Making evidence actually usable | Lindsey Moore

This episode of In Pursuit of Development explores how AI is reshaping the way development organizations learn from evidence, unlocking lessons buried in evaluations and reports, and helping practitioners make better decisions in complex, fast-moving settings. Dan Banik speaks with Lindsey Moore, CEO and Founder of DevelopMetrics, about how ethical AI and predictive analytics can make development evidence genuinely usable — turning decades of evaluations into structured, searchable insight for better decisions.Lindsey draws on her experience in USAID and her work building domain-trained models to explain why the sector’s challenge is not an evidence shortage, but rather an evidence usability gap. Together Lindsey and Dan discuss what it takes to build context-aware systems: transparent taxonomies, careful human labeling, and models grounded in local perspectives rather than default assumptions embedded in general-purpose AI.The conversation also explores how large-scale evaluation archives can be transformed into institutional memory, strengthening professional judgment and helping organizations learn faster, reduce waste, and target interventions more precisely.In this episode:AI for global development beyond hype: What actually works in practice.Why definitions and taxonomies shape results (and power).How to reduce bias and improve context in development AI.Evidence infrastructure, knowledge management, and decision workflows.Resources:When USAID Shut Down, Its Lessons Nearly Vanished. AI Helped Recover Them (Stanford Social Innovation Review, December 2025)Integrating human-centered AI for land use policy: Insights from agricultural interventions in international development (Land Use Policy, 2025) Host:Professor Dan Banik, Centre for Global Sustainability, University of OsloSubscribe:Apple Spotify YouTubehttps://globaldevpod.substack.com/