Artificial Intelligence and Agricultural Risk Management for Smallholder Cowpea Farmers and Processors in Niger State, Nigeria

Agriculture remains a critical pillar of Nigeria’seconomy, employing over 70% of the rural workforce and contributing significantly to national GDP, food security, and livelihoods (FAO, 2023; NBS, 2022). Among key staple crops, cowpea (Vigna unguiculata)—commonly known as black-eyed pea—plays a dual role: as a high-protein dietary staple and as a commercially valuable commodity for both rural farmers and urban markets (Kamilaris and Prenafeta-Boldú, 2018). Nigeria is the world’slargest producer and consumer of cowpea, with an estimated annual production exceeding 3 million metric tonnes (Olawuyi and Ogunniyi, 2023; Adeyemi et al., 2025). Despite its economic and nutritional importance, cowpea production in Nigeria remains highly susceptible to a variety of risks that undermine both productivity and profitability. These risks include unpredictable rainfall patterns, extended dry spells, rising temperatures, and increasing incidences of pest and disease outbreaks, particularly Maruca vitrata and Callosobruchus maculatus (Ibrahim, Shettima and Usman, 2019; Kamai Zakka and Abdulraheem, 2020).

These biotic and abiotic stressors, compounded by market price volatility, low access to formal insurance products, and weak infrastructural support systems, create a hostile operating environment for smallholder cowpea farmers (Joel et al., 2025). Furthermore, the post-harvest segment—dominated by informal processors, many of whom are women—is equally exposed to high levels of risk through poor storage infrastructure, susceptibility to pest damage, and the absence of standardized quality control systems (Maisule et al., 2025). As a result, farmgate profits remain minimal, post-harvest losses are estimated to range between 15% and 30%, and producers struggle to maintain consistent supply to meet both local and export market demands (Ajayi, Fatunbi and Akinbamijo, 2020; Olomola, 2021).

Traditional risk management strategies employed by cowpea farmers and processors in Nigeria tend to be reactive and informal. These include diversified cropping, delayed planting, reliance on indigenous knowledge systems, and limited engagement with formal creditor insurance mechanisms (Ibrahim et al., 2019 Olawumi et al., 2025). While these strategies reflect a high degree of local adaptation, they are often insufficient in the face of increasingly erratic climatic patterns and volatile agricultural markets driven by global and regional trade disruptions. Additionally, smallholder cowpea producers frequently lack timely access to accurate meteorological data, pest forecasts, or market intelligence, which significantly limits their capacity to make informed decisions (Oyediji et al., 2025). In this context, Artificial Intelligence (AI) has emerged globally as a potentially transformative tool for enhancing agricultural risk management by offering predictive, real-time, and data-rich support systems across the agricultural value chain. AI-driven systems are increasingly capable of leveraging large datasets ranging from satellite imagery and weather data to market trends and pest infestation records to generate actionable insights that could help farmers and processors anticipate risks and respond more effectively. For instance, AImodels trained on historical weather patterns can now forecast drought conditions with considerable accuracy, while machine vision tools can identify early signs of pest infestation on leaves through smartphone applications (Kamilaris and Prenafeta-Boldú, 2018; Adebayo, Lawal and Alamu, 2022). In theory, the use of such AItools could dramatically shift the paradigm of risk management from reactive coping to anticipatory planning. However, the real-world integration of AIinto smallholder agricultural systems in Nigeria remains limited and faces a range of critical challenges (Oyediji et al., 2024; Olawumi et al., 2025). The application of AI in agriculture, particularly in smallholder systems in sub-Saharan Africa, is constrained by several interrelated technological, socio-economic, and institutional barriers. First, the digital divide remains a significant obstacle. Many rural areas in Nigeria lack reliable internet connectivity, access to smartphones, or electricity infrastructure, all of which are foundational for AI-enabled platforms to function effectively (Barrett and Rose, 2022). Digital literacy among rural farmers and processors also remains low, further limiting the capacity of these stakeholders to utilize or even trust AI-driven tools. Moreover, many existing AItools in agriculture are designed for commercial agribusinesses or industrial-scale farms and are poorly adapted to the resource constraints and knowledge systems of smallholder farmers. For example, pest detection algorithms that require high-resolution imaging or cloud-based computing may be inaccessible to most farmers in rural northern Nigeria. Even where relevant AItools are available, adoption remains low due to lack of trust, poor user experience, limited training, and the absence of intermediary support systems such as local extension agents equipped to interpret and translate AI-generated information (Hellin and Camacho, 2017; Lai-Solarin et al., 2025). For cowpea processors, the post-harvest segment has received even less attention in AIresearch, despite its critical importance for food security and farmer incomes. Issues such as mold detection, storage optimization, and supply chain monitoring remain underdeveloped in the AIliterature, further illustrating the narrow scope of current technological interventions (Sennuga et al., 2025).

Given these constraints and the unique characteristics of cowpea farming and processing systems in Nigeria, there is an urgent need to better understand how AItechnologies can interface with the specific risk experiences of smallholder actors across the value chain. Cowpea stakeholders are not a homogeneous group; they differ by gender, region, scale of operation, access to inputs, and level of formal education. Furthermore, the informal nature of many cowpea markets and the dominance of unregulated input systems introduce further complexity to risk prediction and mitigation. These factors necessitate a context-specific analysis of both the technological capabilities and the social dynamics that mediate AIadoption and effectiveness. As Nigeria moves forward with its digital agriculture agenda—articulated in the National Agricultural Technology and Innovation Policy (NATIP, 2021–2025)—there is a critical need to generate empirical evidence on how AIcan serve not merely as a technological fix but as a support system that aligns with the everyday realities of rural farmers and processors. Addressing this knowledge gap is essential to ensure that AI-enabled agricultural systems are inclusive, relevant, and responsive to local needs, particularly in under-researched crops like cowpea that are vital for both economic resilience and nutritional security. This study aims to evaluate the interface between AItools and the risk experiences of smallholder cowpea stakeholders in Nigeria. To accomplish this, the following objectives are put forward to: i. Describe the socio-economic characteristics of smallholder cowpea farmers and processors in the study area. ii. Assess the levels of awareness, accessibility, and patterns of adoption of AI-enabled technologies among smallholder cowpea stakeholders in the study area iii. Examine the availability, functionality, and relevance of existing AItools designed to address agricultural risks, with specific attention to their applicability within cowpea-based farming systems in the study area iv. Analyze the socio-economic, demographic, and institutional factors influencing the adoption and effectiveness of AIapplications for risk management in cowpea farming and processing in the study area. v. Assess the challenges faced by smallholder cowpea farmers and processors in adopting AIfor agricultural risk management in the study area.