This phase presents the mapping of the main activities and the structure of the systematic literature review.

A preliminary review of the literature reveals the existence of models and frameworks for digital competencies in artificial intelligence and entrepreneurship within both small and large enterprises, aiming to characterize the emerging challenges faced by new entrepreneurs. Studies deemed to have limited relevance to the objectives of the present research were excluded based on established exclusion criteria.

The central question guiding this study was: What are the main challenges entrepreneurs face when integrating artificial intelligence technologies into their businesses, and how do implementation strategies, staff training, and institutional support influence the ability to overcome these obstacles in order to foster successful entrepreneurship in the era of artificial intelligence?

As entrepreneurs attempt to incorporate new technologies into their ventures, they may face significant hurdles. These include, first and foremost, organizational resistance to change, which can lead to dissatisfaction or discomfort, strict regulatory and validation requirements, and the challenge of implementing ethical AI principles—often hindered by existing policies and organizational practices ( Ali et al., 2023; Sharma, 2023).

The systematic literature review aims to address the following research questions in alignment with the study’s objectives. •

RQ1. What is the impact of organizational culture and entrepreneurial mindset on the adoption and effective use of artificial intelligence technologies by entrepreneurs?

The systematic review defines its scope using the CIMO framework (Context, Intervention, Mechanisms, and Outcomes), as outlined in Table 1, based on the model proposed by Kitchenham et al. (2009).

Based on the CIMO framework and the research questions, exploratory searches were conducted to evaluate the keywords and identify their relevance to the field of study. The ERIC and UNESCO thesauri were used to define the terms or synonyms associated with the keywords used in the information search ( Table 2).

The review protocol, outlined in Figure 1, consists of three phases: (a) inclusion and exclusion criteria, (b) search strategies, and (c) search string.

To select the most relevant studies that contribute to answering the research questions, six inclusion criteria (IC) were established:

Exclusion criteria apply to all documents that do not meet the inclusion standards described above.

The selection of information was based on open-access databases available through the virtual library of Universidad Técnica del Norte: Scopus, Web of Science (WoS), Taylor and Francis, and Scielo. The assessment of quality and relevance within the field of study was guided by the use of specific search strings.

Using the CIMO question framework and considering the discipline of business and management, the keywords, inclusion and exclusion criteria were structured into tailored search strings for each database. These keywords were combined using Boolean operators AND and OR, allowing for multiple query combinations. Quotation marks (“”) were used to include exact term pairs, and the asterisk (*) was applied to capture both singular and plural forms.

Additional inclusion criteria were applied: publication dates between 2020 and 2024; languages included English, Spanish, and Portuguese (either written or translated in draft or final article versions); open-access availability; and document type. The customized search strings for each database are presented in Table 3.

Un A protocol “is an essential component of the systematic review process; it ensures that the review is carefully planned and that this planning is explicitly documented before the review begins” ( Moher et al., 2015, p. 1; as cited in Basantes-Andrade et al., 2022).

The systematic review was developed based on the PRISMA 2020 statement. To ensure methodological rigor, the review protocol was assessed in accordance with PRISMA 2020 guidelines, even though the study does not include a synthesis component. This assessment was intended to uphold the review’s credibility and transparency. The protocol was validated by experts in entrepreneurship and artificial intelligence technologies. “The PRISMA 2020 statement is intended for use in systematic reviews that include synthesis (e.g., pairwise meta-analyses or other methods of synthesis)” ( Page et al., 2021), thereby reinforcing the review’s validity and ethical standards.

During the study selection process, the systematic review was conducted based on primary studies, including an assessment of study quality, as well as the extraction and synthesis of data.

To identify and organize key bibliographic sources, the data were compiled using Microsoft Excel 365. Data from the Web of Science (WoS) and Scielo databases were extracted using the Win tab tool, which downloads records in a text file format ( savedrecs.txt). Files from Taylor and Francis and Scopus were downloaded in comma-separated values (.csv) format.

The selection of primary studies was guided by the flow diagram proposed by Page et al. (2021), shown in Figure 2, which outlines four stages: identification, screening, eligibility, and inclusion.

A total of 721 documents were retrieved: 121 indexed in Scopus, 123 in Web of Science (WoS), 123 in Taylor and Francis, and 429 in Scielo. During the screening phase, duplicate records were removed (n = 68) using Excel’s “Remove Duplicates” function. In the eligibility phase, the titles, abstracts, and keywords of each document were reviewed, applying the predefined inclusion criteria and search strings (n = 108).

Subsequently, the documents were evaluated according to the inclusion and exclusion criteria. A total of 82 studies were excluded for one or more of the following reasons: they did not address the research questions, their context differed from that of entrepreneurship, or the full text was not accessible. In the final phase, 26 articles were included for in-depth reading and detailed analysis. It is declared that there are no conflicts of interest.

The main criteria for the overall assessment of study quality were analyzed comprehensively and are presented through a procedure involving inference by both academic experts and professionals from the business sector. This process employed a quantitative quality checklist, detailed in Table 4. Since the studies had already been evaluated using a qualitative checklist, the quantitative checklist was developed based on the questions proposed by ( Kitchenham et al., 2009).

The checklist includes ten questions, each assigned a value of 1 point and evaluated using a three-option Likert scale: yes, no, and partial, corresponding to scores of 1, 0, and 0.5, respectively. A cutoff score of 7 was established; thus, all studies scoring 7 or higher were included, while those failing to meet this threshold were excluded due to insufficient reliable evidence. Table 5 presents a summary of the 26 selected studies and their corresponding scores, organized by reference to enhance clarity in the presentation of results.

The selected studies, derived from the extraction of primary data, were evaluated and organized in two stages: 1) General metadata were considered for each primary study, categorized by author(s), document title (topic), abstract, keywords, DOI, year of publication, database, document type, and language; and 2) The selected studies were classified according to their purpose, type of study, and the country in which the research was conducted. For the development of this research, training standards and frameworks were used as references to support the study topic: the challenges of entrepreneurship in the era of artificial intelligence.

A systematic review requires categorization by type of study, as it typically relies on predefined inclusion and exclusion criteria established by the researchers. Among these categories—based on data quality, information practices, and privacy classification—relevant considerations were drawn from the work of Jaya et al. (2017), Inverardi et al. (n.d.), and Zhong et al. (2023).

Experimental or Quasi-Experimental Studies (EQ-ES) include randomized controlled trials and quasi-experimental studies (QE-S), which use similar methods but do not randomly assign participants. Observational Studies (OS) may include cohort studies, case-control studies, cross-sectional, and descriptive studies. These do not involve manipulation of variables; instead, they are limited to observing and collecting data. Case Studies (CS) involve detailed and in-depth analyses of one or more individual cases, offering valuable insights into rare or unique phenomena. Qualitative Studies (QS) focus on understanding participants’ experiences, meanings, and perceptions through methods such as in-depth interviews, focus groups, and content analysis to explore complex themes. Mixed Methods Studies (MS) combine both qualitative and quantitative research approaches, addressing complex research questions from multiple perspectives. Meta-analyses or Literature Reviews (SLR), although not primary studies, are essential for synthesizing existing research within a specific area.

Clearly defining the inclusion and exclusion criteria from the outset was essential to ensure that the systematic review of primary studies remained focused and rigorous. This step guaranteed that all selected studies were relevant to the research questions and met the established quality standards, Table 5.

It is important to note that these types of studies may vary in terms of data quality, information practices, and privacy categorization, as highlighted in the works of Inverardi et al. (n.d.), Jaya et al. (2017), and Zhong et al. (2023). Inclusion and exclusion criteria must be clearly defined at the outset of the systematic review to ensure that all selected studies are relevant to the research questions and meet established quality standards.

The review is summarized in the following section based on each of the research questions previously outlined.

As shown in Table 6, findings indicate that organizational culture and entrepreneurial mindset are critical factors shaping the adoption and effective use of AI technologies by entrepreneurs ( 10.1111/jofi.13302). Traditional entrepreneurial thinking has primarily emphasized economic gains, often overlooking broader social and environmental impacts.

Resistance to change within companies often arises in organizations that maintain a traditional culture or a generally conservative mindset ( Jia & Stan, 2021). In contrast, companies that foster a culture of openness to innovation and experimentation are significantly more likely to adopt AI effectively and maximize its benefits ( Guatemala Mariano et al., 2023).

Companies that are structured around artificial intelligence from the outset—such as AI-driven startups—tend to have a greater likelihood of success ( Jia & Stan, 2021). Additionally, certain studies offer a framework for addressing the convergence of the Fourth Industrial Revolution (4IR), the impact of the COVID-19 pandemic, and climate change, emphasizing the need for a renewed approach to entrepreneurship. These analyses suggest that the rapid pace of technological advancement has profound implications for both society and institutions, calling for a reassessment of the fundamental characteristics of entrepreneurship ( Guatemala Mariano et al., 2023).

As summarized in Table 7, the findings highlight the central role of automation and data analytics among the specific artificial intelligence strategies and tools used by entrepreneurs to improve operational efficiency and decision-making within their businesses. AI is employed to streamline operations, generate insights to support decision-making, and offer new ways to engage with both customers and employees. Machine learning and deep learning are frequently cited as key techniques for analyzing data and extracting actionable knowledge ( Weber et al., 2022), in relation to sustainability and education, AI and Big Data are increasingly used to address challenges linked to sustainable business practices. Furthermore, the literature provides an overview of how AI technologies are leveraged as valuable resources to enhance operational efficiency and decision-making in student-led enterprises across the Baltic countries ( Bickley et al., 2024), In relation to decision-making and risk, the development of AI-based risk assessment algorithms for ventures in rural areas is particularly noteworthy. Additionally, the creation of an AI factory is mentioned—an integrated system that combines data, algorithms, and experimentation to drive business growth and support strategic decision-making ( Wu et al., 2022). The section on innovation and marketing suggests that AI tools can be used to enhance operational efficiency and decision-making by analyzing large datasets and generating predictive insights. Moreover, it emphasizes that understanding new paradigms and adopting emerging technologies are essential for the effective implementation of digital marketing—an area where AI tools are increasingly being applied ( Guatemala Mariano et al., 2023). In terms of competitive advantages, the use of AI contributes to improved operational performance and more efficient decision-making within businesses. Specific tools such as data analytics, machine learning, and customer relationship management (CRM) systems are highlighted as key enablers in this process ( Rajagopal et al., 2022).

The results obtained across most of the studies analyzed indicate that entrepreneurs are leveraging AI capabilities to automate processes, analyze data, make informed decisions, and drive innovation and sustainability. These strategies and tools are enabling businesses to enhance operational efficiency, improve decision-making, and gain competitive advantages in an increasingly digital and data-driven business environment. “ Companies that design their organizational structure around AI from the outset—such as AI-driven startups—are more likely to succeed in adopting and leveraging these tools for rapid growth […]” ( Jia & Stan, 2021).

As shown in Table 8, a lack of training, combined with resistance to change, represents a significant barrier to the successful implementation of artificial intelligence technologies by entrepreneurs. The studies reviewed indicate that internal resistance, organizational inertia, and limited understanding of technological transformation often hinder the transition toward AI-driven business models. This underscores the importance of fostering an organizational culture that is open to change as a key enabler of AI adoption ( Jia & Stan, 2021), Moreover, insufficient knowledge, skills, and training in AI severely limit entrepreneurs’ ability to fully leverage these technologies. Several studies highlight the growing demand for AI-related competencies in the business environment and emphasize the role of educational institutions in addressing this skills gap ( Weber et al., 2022). Overall, resistance to change and inadequate training negatively affect the successful implementation of AI, reinforcing the need for continuous learning, collaboration, and capacity building ( Table 8). In summary, entrepreneurs must adopt a proactive approach to overcoming resistance and closing the AI skills gap by promoting an innovation-oriented culture and strengthening collaboration with educational institutions to develop relevant competencies ( Rajagopal et al., 2022).

As detailed in Table 9, government policies and regulations play a critical role in shaping the entrepreneurial and technological landscape, directly influencing investment decisions and the adoption of artificial intelligence. This influence is particularly significant for early-stage and growth-oriented ventures, where access to financial support and regulatory clarity is essential for prioritizing innovation and attracting investment ( Lammers et al., 2022). Public initiatives such as China’s “Mass Entrepreneurship and Innovation” program illustrate how policy frameworks can promote innovative activity and encourage AI adoption among high-tech startups operating in makerspaces (Y. Li et al., 2023). At the same time, AI-driven companies face regulatory challenges related to data governance, cybersecurity, and ethical risks, which can be further complicated by government regulations ( Jia & Stan, 2021) . Despite these challenges, policies and regulations remain essential for fostering sustainable and frugal innovation, even though their impact on investment decisions may vary across contexts ( Cetindamar et al., 2020). Another important factor is the adaptation of educational models and the optimization of the teaching environment within business education, particularly in virtual reality-based learning contexts. Artificial intelligence also underscores the role of governmental policies in supporting the training of entrepreneurs (W. Li et al., 2022). Overall, the findings summarized in Table 9 highlight that government policies and regulations significantly shape entrepreneurs’ investment behavior and technology adoption in the AI domain. These policies can foster innovation, provide financial support, address regulatory challenges, and promote the integration of emerging technologies such as AI within the entrepreneurial ecosystem.

As presented in Table 10, best practices and lessons learned from entrepreneurs who have successfully integrated artificial intelligence emphasize adaptability to continuous technological change as a strategic priority. Developing an agile mindset and maintaining flat organizational structures are identified as key factors that facilitate effective AI implementation and support rapid, informed decision-making ( Jia & Stan, 2021), Partnerships with venture capital investors also emerge as a critical strategy for mitigating the risks associated with AI adoption, as these alliances provide not only financial resources but also strategic guidance and technical expertise ( Jia & Stan, 2021b). Innovation and experimentation are consistently highlighted as essential practices, enabling entrepreneurs to redesign processes and business models in response to digital transformation ( Voronov et al., 2023). Finally, training and the development of digital skills are fundamental for overcoming integration challenges. Acquiring relevant competencies significantly enhances entrepreneurs’ ability to fully exploit the potential of AI within their operations ( Drydakis, 2022). Collectively, the evidence summarized in Table 10 demonstrates that agility, collaboration, experimentation, and continuous learning are central to successful AI integration in entrepreneurial ventures.

The Impact of Organizational Culture on AI Implementation and the Supporting Research on Cultural Differences, Which Contribute to Diverse Organizational Cultures Across Regions and Industrial Sectors That Influence the Adoption and Success of AI Technologies.

A longitudinal study of cultural transformation becomes necessary within the framework of longitudinal research representations to observe how companies can shift their organizational cultures to become more receptive to AI and which specific practices are most effective.

Likewise, the role of Government Policies and Regulations should be shaped based on an international comparative analysis that examines how different governmental policies and regulatory frameworks across countries influence the adoption of AI in startups and small businesses.

The evaluation of innovation support programs aims to study the effectiveness of specific initiatives—such as China’s “Mass Entrepreneurship and Innovation”—in promoting AI technologies, whether through training or skill development in AI.

This includes designing AI educational programs that explore both the creation and implementation of effective training initiatives to close the skills gap among entrepreneurs. Continuous education and its assessment play a key role in lifelong learning, enhancing entrepreneurs’ ability to adapt to and leverage new AI technologies.

Another suggested research avenue is the collaboration between startups and venture capital. This would involve case studies of successful partnerships where startups have worked with venture capital to identify key factors that contribute to the successful adoption of AI.

Another suggested line of research is innovative financing that integrates AI to improve risk and opportunity assessment in investments in tech startups. At the same time, it’s important to explore AI’s role in sustainability and social responsibility—examining how these technologies can be applied to enhance business sustainability and address environmental and social challenges, including the impact of AI on Corporate Social Responsibility (CSR).

Innovation and experimentation with AI are also key, using agile experimentation methods and frameworks for testing AI technologies in startups and small businesses. It is therefore necessary to combine this with innovation in AI-based business models that can be used to transform traditional business models and create new entrepreneurial paradigms.

Risk and security evaluation in AI should include the development and implementation of AI algorithms for risk assessment across various business contexts, with a special focus on rural areas and underrepresented sectors.