ILRI launches AI-powered chatbot to give fast, evidence-based animal health guidance
The new digital tool links farmers and animal health workers directly to peer-reviewed livestock research through simple questions.
KENYA – A new AI-powered chatbot from the International Livestock Research Institute aims to close the gap between livestock science and day-to-day animal health decisions.
Developed by ILRI’s Data and Research Methods Unit and the CGIAR Digital Transformation Accelerator, the chatbot gives instant answers to animal health questions using peer-reviewed research. Kenyan tech startup Fahamu AI worked with the teams on the system, which is now live at animalanswers.ilri.org.
ILRI designed the tool to help farmers, extension officers, veterinarians, and policymakers who often need quick guidance but lack time or access to scientific papers. Livestock diseases continue to affect productivity and incomes across Africa, especially among camel, cattle, sheep, and goat keepers in low-resource settings.
The chatbot draws from CGIAR’s open-access research archive in CGSpace. Users type questions in plain language and receive responses grounded in published studies. Each answer includes clickable citations that link directly to the source material, allowing users to check the evidence themselves.
According to Jean-Baka Domelevo Entfellner, Head of Data and Research Methods at ILRI, the goal focuses on practical use. “To make our research outputs truly actionable by letting anyone ask questions in plain language and receive reliable, sourced answers instantly,” he said. Alan Orth, a systems specialist in the same unit, provided technical support during development.
How the tool works
The chatbot uses a simple interface. Users enter a question and click “Ask,” after which the system generates a response supported by excerpts from scientific publications. The platform shows where each answer comes from, giving users clear traceability.
A feedback feature allows users to rate answers based on accuracy, completeness, and usefulness. This input helps the development team track performance and improve reliability. The chatbot also stores question history, enabling users to review earlier interactions and support ongoing learning.
Early testing and feedback
ILRI rolled out the chatbot in an early pilot phase and continues to refine it using user feedback. On 9 December 2025, researchers, developers, and potential users gathered at the ILRI campus in Nairobi for a hands-on testing workshop.
Participants welcomed the easier access to scientific knowledge but pointed out areas that need work. Some cited sources appeared only as “document excerpt,” which made it harder to confirm full references. Others noted that the system sometimes ranked loosely related sources too highly. A few off-topic questions also produced unclear text.
The team has already started addressing these issues. Users suggested replacing the fixed limit of five source excerpts per answer with a relevance threshold that highlights only the most useful material. Others asked for personal logins, follow-up question threads, and selectable user profiles such as farmer or researcher to adjust tone and depth.
Next steps for digital livestock research
For ILRI, the chatbot represents a first step toward a wider set of digital tools covering livestock genetics, nutrition, and climate adaptation. “We’re starting small, but the vision is big,” Domelevo Entfellner said.
Planned updates include giving users control over answer length, testing additional generative AI systems to improve speed and accuracy, and strengthening back-end performance through GPU-based hosting. ILRI has invited the public to use the chatbot, review the sources, and share feedback to guide its next phase.
Be the first to leave a comment