Farmers in Tanzania are turning to artificial intelligence (AI) to diagnose crop diseases quickly and inexpensively, thanks to a locally developed mobile application that experts say could significantly reduce harvest losses and improve rural livelihoods.
The KilimoAI app, created by computer scientist Dr. Neema Mduma and her colleague Hudson Laizer, uses machine-learning technology to analyse photos of affected leaves and immediately identify likely diseases.
The tool is freely available on the Google Play Store and works on basic Android smartphones – devices already in widespread use by farmers across the country.
Mduma, a lecturer at the Nelson Mandela African Institution of Science and Technology (NM-AIST), said the idea for KilimoAI came from witnessing repeated crop losses among smallholders who lacked access to quick and accurate disease diagnosis.
“The main problem we are addressing is the lack of quick, reliable, and affordable advice regarding disease diagnosis,” she said. “We want farmers to understand what is affecting their crops early and to take the right action before it’s too late.”
Farmers use the app by photographing a diseased leaf and uploading the image for instant analysis. The underlying AI model, trained on thousands of images of both healthy and diseased crops, evaluates visual clues such as leaf coloration and spots to predict the most likely ailment and suggest next steps.
“The farmer only needs to open the app, take a photo of the leaf, and upload it,” Mduma explained, noting that the user interface features simple icons and Swahili language support to make it accessible even to those with limited digital experience.
The tool was developed with input from farmers and agricultural specialists, and works in collaboration with the Tanzania Agricultural Research Institute (TARI) to ensure disease management guidance aligns with approved practices.
“The app focuses on approved pesticides and proper usage instructions while also promoting integrated pest management and non-chemical options where possible,” Mduma added.
Pilot deployments in regions such as Arusha and the Southern Highlands have shown promising uptake. By mid-2025, KilimoAI had reportedly reached tens of thousands of users, including smallholder farmers and agricultural extension officers – with plans to expand to 400,000 farmers by 2030.
In northern Tanzania’s Seela Sing’isi ward, 53-year-old farmer Roland Daniel Sarikikya said the app has altered how he deals with crop problems. Before, he would buy pesticides on impulse, often without knowing the true cause of a disease.
“Previously, I just used pesticides randomly, but this app makes my work easier,” Sarikikya told The Xylom, describing how KilimoAI identified maize streak virus – a serious viral disease – from a leaf photo.
Many farmers have echoed similar sentiments, noting that prompt identification has helped them avoid unnecessary chemicals and better protect their plants. As a result, some have seen improvements in yields, potentially boosting income and household food security.
Mduma and her team continue to refine KilimoAI’s algorithms and expand its disease-detection capabilities. Already, models for diseases such as black sigatoka in bananas, Fusarium wilt in soil-borne crops and late blight in potatoes have been developed, with plans to integrate these into the app’s diagnostic suite.
As Tanzania’s agricultural tech ecosystem evolves, tools like KilimoAI illustrate how locally engineered AI solutions can address long-standing gaps in agricultural extension services — helping farmers make more informed decisions and potentially transform crop health management practices in rural communities.
Credit: Farmersreviewafrica
