Digital transformation in smallholder agriculture: A systematic review of ai adoption and constraints in the Kenyan potato value chain

Abstract:

As Kenya's second most important food crop, Irish potato supports over 3.5 million value chain participants and generates an annual market value of approximately Ksh. 66 billion. However, smallholder farm-level productivity stagnates between 8 and 10 t/ha against a potential of 25 to 40 t/ha due to seed system failures, phytosanitary threats, and erratic rain-fed dependency. This study evaluates how emerging digital frameworks, specifically Artificial Intelligence (AI), can resolve these structural inefficiencies and foster long-term value chain sustainability. Using a systematic review methodology, we synthesized 28 peer-reviewed data sources and institutional records across five distinct thematic domains spanning phytosanitary diagnostics, precision agronomy, climate adaptation, post-harvest mitigation, and market aggregators. The findings indicate that empirical field deployments including Plant Village Nuru, Ujuzi Kilimo Soil Pal, and The Third Eye Project can catalyze yield improvements of 20% to 40% within smallholder systems. Computer vision platforms empower farmers with real-time phytosanitary diagnostics, while IoT soil telemetry networks and drone-based optical sensing optimize micro-irrigation schedules. Additionally, predictive algorithmic platforms streamline digital market linkages to bypass exploitative brokers. However, macro-scale adoption remains severely hindered by the rural digital divide, last-mile infrastructural deficits, and low digital literacy. We conclude that stakeholder interventions must shift from isolated private pilots toward unified national agricultural frameworks integrating technological deployment with aggressive rural capacity-building and strict data governance policies.