For global health and development teams, generating data and evidence is only half the work: developing a reliable way to surface and actively use data is what drives impact. The Women’s Economic Empowerment (WEE) team at a large philanthropic foundation was continuously developing strong research and learning materials and had a rich set of available resources. However, they struggled with access and use when available knowledge management systems were insufficient or static. As leadership was placing increasing demand to ground investments and communications in data, team members needed to know the right person to find the appropriate source with little visibility into what was available or how to make best use of it. While an updated centralized knowledge system was an obvious fix, the main goal was to build a system that could change how the team used data and evidence day-to-day rather than just knowing where the files lived.
Kinaura Partners led the design, build, and roll out of an AI-enabled learning portal, completely tailored to meet the team’s specific needs and priorities. Using an agile, phased approach, the team moved from a minimum viable product through several rounds of development by incorporating user testing and feedback from targeted use cases at each stage. The system was fit-for-purpose, with features that included a document taxonomy and quality-tiering framework that made evidence classification practical for non-specialists, distilling existing standards into five clear criteria for identifying the most rigorous research in the portal. AI capabilities were integrated throughout: document tagging, synthesis, and a chatbot trained on the full repository. All technical decisions went through the foundation’s internal AI governance, websites, and knowledge management teams, to ensure the system would be compliant and sustainable long after launch.
Staff reported meaningful efficiency gains and survey results showed a strong willingness to both use the learning portal on a regular basis as well as help maintain it by uploading new data. This outcome came from both process and design: by collecting feedback and sharing how it informed the portal design, staff participated in change management as part of the build phase. The portal gave the WEE team a concrete way to respond to leadership’s request to use data and evidence across their decision making. The learning portal was intended to be simple, effective, and iterative. Instead of a checkbox exercise, it produced a system that ultimately enabled the team to support data translation, integration, and uptake across all of their work.
