Kenya AI-Driven Agricultural Advisory Platform: Forerunner Phase

MCAAI as national technical reference partner for voice and local-language AI readiness in agricultural advisory

Ongoing · started 2025

Overview

Background and Rationale Kenya is embarking on the design of a national AI-driven agricultural advisory platform, intended to function as a digital public good that enables multiple public, private, and civil-society actors to deliver safe, high-quality, and affordable advisory services to small-scale producers. A central barrier to adoption of existing digital advisory services is language and literacy. While Kenya has strong digital penetration, many farmers, particularly women, elderly farmers, and those in remote regions, are excluded by text-heavy, English-first, or technically phrased advisory systems. The forerunner phase of the Kenya AI Advisory Platform therefore seeks to build foundational public goods and technical readiness for inclusive, multilingual, and voice-enabled advisory, without creating new standalone delivery platforms. This 18-month phase pilots the foundations required for the National AI-enabled Advisory Service. Within this context, Maseno University's Centre for Applied Artificial Intelligence (MCAAI) is uniquely positioned to contribute deep expertise in African language NLP, corpus development, and applied AI research, complementing SAFIC's role as a neutral platform orchestrator. Purpose and Strategic Role of MCAAI The collaboration positions MCAAI as a technical capability partner responsible for advancing local-language and voice AI readiness within the Kenya AI Advisory Platform, ensuring that AI advisory services are inclusive, culturally intelligible, and usable by low-literacy users. The collaboration focuses on prototyping and validating voice and local-language AI building blocks, collaborating on the development of benchmark datasets and evaluation assets for Kenyan languages, and integrating voice capabilities with the National AI advisory service. MCAAI's Core Goal Within the overall programme architecture, MCAAI acts as the national technical reference partner for voice and local-language AI readiness, supporting SAFIC and delivery partners with prototypes, datasets, benchmarks, and integration guidance. This role ensures that language inclusion is embedded early in platform design, voice interfaces are treated as a core accessibility layer and not an afterthought, and Kenya builds sovereign capacity in agricultural language AI without duplicating delivery systems. Scope of Proposed Work 4.1 Local-Language Prioritisation and Readiness. MCAAI will support the platform by developing a methodologically sound language prioritisation approach for Phase I, including a new methodology for language prioritisation. This includes mapping the agricultural relevance of languages by region and crop systems; recommending a small, representative set of pilot languages (English and Kiswahili as the national access layer, plus selected Kenyan languages chosen to ensure diversity and inclusion); and advising on dialect sensitivity and code-switching patterns common in farmer communication. The output is a language prioritisation and readiness note to guide platform pilots and future scale-up. The note will cover the rationale for selecting each language; prioritisation for the first phase and for the long-term phase; the effort and resources needed to collect data for languages that are a priority for agriculture but are low-resource; and MCAAI's language mapping in Kenya compared against existing language resources. Figure 1 in the source document shows the proposed ecosystem and integration architecture for the Kenya AI-enabled advisory services.

Project logic

  1. Aim

    To position MCAAI as the national technical reference partner for voice and local-language AI readiness, so that AI advisory services on the Kenya AI Advisory Platform are inclusive, culturally intelligible, and usable by low-literacy small-scale producers. This ensures that language inclusion is embedded early in platform design, voice interfaces serve as a core accessibility layer, and Kenya builds sovereign capacity in agricultural language AI without duplicating delivery systems.

  2. Context

    Kenya is designing a national AI-driven agricultural advisory platform as a digital public good, enabling public, private, and civil-society actors to deliver safe, high-quality, and affordable advisory services to small-scale producers. Despite strong digital penetration, many farmers, particularly women, elderly farmers, and those in remote regions, are excluded by text-heavy, English-first, or technically phrased advisory systems. Language and literacy are the central barrier to adoption. The 18-month forerunner phase builds foundational public goods and technical readiness for inclusive, multilingual, voice-enabled advisory, without creating new standalone delivery platforms. SAFIC acts as the neutral platform orchestrator, while MCAAI contributes expertise in African language NLP, corpus development, and applied AI research.

  3. Tasks

    Prototype and validate voice and local-language AI building blocks Collaborate on the development of benchmark datasets and evaluation assets for Kenyan languages Integrate voice capabilities with the National AI advisory service Develop a methodologically sound language prioritisation approach for Phase I, including a new language prioritisation methodology Map the agricultural relevance of languages by region and crop systems Map languages in Kenya against existing language resources to identify gaps Recommend a small, representative pilot language set: English and Kiswahili as the national access layer, plus selected Kenyan languages chosen for diversity and inclusion Advise on dialect sensitivity and code-switching patterns common in farmer communication Document the rationale for language selection, with prioritisation for the first phase and the long-term phase Estimate the effort and resources needed to collect data for agriculturally important but low-resource languages Deliver a language prioritisation and readiness note to guide platform pilots and future scale-up Support SAFIC and delivery partners with prototypes, datasets, benchmarks, and integration guidance

Focus areas

  • Agriculture
  • Language Technology

Subscribe to our Newsletter