Written by Miriam Murigi
Millions of Kenyan farmers who rely on basic mobile phones will soon have access to personalized climate-smart farming advice through a new artificial intelligence initiative being advanced by the Bioversity International coalition and CIAT.
Central to the Alliance’s proposal is the development of AI-driven agricultural advisories delivered through USSD technology, allowing farmers to receive customized guidance using a simple handset.
By downscaling complex geospatial and climate data, the system generates local, farm-specific recommendations for cropping schedules, input use, and climate risk.
“This initiative aims to address a long-standing gap in digital agriculture. Tremendous amounts of climate and agricultural data are already being generated, but much of it remains inaccessible to smallholder farmers who lack smartphones, internet connectivity or digital literacy,” said Anastasia Wahome, Alliance research team leader.
The Alliance argues that Kenya already has much of the infrastructure needed to make the system work. The Kenya Integrated Agricultural Management Information System (KIAMIS) digitally registers 6.5 million farmers, creating one of the most comprehensive agricultural databases in the country.
Additionally, the Ministry of Agriculture conducts monthly crop monitoring through extension agents working at district, sub-district and district levels.
“The data, skills and infrastructure are already in place,” Wahome said. “What we have to do now is translate them into practical services that farmers can use. The Alliance brings capabilities in remote sensing, spatial analysis, cloud computing and advanced analytics.”
Through its AI-powered USSD service, the Alliance envisions transforming sophisticated data into simple, actionable messages that are timely, locally relevant. Farmers will be able to interact with the system by responding to prompts and asking basic questions without the need for an internet-enabled device.
The main focus of this proposal is language inclusion. Many Kenyan farmers primarily communicate in local languages, which are not well supported by existing digital tools. Therefore, the Alliance advocates for open and inclusive training of AI models to enable systems to understand and respond to these languages, increasing trust and adoption.
Its ambitions extend beyond individual farmers. The Alliance is driving the integration of AI systems across Kenya’s broader agricultural ecosystem and making data accessible and usable by actors in other sectors.
The group says the goal is to move from basic recommendations to actionable information.
“With accurate and timely information, farmers can reduce production costs, increase yields and improve incomes by applying the right inputs in the right amounts and at the right time. Better climate risk information can also help farmers prepare for shocks such as droughts and delayed rainfall,” she said.
Data can create new opportunities for agribusinesses, insurance companies, and financial institutions. Insurers can use field-level insights to improve risk assessments and increase coverage, while lenders can extend credit with more confidence. Meanwhile, supply chain actors could benefit from improved yield estimation to support storage, processing, and logistics planning.
“If we know in advance what the harvest is likely to be, we can plan for surpluses or shortages,” Wahome said, noting that predictability is the key to supply chain stability and market efficiency.
The organization is currently calling for collaborative partnerships to co-develop, test, and expand the AI-on-USSD system. While we acknowledge that this effort will require significant investment in data, testing, and refinement, the Alliance maintains that it is practical and scalable.
As climate change continues to increase uncertainties in agricultural production, the Alliance argues that the future of agriculture lies not in generating more data, but in making existing data usable at the forefront.
“By delivering intelligence through the simplest technology available, we believe AI-on-USSD can redefine how digital agriculture reaches the last mile and who it ultimately serves,” she said.
