Previous Next

Project Overview

So, what’s the problem

Agricultural development projects generate vast amounts of data, yet much of this data remains underused. Collecting and managing data requires significant time and investment, but when data is not FAIR the value is reduced.

Datasets are often difficult to locate, access or combine. This leads to duplication of effort, fragmented evidence and missed opportunities for learning, innovation and better decision-making. Although frameworks and guidance exist, they are not consistently applied in everyday programme workflows, limiting the reuse of data for analytics, modelling and emerging AI applications.

At the same time, many countries are developing agricultural data strategies and digital policies. However, turning these into practical governance arrangements, clear institutional roles and trusted data-sharing mechanisms remains a challenge. Concerns around data ownership, incentives, privacy, risk and value-sharing further restrict access and reuse.

These issues reduce the return on investment from data-rich programmes and slow progress towards more coordinated, evidence-based agricultural systems. Addressing them requires more than new platforms or technologies. It depends on stronger governance, aligned policies, practical tools and targeted support that help partners apply FAIR and responsible practices in real programme settings.

A systems-thinking, people-centred approach is needed to build interoperable, trusted data ecosystems that enable data to be shared responsibly and used to improve agricultural outcomes.

What is this project doing?

Through this project, CABI is strengthening responsible data practices across agricultural development by improving the quality of FAIR implementation and supporting Gates Foundation grantees/partners to apply best practice in real programme settings. This work is progressed by:

1.) Making FAIR measurable and actionable
Rather than providing a single score, the project shows what is missing from a dataset, why it matters and what to improve first. It also develops simple, domain-specific FAIR “recipes” that give researchers clear, step-by-step guidance on what data to collect, how to organize it and how to share it so it can be reused. More information can be found in the Fair Process Framework.

2.) Strengthening trusted data access
The work supports partners and governments to agree clear rules for data access and sharing. This includes defining who can use data, under what conditions, and how risks such as privacy and misuse are managed. It also helps turn national data strategies into practical actions and institutional roles.

3.) Clarifying what “AI-ready” data means
The project defines minimum requirements for AI-ready datasets, clarifies the boundary between FAIR and AI readiness, and prevents unnecessary over-engineering while ensuring data are suitable for modelling and advanced analytics.

4.) Embedding governance into funding decisions
The project explains what extra steps – if any – are needed for data to be suitable for modelling and AI, and where FAIR data are already sufficient. This helps teams avoid unnecessary work while ensuring data quality.

5.) Providing responsive technical support
This includes the delivery of expert reviews, policy analysis, FAIR-readiness assessments, standards and repository recommendations to address emerging challenges across foundation investments. These resources can be found on the Gates Open Research platform.

This project is a continuation from a previous funding phase.