About IFAD
At the International Fund for Agricultural Development (IFAD) we invest in rural people, empowering them to increase their food security, improve the nutrition of their families and increase their incomes. We help them build resilience, expand their businesses and take charge of their own development.Activities by IFAD
Mapping food systems national pathways and summit dialogues with AI and ML techniques
Global Not Applicable
The analysis utilizes a supervised learning approach to classify sentences within the national pathways and summit dialogues documents. This method allows for the identification of prevalent topics across various dimensions, including food system drivers, components, and outcomes. The taxonomy for classification is derived from the HLPE Food System Framework, ensuring a comprehensive and relevant analytical lens. This analysis was carried out on specific sub-regions to primarily test the potential of this technology for policy learning. The main aim was to uncover insights from publicly available national pathways for food system agenda achievement through artificial intelligence (AI) and machine learning (ML) techniques. Several key lessons were learned. First and foremost, the project underscored the importance of leveraging advanced technologies like AI and ML to dissect and understand complex programmatic documents across diverse socio-economic contexts. The use of supervised learning to classify statements within the national pathways illuminated the distinct priorities and actions of different countries, offering nuanced insights into their food system agendas. A critical takeaway was the recognition of the heterogeneity in food system priorities across the specific sub-regions analysed and countries’ income classifications. This diversity necessitates tailored approaches in policy formulation and implementation to address the unique challenges and opportunities within each context. Furthermore, the project highlighted the significant potential of AI and ML in enhancing policy analysis and decision-making processes. By applying the HLPE Food System Framework for classification, the project demonstrated how structured taxonomies could provide a coherent framework for analysing and interpreting complex data.
Lastly, the analysis revealed the challenges and limitations of data availability and quality, particularly in countries where these documents are not available. This gap underscores the need for continued efforts to promote data generation and sharing across countries to support more comprehensive and informed policy analyses.
Harvest AI Portal
Global Not ApplicableHarvest is IFAD’s internal AI portal with IFAD-specific modules for core operations. Harvest brings custom AI solutions to IFAD for the IFAD context, designed and developed with subject matter experts across the organization. Harvest modules, powered by AI and related technologies, combine advanced tools into simple, intuitive, and user-friendly interfaces that address specific workflows. Current modules include: SECAP-Scan, Mainstreaming-Scan, Youth-Scan, Quality-Scan, Insta-Points, Insta-PR, Insta-FA, Insta-Compare, File-Chat, Content-Edit, POP-Chat.
Activity Type AI Tools/SolutionsInfrastructure/Systems DevelopmentBlossom AI Portal
Global Not Applicable
Blossom is IFAD’s internal AI portal with general-purpose modules
for simple daily tasks. With Blossom, IFAD staff have access to a wide range of common AI technologies through a single portal. Blossom modules, powered by AI and related technologies, are designed to make advanced tools easy to use in day-to-day work with simple, intuitive, and user-friendly interfaces. Current modules include: Text Generation, Text Reasoning, Icon Generation, Summarisation, Text Translation, Document Translation, Transcription, Word Cloud, Table Extraction, OCR Conversion.
Sustainability Action Plan Repository and Chatbot
Global Not Applicable
IFAD’s Sustainability Action Plan aims to provide operational and technical actions to strengthen IFAD’s performance related to sustainability of benefits at the project level through an approach that commonly supports IFAD’s portfolio. Due to the diversity of IFAD projects and the numerous factors that may impact project sustainability, the action plan does not aim to address each one individually, but focuses on providing the necessary building blocks to support widespread knowledge, behaviors, and capacity to improve project sustainability. As interventions to enhance sustainability of project benefits are context specific, a key output of the action plan is an AI-enhanced repository on Sustainability. The repository offers a rich, interactive, and visual interface for country teams to explore IFAD’s resources on sustainability of benefits. Users are able to extract information from curated publications on sustainability, according to the criteria and dimensions detailed in the action plan. The repository allows country teams to develop a tailored approach to increasing sustainability and developing sound exit strategies for better ownership in a wide variety of contexts.
The repository consists of how-to-do notes, analysis, and useful resources developed by IFAD, as well as tools and guidance from external sources. These resources are processed with text-centric AI and automation tools that extract key concepts and generate focused summaries and data visualizations. The repository is also connected to a Generative AI (GenAI) chatbot interface, which allows users to ask any questions about sustainability. The chatbot searches through the repository contents and provides a referenced answer to the question using Large Language Models (LLMs). This solution is implemented as a Retrieval Augmented Generation (RAG) search engine with a natural language interface