Elia Gandolfi, Sasha Tanghe, The Agency Fund

Topic: An AI Evaluation Framework for the Development Sector

AI is rapidly becoming a core tool in global development, but many nonprofits, donors, and policymakers struggle to evaluate which AI interventions truly deliver impact at scale while avoiding harm. This session introduces a practical AI Evaluation Framework developed through the AI Accelerator program to address this gap. We’ll present the framework, share real-world use cases from nonprofit implementers, and explore how organizations are rigorously evaluating AI projects across four key phases—from early pilots to scaled solutions. An interactive discussion will give attendees a chance to ask questions, critique the framework, and shape its next iteration. Grant-makers, program and evaluation leads, product teams, and policymakers will leave with a clear, actionable structure to assess AI projects, along with concrete benchmarks and examples they can apply immediately.