The global food system is stressed due to the increasing pressures of climate change, resource scarcity, population growth, and shifting consumption patterns. Traditional agricultural practices are not adequate to meet rising food demands while simultaneously addressing concerns of environmental sustainability, food security, and rural livelihoods. Consequently, agri-tech startups have emerged as pivotal actors in reshaping food production, distribution, and consumption systems, offering innovative solutions that integrate an end-to-end supply chain with digital technologies, data analytics, and sustainable practices. By leveraging tools such as precision farming, blockchain-enabled supply chains, remote sensing, and artificial intelligence (AI)-driven advisory services, these startups not only enhance agricultural productivity but also contribute to the resilience and sustainability of food systems.

The current global food system faces significant challenges stemming from destructive agricultural practices that risk environmental systems, production chain issues, and social crises, which primarily impact developing countries (Devaux et al., 2022). The population needs sustainable approaches that use innovation to empower local communities through food distribution processes (Lisboa et al., 2024).

Local food systems have gained popularity in recent times because they demonstrate the potential to provide security to underserved populations (Horst et al., 2024). The joint approaches between agriculture market initiatives and community programs produce improved results in fresh produce intake and dietary quality (Garrity et al., 2024). Local food programs face obstacles because their growth potential remains restricted due to barriers to entry and cultural practices, which necessitate careful planning for successful implementation (Raihan et al., 2024). Alternative and local food systems frame the essential strategies for developing sustainable food systems that enhance equality (Ruben et al., 2021). Local food initiatives transform the processes of food production, distribution, and consumption through community-supported agriculture, regenerative farming strategies, and digital direct-to-consumer channels (Glaros et al., 2023).

Identification and assessment of Critical Success Factors (CSFs) are essential to ensure the long-term sustainability of agri-tech startups, particularly in developing economies where resource constraints, infrastructural gaps, and smallholder dependency create unique challenges. CSFs provide a structured lens for identifying the most influential factors that determine whether agri-tech ventures can scale beyond pilot projects and deliver measurable impact. By systematically evaluating CSFs, stakeholders can prioritize scarce resources toward high-leverage enablers, creating multiplier effects across the ecosystem. For instance, investments in digital infrastructure enhance connectivity, enabling financial inclusion, advisory services, and traceability systems. Similarly, robust policy frameworks can unlock markets and attract investor confidence, while innovation ecosystems support the development of context-relevant, affordable technologies for farmers. Thus, CSF assessment bridges the gap between startup goals and systemic transformation, offering policymakers, investors, and entrepreneurs a roadmap for building resilient and scalable agri-tech ecosystems that contribute to food security, rural livelihoods, and climate-smart agriculture. Developing sustainable supply chain value requires agri-tech startups to integrate their innovations and models with sustainable operational and value principles (Rialti et al., 2022). Through responsible supply chains, resilience in the food system can be met to fulfill the Sustainable Development Goals SDG 2, SDG 12, and SDG 13 (Michel et al., 2024).

Although the issue of sustainable agri-food systems is gaining considerable consideration, the available literature is primarily centered on single aspects, including technology adoption, organizational preparedness, or ecological regulations. Limited studies are for a comprehensive and analytical examination of related aspects to one another, especially in the developing economies where the supply chain is fragmented, infrastructure is limited, and uncertainty exists in the set policies. Therefore, this study explores the future of sustainable and alternative food systems through the lens of the Technology-Organization-Environment (T-O-E)-based CSFs. However, T-O-E has been implemented in a few studies on agri-tech; the empirical evidence is limited on how the CSFs interact causally to influence the scalability of agri-tech startups and their sustainability. The study has employed Fuzzy DEMATEL, which helps the researchers and policymakers to identify functional relationships between CSFs while establishing strategic priorities and constructing policies (Zhao et al., 2024).

This study is significant as it helps to fill the gaps by adopting a T-O-E-based CSF assessment using Fuzzy DEMATEL methodology, which enables the determination and modeling of causal linkages in agri-tech ecosystems. This contributes to the theoretical understanding of agri-tech-facilitated food system transformations and provides practical knowledge to policymakers, decision makers, investors, and entrepreneurs. Specifically, it identifies strategic leverage points, such as digital infrastructure, enabling policy frameworks, and innovation capacity that can stimulate the rate of change toward the sustainable food system in resource-constrained settings. The study identifies the following particular aims:

The research contributes to the leverage points that can accelerate the role of agri-Tech startups in achieving the United Nations’ Sustainable Development Goals (SDGs), particularly those related to zero hunger, responsible production, and climate action.

The subsequent parts of this research are as follows: Section 2 discusses a literature review of sustainable food systems and agri-tech innovation. Section 3 details the Fuzzy DEMATEL research methodology. Section 4 discusses results from Fuzzy DEMATEL. Section 5 discusses findings with theoretical and practical implications. Section 6 concludes and provides research recommendations for future study.

The present study employs the Fuzzy DEMATEL method to systematically analyze the interdependencies among the identified CSFs within agri-tech ecosystems. A panel of 15 experts from supply chain management, logistics, operations, and agri-business was consulted to provide pairwise comparisons of the factors using a five-point linguistic scale shown in Table 3. These linguistic assessments were transformed into triangular fuzzy numbers (TFNs) to capture uncertainty and subjectivity in expert judgments. Table 3 presents the linguistic scales.

The aggregated responses were then used to construct the fuzzy direct-relation matrix, which was subsequently normalized to ensure comparability. Through iterative computation, the total relation matrix was derived for both direct and indirect influences among the CSFs. Finally, prominence (D + R) and relation (D–R) values were calculated to classify the CSFs into cause-and-effect groups. The research methodology flow is shown in Figure 1.

The fuzzy DEMATEL method enhances the evaluation of ambiguous empirical data. Utilizing fuzzy logic, decision-makers can assess alternatives according to high, medium, and low assessment tiers. This method, with its causal diagram framework, equips decision-makers with the means to identify critical success factors, hence enhancing the methodical examination of the situation. Fuzzy DEMATEL exhibits superior efficacy compared to hierarchical methods in identifying intricate interrelations within multiple systems, despite its utility in the implementation process. Further, it demonstrates adaptability and responsiveness in decision-making, providing advantages to complex information systems with flexible decision-making capabilities (Sharma S. K. et al., 2024). A questionnaire was designed for experts to conduct pairwise comparisons before the implementation of fuzzy DEMATEL. The information regarding the experts’ details is shown in Table 1.

The expert panel for this study was purposely composed of professionals with backgrounds in supply chain management, operations, and agri-business systems, as these individuals possess direct experience in understanding adoption bottlenecks, resource flows, and systemic challenges within agri-tech ecosystems. Their insights were highly valuable for identifying causal linkages between enablers and barriers, which aligns with the objectives of applying Fuzzy DEMATEL to map interdependencies among CSFs. The steps for implementing Fuzzy DEMATEL have been summarized below:

Fuzzy DEMATEL analysis was carried out to explore the causal links between 25 CSFs that determine the scalability and sustainability of agri-tech startups in developing countries. By evaluating the cause-and-effect dynamics through fuzzy logic and expert input, the study classified the factors into two key groups: cause group (influential factors) and effect group (influenced factors). The Initial Direct-Relation Matrix (Z) was created using Equations 1–3, which were developed based on expert responses applying the linguistic scale outlined in Table 3. The fuzzy direct relation matrix is derived from Equations 4–6. Equations 7–9 yield a comprehensive relation matrix, as presented in Appendix Table S1.

Key influential (cause) factors identified include CRSF3 (Technology Infrastructure), CRSF8 (Government Support), and CRSF21 (Policy Frameworks), all of which have high positive (Di–Ri) values, indicating a strong driving impact on the system. Conversely, effect factors such as CRSF11 (Public Engagement) and CRSF24 (Financial Support Mechanisms) showed high (Ri) but relatively lower (Di), suggesting these are outcomes influenced by other variables. The impact results are shown in Table 4.

Several critical success factors (CSFs) belong to the cause category because they determine the direction of other system factors. CRSF1, CRSF3, CRSF4, CRSF6, CRSF8, CRSF16, CRSF18, CRSF20, CRSF21, CRSF23, and CRSF25. The systems’ central drivers are CRSF21 with Di+Ri of 10.198 and Di−Ri of 0.409, and CRSF8 with 10.193 and 0.391. When designing intervention strategies, CRSF18 (9.805, 0.420) and CRSF6 (9.690, 0.390) are the highest priority elements. The initiative requires these driving factors to achieve success.

The effect group includes the rest of the factors because their influence stems from outside variables. CRSF2, CRSF5, CRSF7, CRSF9, CRSF10, CRSF11, CRSF12, CRSF13, CRSF14, CRSF15, CRSF17, CRSF19, CRSF22, CRSF24 Some notable effect factors include: A high level of dependency exists between CRSF5 (Di−Ri = −0.782) and CRSF22 (−0.554), which demonstrates these CSFs receive substantial influence from other areas of the healthcare system. The high influence rating (Di+Ri > 10) of CRSF11 and CRSF10 indicates that these factors tend to be system-sensitive rather than driving change elements because they react to changes across the system. The performance indicators function as monitors that help assess system changes resulting from enhancements in the cause group.

The causal diagram in Figure 2 illustrates strong interdependencies, emphasizing the importance of investing in core enablers such as digital infrastructure, supportive policies, and institutional frameworks to strengthen the entire smart circular economy ecosystem. These insights offer policymakers a prioritized roadmap for informed strategic decision-making.

The fuzzy DEMATEL analysis enabled a comprehensive assessment of the interdependencies among the 25 CSFs. By categorizing the CSFs into causal and effect groups, the analysis elucidated which factors function as strategic drivers and which are outcomes shaped by systemic interactions.

The cause group comprised CRSF1, CRSF3, CRSF4, CRSF6, CRSF8, CRSF16, CRSF18, CRSF20, CRSF21, CRSF23, and CRSF25, indicating their dominant influence within the system. Notably, CRSF21 (Di − Ri = 0.409), CRSF8 (0.391), CRSF6 (0.390), and CRSF18 (0.420) exhibited both high causality and strong systemic prominence (Di+Ri > 9.6), underscoring their role as key drivers. The analysis reveals crucial aspects about digital infrastructure as well as smart governance and advanced technological capabilities, which enable sustainable transitions (Govindan et al., 2022). Evidence from digital platforms and IoT systems has increased the level of transparency in tracking, which has been supported by CRSF21 and CRSF8 (Singh and Kumar, 2024). The close association between CRSF6 and CRSF18 proves that an efficient stakeholder partnership and information exchange mechanisms are primary factors in the implementation of a circular economy (Wielopolski and Bulthuis, 2023). These drivers are consistent with the strategic enablers discussed in the literature as fundamental core factors that require urgent attention to enhance overall sustainability outcomes (Onukwulu et al., 2025; Falcone et al., 2022; Zhao et al., 2022).

Conversely, the effect group included CRSF5, CRSF7, CRSF9, CRSF10, CRSF11, CRSF12, CRSF13, CRSF14, CRSF15, CRSF17, CRSF19, CRSF22, and CRSF24, which primarily represent outcomes shaped by systemic dynamics rather than initiators of change. Among these, CRSF5 (Di − Ri = −0.782) and CRSF22 (−0.554) exhibited the greatest dependency, showing that they were more sensitive to upstream drivers. This observation aligns with previous studies that have highlighted the performance measurements of circular economies, i.e., waste reduction, emission mitigation, and resource conservation, depend on the successful experience of upstream enablers, i.e., policy frameworks, leadership commitment, and digital infrastructure (Hartley et al., 2023; Mottet et al., 2020).

Notably, CRSF10 and CRSF11 had a high systemic prominence (Di+Ri > 10) but with negative values of net influence (Di [?]). Ri (0), which means that they are at the center of the system, but are driven externally. This trend underscores the importance of continually monitoring such indicators to assess the effectiveness of circular economy strategies (Govindan and Hasanagic, 2018).

The causal factors identified were CRSF21 (supportive policy frameworks) and CRSF8 (government support). This is consistent with the National Digital Agriculture Strategy of India and the Kenya Agricultural Sector Transformation and Growth Strategy, which indicate that well-designed regulations and strategic digital interventions trigger innovation among startups. In other examples, CRSF6 (digital infrastructure) is a key determinant of success, as in Bangladesh, where Grameenphone has contributed to the delivery of last-mile agri-advisory using its extensive network to rural districts. Furthermore, in Nigeria, Hello Tractor has benefited from the rural penetration of connectivity to support its share machinery system.

CRSF18 (stakeholder collaboration) and CRSF3 (technology infrastructure) are organizational-level drivers that can be seen in initiatives like DeHaat (India) and AgUnity (Papua New Guinea and Kenya), which use digital platforms and farmer-producer organization (FPO) linkages to expand outreach and blockchain-enabled tools to improve the belief participants place in transactions, respectively. The examples suggest that technological possibilities become effective only when incorporated with inclusive sharing practices and culture-relevant business concepts (Alfonsi et al., 2024).

On the effect side, CRSF5 (financial constraints) and CRSF22 (market linkages) have high dependency. The case of East Africa-based One Acre Fund demonstrates that even well-designed digital platforms fail to efficiently serve the population unless farmers have access to financing or markets. Similarly, the example of EcoPost (Kenya), which suggests recycling agricultural waste into building materials, demonstrates that innovative approaches to the circular economy require a comprehensive and robust backup support system in the form of policies and consumer recognition. Although the fuzzy DEMATEL analysis provides structural explanations, the transformative possibilities are related to the capability of modifying causal enablers to create conditions, particularly in developing countries. The policymakers and ecosystem actors should emphasize CSFs that expand digital infrastructure, policy coordination, and inclusive financing to achieve the systemic potential of agri-tech startups in building sustainable food systems. Inclusion, particularly of women, smallholders, and marginalized groups who form the workforce of developing country agriculture, is also a feature of sustainable food systems. It is revealed that factors of stakeholder collaboration, digital literacy, and NGO partnerships directly influence equity outcomes.

The greenhouse-in-a-box program designed by Kheyti in India, and peer-to-peer training activities provided to women in Ethiopia by Digital Green can serve as another example of how the benefits of accessible technology, as well as a knowledge-sharing system, can empower women and smallholders. Similarly, instances of inclusive financing can be observed in programs as One Acre Fund in East Africa, where marginalized farmers can take advantage of adopting innovations. The addition of inclusive practices would enable agri-tech startups to provide benefits to agro-efficiency while also contributing to the fairness and sustainability of food systems.

Recent empirical research further strengthens our findings. A study by Isher et al. (2024) shows that incubation services are strongly linked to the success of agri-tech startups in Indian states through mentorship and infrastructure support. Another study by Isher and Gangwar (2025) corroborates this, highlighting the role of incubators in enhancing job creation and attracting venture funding. Further, Zhao et al. (2025) provide compelling evidence that investments in digital infrastructure in China significantly enhance green agricultural productivity and rural incomes. Similarly, Ofosu-Ampong et al.’s (2025) review reveals that digital agro-advisory tools in the Global South emphasize how behaviorally informed tool design increases farmer uptake.

The empirical grounding of the causal dynamics we identified via fuzzy DEMATEL is further strengthened by the following relevant studies: Kantoğlu et al. (2025) exhibited in Sub-Saharan Africa that agri-tech and value-added agricultural inputs significantly drive sustainability outcomes and emissions mitigation, reinforcing the current study model’s cause group factors. Choruma et al. (2024) and Nxumalo and Chauke (2025) highlight on-ground barriers (digital literacy, infrastructure, financial constraints) alongside success stories, which align with the current study’s effect group and practical barriers. From the Indian context, Dora et al. (2022) identified technology readiness, regulatory compliance, and perceived benefits among key CSFs of AI adoption in the food supply chain, which mirror several of the current study’s CSFs in the technology and organizational categories. Finally, Quayson et al. (2024) applied a fuzzy DEMATEL approach to the cocoa supply chain in Ghana to identify influence relationships among barriers, such as lack of management support, low awareness, and regulatory challenges; these can directly support our recommendations for interventions.

Recent state-level implementations in India also reinforce these trends. For instance, Maharashtra’s MahaAgri-AI Policy (2025–2029) and Andhra Pradesh’s fully digital APAIMS 2.0 scheme employ digital platforms and AI to enhance farmer services and data exchange, highlighting the catalytic role of policy-driven infrastructure (TOI). Lastly, Zafar et al. (2025) suggest methodological advances in data governance, such as privacy-preserving data linkage frameworks, address key ethical and implementation concerns in public–private data collaboration.

This study employed the T-O-E framework and fuzzy DEMATEL to analyze 25 CSFs for agri-tech startups in developing economies. The findings reveal that systemic enablers such as supportive policy frameworks, robust digital infrastructure, and innovation capacity emerge as dominant enablers. Conversely, financial access, market linkages, and consumer adoption appear to be dependent factors that improve only when causal enablers are addressed. The findings highlight that startups alone cannot drive transformation; supportive policies, robust infrastructure, and inclusive innovation systems are required to achieve scalable and sustainable outcomes in developing economies. The findings not only offer theoretical insights but also provide actionable guidance for stakeholders, including policymakers, investors, and startup founders, who are working toward building resilient, responsible, and inclusive food systems. For policymakers, this study suggests investing in rural connectivity, creating national digital agriculture strategies, and ensuring inclusive regulatory frameworks that reduce adoption barriers for smallholders and women farmers. Investors and funding agencies must de-risk early-stage ventures through blended finance, green funds, and incubation support, while aligning capital flows with measurable SDG outcomes. Startup founders, in turn, need to focus on affordable, context-sensitive business models, collaborate with cooperatives and FPOs to build trust, and embed user-driven R&D to ensure scalability. Concisely, these targeted actions can strengthen agri-tech ecosystems and position startups as catalysts for resilient and equitable food systems.

However, this study has a few limitations. First, the analysis relies on expert judgment. Although expert inputs capture structured knowledge, they may introduce subjectivity and potential bias, especially when experts share similar professional backgrounds or perspectives. Second, the sample of experts undertaken is limited in size and diversity; this may underrepresent other important ecosystem actors such as grassroots farmers, community organizations, and policymakers at local levels. Third, the study does not incorporate empirical validation from farmer surveys, startup field data, or impact assessments of ongoing agri-tech interventions. Future research should address these gaps by integrating larger and more diverse stakeholder groups, including farmers, agri-startups, and rural extension workers, to improve representativeness. The methodology can be extended by combining expert judgment with real-world startup performance data (e.g., investment levels, user growth, supply chain emissions) to validate and refine the fuzzy DEMATEL model. In addition, a mixed-method approach combining expert-based modeling with field surveys, case studies, and secondary adoption data would strengthen the credibility and practical relevance of the findings. Finally, the dynamic nature of agri-tech ecosystems may lead to changes in these success factors over time that need to be assessed. While technology is a central theme, this study does not explore in-depth distinctions between types of digital tools (e.g., AI, IoT, blockchain), which may have different levels of impact on agri-food sustainability. Future researchers can conduct time-series analyses to observe how the influence of CSFs evolves as agri-tech startups mature and external conditions (e.g., policies, market forces) change. The study can be replicated in a broader sample of countries, especially emerging economies in Latin America, Southeast Asia, and Africa. It can also be extended to explore sector-specific startups (e.g., aquaculture, urban farming). The study is insightful for developing systems dynamics or agent-based simulation models to explore how different policy interventions affect the readiness and impact of agri-tech startups over time.