Digital Green has designed FarmerChat, an AI chatbot tailored to assist to farmers and extension workers. FarmerChat has nearly 600,000 users across 120 countries, where the tool works in 14 languages. FarmerChat provides localized solutions to a wide variety of crop and poultry farmers through voice, text and image -based searches. The app also provides weather updates according to specific regions and guides farmers on how to protect their farms from climatic fluctuations to build their resilience .
60 Decibels partnered with Digital Green to understand 450 Kenyan farmers’ experience with the chatbot, and its impact on their agricultural practices.
83% of users report engaging with services like FarmerChat for the first time. 87% of farmers report no good alternatives to FarmerChat, positioning it as a unique digital offering in the market. Among those who do have alternatives, over two-thirds believe that FarmerChat provides a significantly better service compared to other options.
99% say FarmerChat’s recommendations are trustworthy. Most say the information is easy to understand, and provides complete answers to their questions. 7 in 10 farmers report applying FarmerChat recommendations in the last 30 days, and those who haven’t applied plan to apply them in the future.
Meaningful use is a value-based approach to assess whether farmers are engaging with a digital service to the extent that it leads to real changes in how they manage their farm. 61% of farmers are using FarmerChat’s services meaningfully which is higher than our Kenya benchmark of 27% (based on 60dB’s 2025 Meaningful Use Study, read more here).
Female farmers are more likely to recommend FarmerChat than their male counterparts (NPS 75 vs. 57). This is likely driven by female farmers reporting greater trust in the information from FarmerChat. They are also likely to face fewer challenges than male farmers (93% vs. 88%). Female farmers are more likely to report that FarmerChat met ‘all’ of their farming needs (51% vs. 37%) and are also more likely to be meaningful users (72% vs. 58%).
10% of farmers report experiencing a challenge with FarmerChat, mostly related to technical issues like slow performance, updates, or image recognition. Some farmers noted that the responses were too complex or lengthy to understand.
To gain deeper insight into the relationship between the chatbot’s recommendations and farmer behavior, we cross referenced backend data from Digital Green containing farmers’ original queries and the AI recommendations.
60dB used Digital Green’s backend data on the last three queries from farmers and FarmerChat’s responses to those specific queries. Farmers were asked if they recalled any of the three themes and, if so, what actions they had taken on the farm based on FarmerChat’s advice ( Application ).
To understand the link between farmers’ self-reported actions and the recommendation from FarmerChat (Digital Green backend data), 60dB analyzed the extent of qualitative overlap ( Attribution ) between the two.
Researchers only had access to the farmer’s query, not the response, to minimize bias. Farmers were asked if they recalled the query and any actions taken, without being prompted with responses.
Researchers only had access to the farmer’s query, not the response, to minimize bias. Farmers were asked if they recalled the query and any actions taken, without being prompted with responses.
96% recalled at least one of their 3 most recent queries to FarmerChat. 66% reported taking an action based on FarmerChat’s response to their query. Among those who took an action, 80% described an action that partially or fully aligned with what FarmerChat had recommended.
Check out some examples below!
