises, preferably SMEs; empirical or conceptual foundation; publication language German or English; relation to technological, organizational, or environmental factors. Excluded were grey literature, magazine articles, blog posts, and studies with exclusive focus on large enterprises, to ensure scientific quality and contextual relevance of the analysis. Overall, more than 100 initial hits were identified, of which 35 papers were selected for in-depth analysis after title and abstract screening. After thorough examination for thematic fit and methodological quality, 15 core studies remained that fully met the defined criteria and form the basis for the subsequent analysis.
The identified studies were content-analyzed using both deductive and inductive coding strategies (Mayring, 2014). In a first step, central findings on barriers and success factors of AI adoption were extracted and assigned to the a priori defined TOE dimensions and DOI attributes (deductive coding). In a second step, emergent themes and patterns were identified that went beyond the predefined categories (inductive coding). The use of two independent coders for study selection and analysis enhanced the reliability of findings; inter-rater reliability according to Cohen's Kappa was .84, which can be interpreted as very good agreement (Landis & Koch, 1977). Discrepancies were resolved through discussion and consensus building.
The analyzed studies paint a consistent picture: AI adoption in SMEs is influenced by a complex bundle of interdependent factors that can be differentiated along technological, organizational, and environmental dimensions (Pan & Jang, 2008). The complexity of the phenomenon manifests in that individual factors rarely operate in isolation but exhibit multiple interactions and reinforcement effects. Figure 1 visualizes the frequency of barriers identified in the literature, organized by TOE dimension. The following sections systematically present the identified barriers and success factors, illustrated with exemplary findings from primary studies.
Note. Percentages refer to the proportion of the 15 analyzed studies in which the respective barrier was identified as relevant. Technology: Data Quality (87%), Complexity (73%), Explainability (67%); Organization: Competencies (93%), Resources (80%), Culture (53%); Environment: Regulation (60%), Infrastructure (40%), External Support (33%).
Insufficient data quality and availability are consistently identified in the literature as the primary obstacle for AI implementations in SMEs (Hansen & Bøgh, 2021; Ghobakhloo et al., 2022; Janssen, van der Voort & Wahyudi, 2017). This barrier manifests at multiple levels: adequate historical datasets required for training machine learning models are often lacking; existing data is frequently fragmented across heterogeneous legacy systems and exhibits inconsistencies, missing values, or quality deficiencies (Wang & Strong, 1996). Additionally, many SMEs lack data competency to systematically capture, prepare, and make existing data usable for AI models (Sánchez et al., 2025; Provost & Fawcett, 2013).
In the survey by Hansen and Bøgh (2021) among 257 manufacturing SMEs in Denmark, 72 percent of respondents indicated that insufficient data quality represented the greatest obstacle for implementing AI-supported predictive maintenance. Many of the examined companies had machine and sensor data, but this was often unstructured, inconsistently captured, or stored in proprietary formats that complicated integration into comprehensive analysis platforms (Lee, Kao & Yang, 2014). Similar findings are reported by Ulrich et al. (2021) for German SMEs: 58 percent of surveyed enterprises identified missing or fragmented data as the main barrier for AI projects.
SMEs face the inherent complexity of many AI tools (Thompson, Higgins & Howell, 1991). Compared to established software solutions like ERP or CRM systems, modern machine learning models, particularly deep learning approaches, often appear as opaque systems whose functioning is difficult to comprehend even for technically versed users (Füller et al., 2022; Arrieta et al., 2020). The lack of explainability and traceability of algorithmic decisions – often referred to as the black-box problem – leads to hesitation among decision-makers who express concerns regarding reliability, fairness, and legal conformity of AI-based decisions (Enholm et al., 2022; Gunning et al., 2019).
A cross-industry analysis by Baabdullah et al. (2021) notes that lacking interpretability together with concerns regarding data protection and algorithmic fairness substantially restricts practical application possibilities in SMEs. These concerns are reinforced by the regulatory context: The EU AI Act explicitly requires transparency and explainability for high-risk AI systems, which can pose considerable challenges for SMEs without specialized compliance expertise (Mittelstadt, Allo, Taddeo, Wachter & Floridi, 2016). Ghobakhloo et al. (2022) identify technical complexity in their systematic review as one of the three most common adoption barriers, mentioned in 73 percent of analyzed studies.
The perceived relative advantage of AI constitutes an essential success factor and represents the complementary side of the previously discussed barriers (Davis, 1989). SMEs demonstrate higher adoption readiness when concrete use cases with clearly communicated benefits are identified – such as automated customer services through chatbots, personalized marketing campaigns based on customer behavior data, or efficiency gains in manufacturing and logistics through predictive maintenance and demand forecasting (Chatterjee et al., 2021; Baabdullah, 2024; Mikalef & Gupta, 2021). This corresponds with the DOI construct of relative advantage: The clearer the anticipated benefit is quantifiable and related to the specific challenges of the company, the more likely a positive investment decision (Horani et al., 2023; Karahanna et al., 1999).
The empirical study by Wamba-Taguimdje et al. (2020) underscores the importance of benefit perception: In their investigation of 299 companies of various size classes, perceived business value proved to be the strongest predictor for AI adoption, with a standardized path coefficient of β = .47 (p < .001). Companies that had identified clear use cases and calculated their expected ROI showed significantly higher adoption probability than those that viewed AI as generic technology without specific application reference. Chatterjee et al. (2021) replicate this finding in the context of Indian manufacturing companies and emphasize the necessity of translating abstract technology promises into concrete, measurable business benefits.
A significant technological trend in recent years is the proliferation of low-code and no-code development platforms that make complex AI functionalities accessible through graphical user interfaces and preconfigured modules (Liwanag et al., 2025; Sahay et al., 2020). Such tools function as technology enablers by reducing development complexity and enabling employees without deep programming knowledge – so-called citizen developers – to independently create AI applications (Rymer & Koplowitz, 2019). Platforms like Microsoft Power Platform, Google AutoML, or specialized SME solutions offer prefabricated AI building blocks for common use cases such as text classification, image analysis, or forecasting models.
Liwanag et al. (2025) demonstrate in a systematic review of 156 studies that the combination of AI and low-code development increasingly enables citizen-led innovation and can partially compensate for the acute shortage of IT specialists. In 78 percent of examined studies, positive effects on the accessibility of AI technologies were reported. The democratization of AI through such platforms not only lowers technical barriers but also increases the compatibility of new solutions with existing business processes, as domain-knowledgeable specialists can design and iteratively adapt the solutions themselves (Mariani, Perez-Vega & Wirtz, 2022). Sanchis, García-Perales, Fraile and Poler (2020) identify low-code approaches as one of the most promising levers for accelerating digital transformation in SMEs.
Lacking expertise and low digital maturity are considered perhaps the most significant barriers at the organizational level (Schwäke et al., 2024; Arroyabe et al., 2024; Westerman, Bonnet & McAfee, 2014). Digital maturity describes not only the technical infrastructure of a company but also encompasses organizational capabilities such as strategic orientation, leadership qualities, and cultural openness for data-driven decision processes (Kane, Palmer, Phillips, Kiron & Buckley, 2015). Many SMEs do not have specialized data scientists or AI experts required for the conception, implementation, and continuous optimization of AI systems (Davenport & Patil, 2012). Even existing IT departments are often occupied with operational day-to-day business and possess neither capacity nor expertise for strategic AI projects.
Over 60 percent of surveyed German SMEs in the study by Ulrich et al. (2021) identify the lack of employee competencies as the primary challenge for AI initiatives. Schwäke et al. (2024) extend this perspective in their international investigation of 847 SMEs: They show that digital maturity not only influences adoption probability but also moderates the success of implemented AI projects. Companies with higher digital maturity were able to implement AI technologies faster, achieve better results, and constructively use failures for organizational learning. Pingali, Singha, Arunachalam and Pedada (2023) note in their study on digital readiness of SMEs in emerging markets that digital readiness represents a necessary prerequisite for successful AI adoption: Companies with higher digitization levels have better infrastructure, established data management practices, and a culturally conditioned openness for technological innovations (Fitzgerald, Kruschwitz, Bonnet & Welch, 2014).
SMEs typically have limited financial buffers and can mobilize fewer slack resources for experimental projects with uncertain outcomes than large enterprises (Ghobakhloo et al., 2022; Welsh & White, 1981). This structural resource scarcity has far-reaching implications for AI adoption: High initial investments in AI systems, cloud infrastructure, or specialized personnel are difficult to justify as long as immediate benefits appear uncertain and no empirical values from one's own company or comparable industry peers exist (Barney, 1991). The typical cost structures of AI projects – high upfront investments with delayed and often uncertain benefit generation – contradict the financial planning horizons of many SMEs, which are often characterized by short-term liquidity orientation.
Unlike large enterprises, SMEs can absorb failures less readily, leading to a fundamentally conservative innovation attitude (Füller et al., 2022; March, 1991). This risk aversion is reinforced in owner-managed businesses by personal loss fears of the owner, whose private assets are often closely linked to business success (Nooteboom, 1994). Arroyabe et al. (2024) note based on their investigation of 1,284 European SMEs that absorptive capacity – understood as the ability to absorb and economically exploit external knowledge – is systematically limited in SMEs when the financial environment leaves little room for experiments (Cohen & Levinthal, 1990). The authors further show that this resource constraint is moderated by the environmental context: In industries with high competitive pressure or when external funding was available, SMEs were more willing to invest in AI despite limited resources.
Internal attitude toward AI constitutes an often underestimated influencing factor that is receiving increasing attention in the literature (Schein, 2010). In numerous SMEs, a culture of preservation over change dominates, shaped by long-standing successes with established business models and processes. This cultural inertia can lead to even technologically and economically sensible innovations encountering internal resistance (Kotter, 1995). In the specific context of AI, additional concerns regarding job loss through automation play a central role: Employees fear that their activities could be substituted by intelligent algorithms, which can lead to active or passive resistance against AI projects (Knoblach, Vogl, Wehnert, Carbon, Franz & Schmid, 2025; Autor, 2015).
These concerns are not unfounded, as AI technologies do indeed change work requirements and job profiles. However, empirical research paints a more nuanced picture: While certain routine activities are substituted through automation, new tasks simultaneously emerge in the monitoring, interpretation, and optimization of AI systems (Brynjolfsson & McAfee, 2017). Change management thus plays a critical role: Without early involvement of the workforce in planning and implementation processes and without transparent communication about opportunities and limitations of AI, acceptance problems threaten to derail even technically successful implementations (Venkatesh & Bala, 2008). Knoblach et al. (2025) report from their evaluation of a Bavarian AI demonstration center that initial prejudices against AI could only be reduced through practical experience, training, and open dialogue with AI experts. The authors emphasize the importance of hands-on experiences that enable employees to test AI systems in a protected environment and explore their possibilities and limitations themselves.
Among the most significant success factors at the organizational level is active backing from management (Horani et al., 2023; Schwäke et al., 2024; Liang, Saraf, Hu & Xue, 2007). Top management support describes more than passive consent to AI initiatives; it encompasses active prioritization of AI projects, provision of necessary resources, communication of a clear vision, and personal involvement of the leadership level in strategic decisions (Hambrick & Mason, 1984). When owners or managing directors are themselves convinced of AI and actively act as innovation champions, success probabilities increase considerably, as resistance at middle management levels can be more easily overcome and necessary resource reallocations can be legitimized.
Horani et al. (2023) empirically demonstrate that top management support exerts a significantly positive influence on adoption readiness, with a standardized path coefficient of β = .38 (p < .001). This support manifests in concrete actions: releasing budgets for pilot projects, forming interdisciplinary AI teams, initiating external collaborations with research institutions or technology providers, and establishing metrics systems for success measurement (Damanpour & Schneider, 2006). Closely linked to top management support is the presence of a clear AI strategy as part of long-term business orientation (Baabdullah, 2024). Companies that have anchored AI as a strategic topic not only show higher adoption rates but also achieve better implementation results (Ransbotham, Khodabandeh, Fehling, LaFountain & Kiron, 2019).
The external macro-environment significantly influences SMEs, as these typically have fewer resources for environmental monitoring and proactive adaptation than large enterprises (DiMaggio & Powell, 1983). Currently, companies face dynamic technological development alongside uncertain, evolving regulations (Jobin, Ienca & Vayena, 2019). Many German SMEs are closely watching the design and implementation of the EU AI Act, as certain AI applications – particularly those classified as high-risk – will be subject to extensive documentation, testing, and certification requirements in the future (Sánchez et al., 2025; Veale & Borgesius, 2021).
These legal uncertainties act as a brake on investment decisions, as companies may hold back expenditures until binding framework conditions exist and legal certainty is guaranteed (Blind, 2012). The uncertainty relates not only to the regulatory framework itself but also to its enforcement and the specific compliance requirements for different use cases. Horani et al. (2023) note based on their empirical investigation that government regulation can have a negative influence on adoption intentions (β = -.19, p < .05), as strict requirements are perceived as additional cost and compliance factors that reduce the expected net benefit of AI investments. On the other hand, clear regulation can also create legal certainty and strengthen trust in AI technologies – an effect that should be investigated in future studies after full implementation of the EU AI Act.
Competitive pressure functions as an external factor that significantly promotes AI adoption and changes the calculus of potential adopters (Porter, 1985). When competitors successfully deploy AI and thereby gain competitive advantages – whether through cost reduction, improved product quality, or innovative service offerings – pressure arises on other market participants to follow suit to avoid falling behind (Arroyabe et al., 2024; Iacovou, Benbasat & Dexter, 1995). This mimetic isomorphism, where companies imitate the behavior of successful competitors, can motivate even resource-constrained SMEs to adopt AI when the alternative is perceived as an existential competitive disadvantage.
Horani et al. (2023) empirically demonstrate that perceived competitive pressure has a positive effect on adoption decisions (β = .31, p < .001). Customer pressure can operate similarly: When major clients demand data-driven analyses, standardized digital interfaces, or AI-supported quality certifications, SMEs must respond to avoid jeopardizing business relationships and losing supplier qualifications (Teo, Wei & Benbasat, 2003). In automotive industry supply chains, manufacturers increasingly expect their suppliers to use AI, for example for automated quality inspection using image recognition or predictive maintenance to avoid delivery failures (Dolgui & Ivanov, 2022). This form of institutional pressure can move even SMEs with low intrinsic innovation propensity toward AI adoption when the alternative is loss of important customer relationships.
The availability of external support offerings constitutes a significant enabler, particularly relevant for resource-constrained SMEs (Chau & Tam, 1997). External support can take various forms: government funding programs that reduce investment risks through subsidies or favorable loans; technology providers and consulting services offering SME-tailored AI solutions; industry associations and competence centers enabling knowledge transfer and networking; as well as collaborations with universities or research institutes (Knoblach et al., 2025; Chesbrough, 2003). These external resources can compensate for internal competency deficits and enable access to expertise that SMEs could not build on their own.
Horani et al. (2023) show in their empirical investigation that vendor support and resource availability positively correlate with adoption propensity (β = .24, p < .01). Collaborations with universities or research institutes enable knowledge transfer and increase trust and competence on the company side, particularly when these collaborations are application-oriented and address concrete business problems (Perkmann & Walsh, 2007). The Bavarian AI demonstration center described by Knoblach et al. (2025) exemplifies how university and enterprises can jointly work through use cases and thereby reduce adoption barriers: Through hands-on workshops, individual consulting, and provision of test environments, participating SMEs could gain practical experience and initiate concrete implementation projects (Etzkowitz & Leydesdorff, 2000). The authors report that 65 percent of participating companies started first AI projects within 12 months after the workshop.
The three dimensions of the TOE framework mutually influence each other and generate complex interdependencies that make isolated consideration of individual factors appear insufficient (Markus & Robey, 1988). Strong competitive intensity in the environment can, for example, internally function as a top management priority, which in turn frees resources for addressing technical challenges and overcoming organizational resistance. Conversely, a deficiency in one dimension – such as lacking data infrastructure at the technological level – frequently leads to progress in other dimensions being thwarted, even when top management support is present and external funding is available (Sánchez et al., 2025).
These interdependencies illustrate that AI adoption in SMEs must be understood as a socio-technical system in which technology, people, and environment interact in complex ways (Bostrom & Heinen, 1977). Successful implementations ideally address all dimensions simultaneously: They select suitable, accessible tools compatible with the company's technical infrastructure and data holdings; they train and motivate employees through transparent communication and participation; and they strategically utilize market opportunities and external support structures to compensate for internal resource constraints (Orlikowski, 1992). Analysis of the 15 core studies shows that companies with a holistic adoption approach – which addresses technological, organizational, and contextual factors in an integrated manner – exhibit significantly higher success rates than those that view AI adoption primarily as a technical project.
Application of the integrated TOE-DOI framework proved helpful for structured analysis of adoption factors and confirms findings from earlier studies on technology adoption in organizational contexts (DePietro, Wiarda & Fleischer, 1990). In agreement with Chatterjee et al. (2021), the combination of both theories provides a more comprehensive picture than isolated application of individual approaches. While the TOE framework captures contextual conditions under which adoption decisions are made, DOI theory supplements the subjective evaluation of innovation by decision-makers. This dual perspective appears particularly suitable for the SME context, as here – unlike in large enterprises with specialized decision-making bodies – individuals frequently make strategic technology decisions whose individual perceptions and risk preferences exert considerable influence.
The results confirm that technological factors such as complexity and compatibility as well as organizational factors such as management support and resource endowment are equally critical (Zhu et al., 2006). DOI features such as relative advantage and complexity appear implicitly or explicitly in nearly all analyzed studies: Companies consistently evaluate the expected benefit and effort of an innovation and weigh these against alternatives (Fichman, 2004). The present analysis extends theoretical discourse with the factor of technological democratization. The advent of low-code and no-code AI tools is a relatively new phenomenon not yet explicitly represented in classical adoption models (Sanchis et al., 2020). Results suggest that such platforms could act as a moderator that reduces technical complexity and thereby removes a central adoption barrier (Liwanag et al., 2025). This insight suggests that established adoption models should be supplemented with the concept of technological democratization.
A common thread in the analyzed studies is the recognition that SMEs operate under different conditions than large enterprises and therefore require specific analytical perspectives (Schwäke et al., 2024; Arroyabe et al., 2024; Rothwell & Dodgson, 1994). While corporations benefit from extensive data pools, dedicated AI departments, and internal expertise, while simultaneously struggling with bureaucratic structures, silo mentalities, and lengthy decision processes, SMEs are characterized by higher agility, shorter decision paths, and closer customer relationships, but suffer from resource scarcity and limited specialization (Acs & Audretsch, 1990). The personality and vision of top management is often more decisive in SMEs than in large enterprises, as a single owner can exert considerable influence on innovation orientation and is not constrained by committee decisions or shareholder interests (Lubatkin, Simsek, Ling & Veiga, 2006).
These findings imply that theories and models of technology adoption should not be transferred unmodified from large enterprises to SMEs (Curran & Blackburn, 2001). The specific characteristics of SMEs – limited slack resources, person-dependent decision processes, informal organizational structures, and high flexibility – require adapted theoretical perspectives and practical intervention strategies. At the same time, some of these characteristics also open opportunities: The lower organizational complexity can enable faster implementation cycles, and the proximity of management to operational processes facilitates identification of relevant use cases and overcoming internal resistance.
Several limitations should be considered when interpreting the results. First, most evaluated studies are based on survey data or conceptual analyses (Podsakoff, MacKenzie, Lee & Podsakoff, 2003). Long-term observations that trace the actual success or failure of AI projects in SMEs and identify causal mechanisms are available only to a limited extent. Conclusions are thus primarily based on perceived barriers and intended adoption, less on ex-post success evaluations (Sheppard, Hartwick & Warshaw, 1988). This limitation is characteristic of a young research field and should be addressed through future longitudinal studies.
Second, the analysis focuses on Germany or comparable European industrial countries; for other cultural and institutional contexts, individual factors may possess different relevance (Pingali et al., 2023; Hofstede, 2001). The strong expression of the Mittelstand concept in Germany, the specific support structures, and cultural values such as risk aversion and long-term orientation limit generalizability to other national contexts. Third, the dynamics of the research subject should be considered: The rapid advancement of AI technologies, particularly in the area of generative AI, as well as evolving regulatory frameworks mean that barriers identified today may already lose relevance tomorrow, while new challenges emerge.
The findings imply that successful AI implementation is not merely a technical project but requires holistic organizational transformation (Westerman et al., 2014). Management should assume an active role, develop visions for AI applications, and establish realistic implementation plans that address technological, organizational, and contextual factors in an integrated manner (Fountaine, McCarthy & Saleh, 2019). Results consistently show that top management support is one of the strongest predictors of adoption success – passive delegation to IT departments appears as an inadequate strategy.
Essential is investment in employee qualification to reduce apprehension, build internal competencies, and address cultural resistance (Agrawal, Gans & Goldfarb, 2018). AI introduction should be designed as a gradual process, beginning with pilot projects that deliver quick, visible successes and enable organizational learning (Ries, 2011). The use of low-code platforms enables prototype development with minimal resource commitment and limited risk, while involving specialist departments in solution development (Liwanag et al., 2025). External partnerships with technology providers, consultants, or research institutions can close critical competency gaps and enable access to expertise that cannot be built internally (Chesbrough, 2003). Finally, SMEs should actively utilize available funding programs and support structures to reduce investment risks and benefit from others' experiences.
Results show that political and infrastructural framework conditions exert significant influence on AI adoption in SMEs (Mazzucato, 2018). To anchor AI in the Mittelstand, economic policymakers should address several action areas: First, expansion of digital infrastructure, particularly in rural regions where many medium-sized companies are located (Prieger, 2013). Fast internet connections and access to cloud infrastructure are necessary prerequisites for many AI applications. Second, provision of accessible funding programs for AI investments in SMEs that cover not only hardware and software but also consulting services and qualification measures.
Third, support for regional AI competence centers that provide demonstrators, offer training, and act as intermediaries between SMEs and research (Knoblach et al., 2025; Nauwelaers & Wintjes, 2002). Such centers can assume an important bridging function by translating abstract technology knowledge into practically applicable solutions and enabling SMEs a low-risk entry into AI technologies. Fourth, development of clear regulatory guidelines and quality standards that provide orientation and strengthen trust in AI technologies without creating disproportionate compliance burdens for SMEs (Floridi et al., 2018). Fifth, promotion of education initiatives at all levels – from vocational training through university curricula to continuing education offerings – to increase the supply of skilled workers in the AI field (Autor, 2015).
This systematic literature review has examined AI implementation in German SMEs through the lens of the integrated TOE-DOI framework. The past five years have produced considerable research activity that consistently demonstrates: SMEs face significant hurdles in AI adoption, yet recognizable levers exist for overcoming these barriers (Dwivedi et al., 2021). The systematic analysis of 15 core studies enables consolidation of a previously fragmented state of research and derivation of differentiated insights for science and practice.
As central barriers were identified: the shortage of high-quality data and AI specialists (mentioned in 93% and 87% of studies respectively), technical and organizational complexity of AI systems, limited financial resources for experimental projects, and an uncertain regulatory environment. These barriers do not operate in isolation but mutually reinforce each other in complex interdependencies: Lacking data infrastructure complicates demonstration of concrete benefit potentials, which in turn inhibits investment decisions and delays building of internal competencies.
As essential success factors emerged: engaged leadership with clear digital vision and active support of AI initiatives, clear benefit communication through identification of concrete, quantifiable use cases, deployment of easily manageable AI tools like low-code platforms, and integration of external support structures (Pumplun, Tauchert & Heidt, 2019). Particularly noteworthy is the potential of low-code and no-code platforms, which lower technical entry barriers and enable even resource-constrained enterprises to develop AI applications. These democratizing technologies could initiate a paradigm shift by shifting AI development from specialized experts to domain-knowledgeable specialist users.
For future research, several promising directions emerge: (1) empirical studies on the effectiveness of low-code platforms for AI adoption in SMEs that measure actual implementation successes beyond adoption intentions; (2) longitudinal studies that track the long-term success of AI implementations and trace dynamic development paths; (3) industry-specific analyses to identify context-dependent adoption factors in different Mittelstand sectors; (4) comparative investigations of the effectiveness of different funding instruments and policy interventions; and (5) research on the role of generative AI technologies that fundamentally expand the application spectrum of AI and could create new adoption dynamics (Makridakis, 2017).
AI adoption in SMEs remains a complex transformation process affecting technology, people, and structures equally (Bharadwaj, El Sawy, Pavlou & Venkatraman, 2013). Companies that invest in qualification today, build data infrastructures, and initiate pilot projects create the foundation for sustainable competitive advantages in an increasingly AI-shaped economy. The present analysis serves as an orientation framework that systematically identifies critical obstacles and promising solution approaches and offers action-relevant insights for both business practice and economic policymakers.
Acs, Z. J., & Audretsch, D. B. 1990. Innovation and small firms. Cambridge, MA: MIT Press.
Agarwal, R., & Prasad, J. 1998. A conceptual and operational definition of personal innovativeness in the domain of information technology. Information Systems Research, 9(2): 204–215.
Agrawal, A., Gans, J., & Goldfarb, A. 2018. Prediction machines: The simple economics of artificial intelligence. Boston: Harvard Business Review Press.
Ajzen, I. 1991. The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2): 179–211.
Alsheibani, S., Cheung, Y., & Messom, C. 2018. Artificial intelligence adoption: AI-readiness at firm-level. PACIS 2018 Proceedings, 37.
Arrieta, A. B., et al. 2020. Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges. Information Fusion, 58: 82–115.
Arroyabe, M. F., Arranz, N., de Arroyabe, J. C. F., & de Arroyabe, F. I. 2024. Analyzing AI adoption in European SMEs. Technology in Society, 79: 102733.
Autor, D. H. 2015. Why are there still so many jobs? Journal of Economic Perspectives, 29(3): 3–30.
Awa, H. O., Ukoha, O., & Emecheta, B. C. 2016. Using T-O-E theoretical framework to study the adoption of ERP solution. Cogent Business & Management, 3(1): 1196571.
Baabdullah, A. M. 2024. The precursors of AI adoption in business. International Journal of Information Management, 75: 102745.
Baabdullah, A. M., Alalwan, A. A., Slade, E. L., Raman, R., & Khatatneh, K. F. 2021. SMEs and artificial intelligence (AI). Industrial Marketing Management, 98: 255–270.
Baker, J. 2012. The technology–organization–environment framework. In Y. K. Dwivedi et al. (Eds.), Information systems theory: 231–245. New York: Springer.
Barney, J. 1991. Firm resources and sustained competitive advantage. Journal of Management, 17(1): 99–120.
Bharadwaj, A., El Sawy, O. A., Pavlou, P. A., & Venkatraman, N. 2013. Digital business strategy. MIS Quarterly, 37(2): 471–482.
Bitkom. 2024. IT-Fachkräftemangel in Deutschland. Berlin: Bitkom Research.
Bley, K., Leyh, C., & Schäffer, T. 2016. Digitization of German enterprises in the production sector. AMCIS 2016 Proceedings.
Blind, K. 2012. The influence of regulations on innovation. Research Policy, 41(2): 391–400.
Bostrom, R. P., & Heinen, J. S. 1977. MIS problems and failures: A socio-technical perspective. MIS Quarterly, 1(3): 17–32.
Brynjolfsson, E., & McAfee, A. 2017. Machine, platform, crowd. New York: W. W. Norton.
Bughin, J., et al. 2018. Notes from the AI frontier: Modeling the impact of AI on the world economy. McKinsey Global Institute.
Chalmers, D., MacKenzie, N. G., Carter, S., & Quail, S. 2020. Artificial intelligence and entrepreneurship. Entrepreneurship Theory and Practice, 45(5): 1022–1053.
Chatterjee, S., Rana, N. P., Dwivedi, Y. K., & Baabdullah, A. M. 2021. Understanding AI adoption in manufacturing using TAM-TOE. Technological Forecasting and Social Change, 170: 120880.
Chau, P. Y. K., & Tam, K. Y. 1997. Factors affecting the adoption of open systems. MIS Quarterly, 21(1): 1–24.
Chesbrough, H. W. 2003. Open innovation. Boston: Harvard Business School Press.
Cohen, W. M., & Levinthal, D. A. 1990. Absorptive capacity. Administrative Science Quarterly, 35(1): 128–152.
Curran, J., & Blackburn, R. A. 2001. Researching the small enterprise. London: Sage.
Damanpour, F., & Schneider, M. 2006. Phases of the adoption of innovation in organizations. British Journal of Management, 17(3): 215–236.
Davenport, T. H., & Patil, D. J. 2012. Data scientist: The sexiest job of the 21st century. Harvard Business Review, 90(10): 70–76.
Davenport, T. H., & Ronanki, R. 2018. Artificial intelligence for the real world. Harvard Business Review, 96(1): 108–116.
Davis, F. D. 1989. Perceived usefulness, perceived ease of use, and user acceptance of IT. MIS Quarterly, 13(3): 319–340.
DePietro, R., Wiarda, E., & Fleischer, M. 1990. The context for change. In L. G. Tornatzky & M. Fleischer (Eds.), The processes of technological innovation: 151–175.
Destatis. 2023. Unternehmensstatistik Deutschland. Wiesbaden: Statistisches Bundesamt.
DiMaggio, P. J., & Powell, W. W. 1983. The iron cage revisited. American Sociological Review, 48(2): 147–160.
Dolgui, A., & Ivanov, D. 2022. 5G in digital supply chain. International Journal of Production Research, 60(2): 442–451.
Dwivedi, Y. K., et al. 2021. Artificial Intelligence (AI): Multidisciplinary perspectives. International Journal of Information Management, 57: 101994.
Dwivedi, Y. K., et al. 2023. Opinion Paper: "So what if ChatGPT wrote it?" International Journal of Information Management, 71: 102642.
Enholm, I. M., Papagiannidis, E., Mikalef, P., & Krogstie, J. 2022. AI and business value. Information Systems Frontiers, 24(6): 1709–1734.
Etzkowitz, H., & Leydesdorff, L. 2000. The dynamics of innovation. Research Policy, 29(2): 109–123.
European Commission. 2003. Commission Recommendation 2003/361/EC. Official Journal of the European Union, L 124: 36–41.
Fichman, R. G. 2004. Going beyond the dominant paradigm for IT innovation research. JAIS, 5(8): 314–355.
Fitzgerald, M., Kruschwitz, N., Bonnet, D., & Welch, M. 2014. Embracing digital technology. MIT Sloan Management Review, 55(2): 1–12.
Floridi, L., et al. 2018. AI4People—An ethical framework for a good AI society. Minds and Machines, 28(4): 689–707.
Fountaine, T., McCarthy, B., & Saleh, T. 2019. Building the AI-powered organization. Harvard Business Review, 97(4): 62–73.
Frambach, R. T., & Schillewaert, N. 2002. Organizational innovation adoption. Journal of Business Research, 55(2): 163–176.
Füller, J., Hutter, K., Wahl, J., Bilgram, V., & Tekic, Z. 2022. How AI revolutionizes innovation management. Technological Forecasting and Social Change, 178: 121598.
Gangwar, H., Date, H., & Ramaswamy, R. 2015. Understanding cloud computing adoption using TAM-TOE. Journal of Enterprise Information Management, 28(1): 107–130.
Gartner. 2019. Gartner predicts 85% of AI projects will deliver erroneous outcomes. Stamford, CT: Gartner Inc.
Ghobakhloo, M., Iranmanesh, M., Vilkas, M., Grybauskas, A., & Amran, A. 2022. Drivers and barriers of Industry 4.0 adoption. Journal of Manufacturing Technology Management, 33(6): 1029–1058.
Greenhalgh, T., & Peacock, R. 2005. Effectiveness of search methods in systematic reviews. BMJ, 331(7524): 1064–1065.
Gunning, D., et al. 2019. XAI—Explainable artificial intelligence. Science Robotics, 4(37): eaay7120.
Hambrick, D. C., & Mason, P. A. 1984. Upper echelons. Academy of Management Review, 9(2): 193–206.
Hansen, E. B., & Bøgh, S. 2021. AI and IoT in small and medium-sized enterprises. Journal of Manufacturing Systems, 58: 362–372.
Hofstede, G. 2001. Culture's consequences (2nd ed.). Thousand Oaks, CA: Sage.
Horani, O. M., et al. 2023. Critical determinants impacting AI adoption. Information Development, Online First.
Hsu, P. F., Kraemer, K. L., & Dunkle, D. 2006. Determinants of e-business use. International Journal of Electronic Commerce, 10(4): 9–45.
Iacovou, C. L., Benbasat, I., & Dexter, A. S. 1995. EDI and small organizations. MIS Quarterly, 19(4): 465–485.
Janssen, M., van der Voort, H., & Wahyudi, A. 2017. Factors influencing big data decision-making quality. Journal of Business Research, 70: 338–345.
Jeyaraj, A., Rottman, J. W., & Lacity, M. C. 2006. IT innovation adoption research. Journal of Information Technology, 21(1): 1–23.
Jobin, A., Ienca, M., & Vayena, E. 2019. The global landscape of AI ethics guidelines. Nature Machine Intelligence, 1(9): 389–399.
Kane, G. C., Palmer, D., Phillips, A. N., Kiron, D., & Buckley, N. 2015. Strategy, not technology, drives digital transformation. MIT Sloan Management Review, 56(4): 1–25.
Karahanna, E., Straub, D. W., & Chervany, N. L. 1999. IT adoption across time. MIS Quarterly, 23(2): 183–213.
Kitchenham, B. 2004. Procedures for performing systematic reviews. Keele University Technical Report.
Knoblach, J., Vogl, L., Wehnert, C., Carbon, C.-C., Franz, J., & Schmid, U. 2025. The SME AI Experience Center. KI – Künstliche Intelligenz, Online First.
Kotter, J. P. 1995. Leading change. Harvard Business Review, 73(2): 59–67.
Landis, J. R., & Koch, G. G. 1977. The measurement of observer agreement. Biometrics, 33(1): 159–174.
Lee, J., Kao, H. A., & Yang, S. 2014. Service innovation and smart analytics for Industry 4.0. Procedia CIRP, 16: 3–8.
Levy, Y., & Ellis, T. J. 2006. A systems approach to conduct literature review. Informing Science, 9: 181–212.
Liang, H., Saraf, N., Hu, Q., & Xue, Y. 2007. Assimilation of enterprise systems. MIS Quarterly, 31(1): 59–87.
Liwanag, G. L. L., Ebardo, R. A., & Cheng, D. C. 2025. Low-Code and No-Code Development in the Era of AI. Data & Metadata, 4(4): 1218.
Lubatkin, M. H., Simsek, Z., Ling, Y., & Veiga, J. F. 2006. Ambidexterity and performance in SMEs. Journal of Management, 32(5): 646–672.
Makridakis, S. 2017. The forthcoming AI revolution. Futures, 90: 46–60.
March, J. G. 1991. Exploration and exploitation in organizational learning. Organization Science, 2(1): 71–87.
Mariani, M. M., Perez-Vega, R., & Wirtz, J. 2022. AI in marketing. Psychology & Marketing, 39(4): 755–776.
Markus, M. L., & Robey, D. 1988. IT and organizational change. Management Science, 34(5): 583–598.
Mayring, P. 2014. Qualitative content analysis. Klagenfurt: SSOAR.
Mazzucato, M. 2018. The entrepreneurial state. London: Penguin.
Mikalef, P., & Gupta, M. 2021. Artificial intelligence capability. Information & Management, 58(3): 103434.
Mittelstadt, B. D., Allo, P., Taddeo, M., Wachter, S., & Floridi, L. 2016. The ethics of algorithms. Big Data & Society, 3(2): 1–21.
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. 2009. PRISMA statement. PLoS Medicine, 6(7): e1000097.
Moore, G. C., & Benbasat, I. 1991. Development of an instrument to measure IT adoption perceptions. Information Systems Research, 2(3): 192–222.
Müller, J. M., Buliga, O., & Voigt, K. I. 2021. The role of absorptive capacity and innovation strategy in Industry 4.0. European Management Journal, 39(3): 333–343.
Nauwelaers, C., & Wintjes, R. 2002. Innovating SMEs and regions. Technology Analysis & Strategic Management, 14(2): 201–215.
Nooteboom, B. 1994. Innovation and diffusion in small firms. Small Business Economics, 6(5): 327–347.
Oliveira, T., & Martins, M. F. 2011. Literature review of IT adoption models at firm level. EJISE, 14(1): 110–121.
Orlikowski, W. J. 1992. The duality of technology. Organization Science, 3(3): 398–427.
Pan, M. J., & Jang, W. Y. 2008. Determinants of ERP adoption within TOE. JCIS, 48(3): 94–102.
Perkmann, M., & Walsh, K. 2007. University–industry relationships. International Journal of Management Reviews, 9(4): 259–280.
Pingali, S. R., Singha, S., Arunachalam, S., & Pedada, K. 2023. Digital readiness of SMEs in emerging markets. Journal of Business Research, 164: 113973.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. 2003. Common method biases. Journal of Applied Psychology, 88(5): 879–903.
Porter, M. E. 1985. Competitive advantage. New York: Free Press.
Premkumar, G., & Roberts, M. 1999. Adoption of new IT in rural small businesses. Omega, 27(4): 467–484.
Prieger, J. E. 2013. The broadband digital divide. Telecommunications Policy, 37(6–7): 483–502.
Priyono, A., Moin, A., & Putri, V. N. A. O. 2020. Identifying digital transformation paths in the business model of SMEs during the COVID-19 pandemic. Journal of Open Innovation, 6(4): 104.
Provost, F., & Fawcett, T. 2013. Data science and data-driven decision making. Big Data, 1(1): 51–59.
Pumplun, L., Tauchert, C., & Heidt, M. 2019. A new organizational chassis for AI. ECIS 2019 Proceedings.
Ramdani, B., Kawalek, P., & Lorenzo, O. 2009. Predicting SMEs' adoption of enterprise systems. JEIM, 22(1/2): 10–24.
Ransbotham, S., Khodabandeh, S., Fehling, R., LaFountain, B., & Kiron, D. 2019. Winning with AI. MIT Sloan Management Review, 61(1): 1–17.
Ransbotham, S., Kiron, D., Gerbert, P., & Reeves, M. 2017. Reshaping business with AI. MIT Sloan Management Review, 59(1): 1–17.
Ries, E. 2011. The lean startup. New York: Crown Business.
Rogers, E. M. 2003. Diffusion of innovations (5th ed.). New York: Free Press.
Rothwell, R., & Dodgson, M. 1994. Innovation and size of firm. In M. Dodgson & R. Rothwell (Eds.), Handbook of industrial innovation: 310–324.
Russell, S., & Norvig, P. 2021. Artificial intelligence: A modern approach (4th ed.). Upper Saddle River, NJ: Pearson.
Rymer, J. R., & Koplowitz, R. 2019. The Forrester Wave™: Low-code development platforms. Forrester Research.
Sahay, A., Indamutsa, A., Di Ruscio, D., & Pierantonio, A. 2020. Supporting low-code development platforms. SEAA 2020 Proceedings, 171–178.
Sánchez, E., Calderón, R., & Herrera, F. 2025. AI Adoption in SMEs: TOE–DOI Framework. Applied Sciences, 15(12): 6465.
Sanchis, R., García-Perales, Ó., Fraile, F., & Poler, R. 2020. Low-code as enabler of digital transformation. Applied Sciences, 10(1): 12.
Schein, E. H. 2010. Organizational culture and leadership (4th ed.). San Francisco: Jossey-Bass.
Schwäke, J., Peters, A., Kanbach, D. K., Kraus, S., & Jones, P. 2024. The new normal: AI adoption in SMEs. Journal of Small Business Management, 63(1): 1–35.
Sheppard, B. H., Hartwick, J., & Warshaw, P. R. 1988. The theory of reasoned action. Journal of Consumer Research, 15(3): 325–343.
Soto-Acosta, P. 2020. COVID-19 pandemic: Shifting digital transformation to a high-speed gear. Information Systems Management, 37(4): 260–266.
Teo, H. H., Wei, K. K., & Benbasat, I. 2003. Predicting intention to adopt interorganizational linkages. MIS Quarterly, 27(1): 19–49.
Thompson, R. L., Higgins, C. A., & Howell, J. M. 1991. Personal computing. MIS Quarterly, 15(1): 125–143.
Thong, J. Y. 1999. An integrated model of IS adoption in small businesses. JMIS, 15(4): 187–214.
Tornatzky, L. G., & Fleischer, M. 1990. The processes of technological innovation. Lexington, MA: Lexington Books.
Tornatzky, L. G., & Klein, K. J. 1982. Innovation characteristics and innovation adoption-implementation. IEEE Transactions on Engineering Management, 29(1): 28–45.
Tranfield, D., Denyer, D., & Smart, P. 2003. Towards a methodology for systematic review. British Journal of Management, 14(3): 207–222.
Ulrich, P., Frank, V., & Kratt, M. 2021. Adoption of AI technologies in German SMEs. In S. Hundal et al. (Eds.), Corporate governance: 76–84.
Veale, M., & Borgesius, F. Z. 2021. Demystifying the Draft EU AI Act. Computer Law Review International, 22(4): 97–112.
Venkatesh, V., & Bala, H. 2008. Technology acceptance model 3. Decision Sciences, 39(2): 273–315.
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. 2003. User acceptance of IT: Toward a unified view. MIS Quarterly, 27(3): 425–478.
Vial, G. 2019. Understanding digital transformation. Journal of Strategic Information Systems, 28(2): 118–144.
Wamba-Taguimdje, S. L., Wamba, S. F., Fosso Wamba, J. R. K., & Tchatchouang Nguegang, E. T. 2020. Influence of AI on firm performance. Business Process Management Journal, 26(7): 1893–1924.
Wang, R. Y., & Strong, D. M. 1996. Beyond accuracy: What data quality means. JMIS, 12(4): 5–33.
Webster, J., & Watson, R. T. 2002. Analyzing the past to prepare for the future. MIS Quarterly, 26(2): xiii–xxiii.
Welsh, J. A., & White, J. F. 1981. A small business is not a little big business. Harvard Business Review, 59(4): 18–27.
Westerman, G., Bonnet, D., & McAfee, A. 2014. Leading digital. Boston: Harvard Business Review Press.
Zhu, K., & Kraemer, K. L. 2005. Post-adoption variations in e-business. Information Systems Research, 16(1): 61–84.
Zhu, K., Kraemer, K. L., & Xu, S. 2006. The process of innovation assimilation. Management Science, 52(10): 1557–1576.
Dr. Konstantin Krebs · Dr. Justus & Partners · Jürgensallee 53 · 22609 Hamburg · +49 40 285 390 21
JP