Projects & Engagements

Kilimo Hakika served as a live innovation testbed for CIMMYT, running head-to-head field trials of generative-AI farm advisory tools, pitting its own WhatsApp-based AgroTutor against FarmerChat, Gemini, and ChatGPT under real farming conditions. The team gathered direct feedback from more than 100 farmers and synthesized the results to guide the next stage of tool development.

Under the UK-funded Strengthening Digital Communities Project, Kilimo Hakika designed and delivered a focused digital agriculture training program for Community Digital Champions (CDCs) and agriculture officers. The program embedded validated digital agriculture tools into participants' existing knowledge base, strengthening the guidance they give farmers on value-chain decisions.

Kilimo Hakika led detailed data collection on crop and livestock practices across three counties to inform refinement of ILRI's Intelligent Agricultural Systems Advisory Tool (iSAT). The engagement combined household-level data collection with short stakeholder surveys of crop and livestock officers, grounding the platform in current, verified farming practice.

In partnership with UKDAP, ACWICT, ACTIP, and the Busia Women Cross-Border Traders' Cooperative, Kilimo Hakika trained Community Digital Entrepreneurs across five districts in Eastern Uganda, equipping them with practical skills to use digital tools and platforms to strengthen agricultural value chains.

EDACR was Kilimo Hakika's flagship digital-skilling program, delivered with KICTANet and funded by the UK FCDO through the British High Commission, in partnership with the Otermans Institute. The project mobilized, trained, and supported Community Digital Champions across eight arid and semi-arid (ASAL) counties, transitioning them into Community Digital Entrepreneurs through a Hybrid Human–AI skilling model spanning Digital Agriculture, IT Technician, Gig Economy, and Creative Economy tracks, with a deliberate focus on reaching women and persons with disabilities.

Kilimo Hakika paired an offline AI tutor with weekly facilitator-run device clinics, proactive check-ins, and 48-hour outreach whenever a learner missed a login. The support model was built specifically to catch and re-engage learners before they dropped out silently.

Kilimo Hakika partnered with CAB International to design and build an AI-based Soil Fertility Prediction and Advisory System (Soil-FPAS), integrating ground-truth soil data, IoT sensor streams, weather-station outputs, drone and satellite imagery, and laboratory soil reports into a single platform. The system produces real-time predictive models of soil fertility and delivers location-specific, actionable recommendations to farmers through mobile platforms.

Commissioned by FCDO through the Evidence Fund and delivered with CABI, this mixed-methods research study examines how artificial intelligence can support climate-smart agriculture and food security in Kenya. It investigates four questions: which agricultural datasets are missing or inaccessible and how that constrains AI tools; how usable and impactful existing AI advisory tools are for farmers; the institutional, economic, infrastructural, regulatory, and social barriers to scaling, with particular attention to women, youth, and marginalized groups; and what data-sharing and governance models can support AI innovation while protecting farmers' rights.

Findings will directly inform Kenya's Ministry of Information, Communications and the Digital Economy (MoICDE), the Ministry of Agriculture and Livestock Development (MoALD), and FCDO investment decisions.

Kilimo Hakika is building an East African Livestock Data Corpus to power ILRI's Digital Transformation Accelerator and Inception 42's (In)Livestock AI assistant for smallholder livestock keepers, an agentic AI tool spanning animal health, nutrition and ration, climate and environment, and knowledge and policy. The work covers surveying and inventorying existing Kenyan livestock datasets, securing legal and ethical clearance for reuse under the Data Protection Act (2019) and NACOSTI permits, closing priority data gaps through targeted field collection, and preparing an annotation-ready corpus. Kilimo Hakika will also support an expert annotation workshop in Kenya by the end of 2026.

Over a four-month consultancy, Kilimo Hakika laid the technical and content groundwork for the AgroTutor GenAI platform: building a bilingual (Swahili/English) agricultural knowledge corpus, drafting digitization-ready course materials, and running farmer-centered testing of the chatbot. The team analyzed and integrated farmer feedback, expanded the corpus to close knowledge gaps surfaced by real farmer questions, supported digitization of the AgroTutor Academy curriculum, recruited and supported over 100 farmers to test the finished courses, and iteratively tested the GenAI model after each revision.