Microsoft is backing the use of artificial intelligence to improve climate information for African farmers, as increasingly unpredictable rainfall, prolonged dry spells and recurrent flooding threaten agricultural production and the livelihoods of millions across the continent.
The technology company is supporting research and initiatives aimed at making weather forecasts more reliable, translating climate information into practical agricultural advice and delivering guidance to farmers in local languages, according to the Head of Microsoft’s AI for Good Lab in Nairobi, Dr Girmaw Tadesse.
The push comes as farmers in countries such as Nigeria face growing difficulties in determining when to plant, manage agricultural inputs and prepare for adverse weather conditions, with limited access to timely and locally relevant climate information adding to the risks facing agricultural production.
For millions of farmers across Africa, the timing of rainfall can make the difference between a successful harvest and financial losses. Planting before the rains begin risks seed failure, while late planting may deny crops the adequate growing period needed to reach maturity.
Tadesse said the challenge was no longer simply obtaining climate information but ensuring that farmers and public institutions could translate forecasts into practical decisions before climate risks developed into crises.
“As climate variability intensifies, the challenge is no longer simply accessing climate information, but translating it into practical, timely decisions that protect agricultural productivity, strengthen resilience and safeguard livelihoods,” he said.
Artificial intelligence could help tackle some of these challenges by enhancing current weather forecasting systems and tailoring climate information to the specific agricultural needs of local communities.
One approach highlighted by Tadesse is Probabilistic Bias Correction, a machine-learning framework developed by Microsoft Research to correct recurring errors in existing weather models and improve forecasts over the medium term.
Improved forecasting accuracy could provide farmers with timely information ahead of planting seasons, enabling them to determine the right time to sow, plan the use of farm inputs and prepare for potential dry spells.
However, improved forecasting alone would not guarantee better harvests. Agricultural advice must account for differences in soil conditions, crop varieties, water availability, production costs and farmers’ existing knowledge.
Technology-enabled agricultural services illustrate how digital platforms can connect farmers with resources needed to respond to production challenges. Nigerian agritech company Farmcrowdy, for instance, developed a digital platform connecting farmers with financing and agricultural support. Although it is an example of digital agricultural finance rather than AI-powered weather forecasting, it demonstrates how technology can help address financing constraints facing smallholder farmers.
More directly focused on climate information is Nigeria’s Digital Climate Advisory Services initiative, involving the Nigerian Meteorological Agency and MTN Nigeria. Designed to deliver location-specific weather guidance through mobile channels, the initiative aims to help farmers make decisions about planting, irrigation and harvesting.
The programme illustrates how meteorological expertise can be combined with telecommunications infrastructure to bring climate information closer to farmers who may otherwise have limited access to formal forecasting services.
Data-driven agricultural insurance is another area where technology is being deployed to manage climate risks. In Nigeria, Pula, Leadway Assurance, Bayer Foundation and government partners are advancing agricultural insurance and climate-risk management initiatives intended to strengthen the resilience of about 250,000 smallholder farmers across eight states.
These initiatives link climate information to practical support that enables farmers to cope with the financial impact of extreme weather, including insurance coverage and access to essential agricultural resources.
For governments, the benefits of better climate intelligence extend beyond individual farms. More timely information could help public institutions identify vulnerable communities, anticipate agricultural disruptions and direct resources to areas most exposed to droughts, flooding and other climate-related risks.
But the effectiveness of these systems depends on whether information reaches farmers in a form they can understand and use. Language barriers, the cost of connectivity, limited digital literacy and shortages of agricultural extension services remain obstacles in Nigeria and other African markets.
This is where initiatives such as LINGUA Africa become relevant. Supported by Microsoft’s AI for Good Lab and partners, the initiative is advancing the development of artificial intelligence tools that reflect Africa’s linguistic diversity.
In Nigeria, projects such as AgroLingua Africa are exploring how AI can deliver agricultural information in local languages, including Hausa, Yoruba and Igbo. HausaNLP’s DialectBridge is also working to improve artificial intelligence capabilities across regional Hausa dialects.
Tadesse said the development of agricultural AI must take account of farmers’ knowledge and local conditions, rather than rely solely on technological predictions.
“The principle is simple: agricultural AI must listen before it advises,” he said.
The expert argued that farmers, researchers and agricultural extension workers possess knowledge and local context that technology would struggle to replace, adding that AI should strengthen their decision-making rather than override their experience.
The need to make climate information accessible is also linked to wider institutional challenges. Agricultural records are often fragmented, government agencies may lack specialised technical capacity, and institutions responsible for agriculture, meteorology and disaster management can struggle to share information.
Artificial intelligence cannot resolve these structural weaknesses on its own, making investment in data infrastructure, institutional cooperation and technical skills important to the success of climate adaptation programmes.
As African governments and development partners discuss climate financing at COP31, Tadesse said agricultural data systems should be treated as essential infrastructure for adapting to climate change.
He said adaptation financing should support the integration of agricultural surveys, interoperable databases, affordable connectivity and continuous training for government analysts and agricultural extension workers.
Responsible data-sharing and national ownership of these systems would also be necessary to ensure that they remain useful beyond individual development projects.
Tadesse cautioned that artificial intelligence should not be viewed as a substitute for the physical infrastructure and financial support required to protect agricultural production.
“AI alone cannot prevent droughts, replace irrigation or eliminate farmers’ financial pressures. It can, however, help farmers and governments anticipate risks and respond more effectively,” he said.
He added that the value of climate information ultimately depended on the decisions farmers and governments could make with it before adverse weather conditions arrived.

No comments yet. Start the conversation.