<aside>
Principal Authors: Malcolm Durosaye, Mariam Adeyemo, Ahmad Raji, Ridwan Sorunke, Shuko Musemangezhi, and Femi Royal
</aside>
The transformative potential of artificial intelligence in agriculture is increasingly recognized, yet most large language models (LLMs) remain inaccessible to small-scale producers (SSPs) in Africa due to high computational demands and limited localization.
This paper explores the development and deployment of affordable, locally fine-tuned large language models (Agri-LLMs) through multi-stakeholder public-private partnerships integrated with existing agricultural extension systems, bridging the AI adoption gap for small-scale producers.
It examines how both African and Indian farmers can benefit from localized, low-power, and cost-efficient AI models that respond to their linguistic, cultural, and operational realities. The paper also analyzes India’s Digital Public Infrastructure (DPI) approach to agricultural data and localization, alongside Africa’s emerging ecosystem of agricultural AI innovation.
<aside>
Keywords: Localized Agri-LLMs, Smallholder Farmers, Accessibility and Affordability, Agricultural AI Ecosystem, Policy and Governance.
</aside>
Small-scale producers (SSPs) are essential to agriculture, accounting for one-third of global food production and 84% of the world’s agricultural holdings (Ritchie 2021). Their role is very important for food security and economic growth, especially in sub-Saharan Africa, where farming supports livelihoods and national economies. Given the challenges of extreme weather events post for over 95% of African’s rain-fed farms and weak extension systems, digital services offer powerful tools to address these and other challenges farmers face. In countries like Nigeria, SSPs produce up to [90%](https://www.pwc.com/ng/en/assets/pdf/afcfta-agribusiness-current-state-nigeria-agriculture-sector.pdf#:~:text=More than 80% of Nigeria’s farmers are smallholder,in 2019 compared to N549.3 billion in 2018.) of the food consumed domestically (PWC 2020). In India, small and marginal farmers represent nearly 86 % of all agricultural holdings (Gupta and Li 2021; Oxfam India 2022).
Despite their significance, SSPs struggle to access the technology and information needed to improve their productivity. Many farmers face challenges associated with technological infrastructure, language differences, and a lack of digital skills. This situation makes it hard for them to adopt modern farming practices to improve productivity.
Along the value chain, knowledge gap remains a concern. many agricultural extension agents - responsible for bridging the knowledge gaps of small-scale farmers - have little to moderate knowledge in the latest technologies, such as IoTs and Artificial Intelligence that could aid in decision-making, crop management, and sustainability efforts (Asante 2025). This situation limits both the reach and effectiveness of new agricultural innovations.
Emerging technologies such as Artificial Intelligence (AI) present a scalable way to deliver agricultural information and support. Technologies like Large Language Models (LLMs) and computer vision have the potential to change how farmers access timely and relevant advice using natural language or visual inputs. These technologies support the creation of tools that understand local dialects and languages, making information easier to access for diverse rural communities (Yang et al. 2023; Almufareh et al. 2025).
Natural disasters such as extreme drought and flooding, intensified by climate change, have led to increased pest and disease infestations, threatening food security. AI-based innovations, combined with remote sensing, have contributed to climate-resilient agriculture by equipping farmers with data-driven insights to optimize resource allocation and planning (Pathania et al. 2025).
Even with existing technological innovations and advancements, many small-scale farmers remain excluded due to high costs, language barriers, and limited digital skills (Choruma et al. 2024). These challenges leave them marginalized and unable to access crucial agricultural information including diseases, climate and weather, and market data. Without intervention, these farmers risk falling further behind, threatening their livelihoods and local food security.
Given the threats posed by climate change and market fluctuations, empowering both small-scale farmers and their extension agents with cost-effective information dissemination channel and advisory companion tailored to their specific agricultural needs and regional contexts has become increasingly important. The vision of this discussion is straightforward and transformative, offering compelling value propositions for multiple stakeholders.
At the individual level, the efficiency and effectiveness of extension services can be optimized if local, in-country extension agents have per user, low-cost access to low-powered, leading-edge LLMs specifically fine-tuned for small-scale producer contexts. This would fundamentally transform the advisory capacity of extension services, enable highly localized, real-time farming advice, and catalyze improvements in productivity, resilience, and sustainability among smallholders (Prestegaard-Wilson and Vitale 2024). At the same time, this vision presents a compelling opportunity for investors and innovators to deliver and scale affordable AI solutions tailored to small-scale producers. By making advanced AI both accessible and cost-effective, they can unlock new value for farmers, foster inclusive innovation, and drive sustainable growth across rural communities and the broader agricultural ecosystem.
At a time when global food systems face increasing pressures, this paper advocates for AI solutions that address the core issues of accessibility, affordability, and contextual relevance. By enhancing the capacity of extension services, these solutions aim to ensure that critical agricultural knowledge reaches the small-scale producers who need it most. The paper aims to accelerate the much-needed collaboration among key stakeholders including governments, research institutions, technology developers, agricultural stakeholders, and donors on the responsible and equitable use of AI in agriculture. It highlights the current landscape, identifies emerging opportunities and persistent challenges, and frames key questions necessary to guide inclusive and sustainable innovation.