Tendai Chiguware and Binganidzo Muchara
Abstract
Smallholder farmers in Southern Africa heavily rely on agricultural extension services for critical knowledge. However, language and literacy barriers limit the effectiveness of conventional and digital advisory tools. This paper explores the institutionalisation of AI-powered multilingual chatbots on mobile messaging platforms as a scalable, inclusive solution for enhancing agricultural knowledge dissemination. Using data from the Malawi Fifth Integrated Household Survey (IHS5) and documented chatbot case studies (e.g., Ulangizi AI in Malawi and Darli AI by Farmerline), we find that while mobile phone access is widespread, actual use of digital advisory services remains low due to content being inaccessible in local languages or requiring high literacy. AI chatbots offer potential to deliver timely, relevant, and localized advice via voice and text in indigenous languages, helping overcome these barriers. Evidence shows increased farmer engagement, improved access to extension support, and stronger trust when services are institutionalised within community context and local languages. Nonetheless, issues around data privacy, infrastructure, content validation, and sustainability persist. We argue for institutionalizing these AI tools within formal extension systems, supported by government frameworks and partnerships, to ensure scale, credibility, and long-term impact. This approach can institutionalise access to agricultural knowledge and contribute meaningfully to rural development.