
0
Bring research and evidence into classroom products
April 20, 2026
AI Analysis: Bandwagon
Posted 2 hours ago by
How do you build products that work? We have decades of accumulated science of learning research, but it can be hard to get that research into the hands of classroom teachers. I met with Sandra Liu Huang, Learning Commons’ president, to discuss building the infrastructure to bring learning science into product development and empower educators with better tools.

We talked about making research more usable for developers and educators, why shared infrastructure matters, and how we can ensure learning science actually reaches classrooms. Auditi: Something I have long been fascinated by is the gap between established learning science and what reaches teachers and students through classroom products. What are the biggest challenges in translating research into classroom tools? Sandra: Let me start with the positive. We actually know a great deal about how learning happens—about the conditions needed for optimal learning and the instructional strategies that work best. The challenge is translating research into tools and materials teachers can use every day. Much of the research lives in journals and is often incremental, meaning you have to synthesize findings across decades of studies. So we’re asking teachers to do the impossible: continuously review academic literature and determine how to integrate it into their lesson plans, while tailoring those lesson plans in real time for every student. Educators need better resources grounded in learning science, with the flexibility to adapt to each student’s needs. Auditi: That resonates. At AERDF, we focus on how research informs the development of new solutions. It’s not just about generating new knowledge—it’s about making that knowledge usable. How can we bring more evidence into product development? Sandra: The education field has an opportunity to build on years of work to advance learning science and translate research into practice. However, that process can be difficult. What’s different now is that new technologies, including AI, create opportunities to help educators synthesize research and apply it more coherently for classroom needs. But that only works if AI systems draw upon high-quality data. Tools need to be connected to curriculum, academic standards, and learning science in ways that reflect how students actually learn. That’s why the field needs shared infrastructure that creates a baseline for quality. AI isn’t a panacea, but it can be a powerful lever if it reflects the best of learning science. Auditi: What you’re describing—building shared infrastructure rather than proprietary solutions—feels like a meaningful shift. Traditionally, philanthropy funds programs with clear outcomes and timelines. Infrastructure work is different. It’s slower, shared, and its impact spreads across the field. Why is that work worth doing? Sandra: Combining grants, partnerships, and technology can help the education sector shape how tools develop. By working with experts in learning science and classroom practice, we can translate their knowledge into useful developer resources that improve the whole sector. That allows their work to reach far beyond individual research projects. Ultimately, the goal is to ensure all students have access to rigorous, motivating instruction. Auditi: Organizations like ours are generating deep research about how students learn. But generating research alone isn’t enough. What’s exciting about partnerships like the one between Learning Commons and Magpie Literacy, a nonprofit reading program we’ve supported, is that they help translate insights into shared infrastructure, like Knowledge Graph. That kind of work extends impact beyond one organization’s products to strengthen the whole field. It’s the difference between building one tool and laying a foundation. What does it take to make research frameworks usable for developers? Sandra: Our latest round of partnerships is focused on expanding math, science, and literacy datasets that connect academic standards, curriculum, and learning science. Many edtech systems rely on data that isn’t granular enough or structured in ways machines can interpret. Step one is breaking academic standards into the smaller skills students need to learn. Then we connect those skills to curriculum and research. That structure helps AI systems understand how concepts relate to one another, and how learning progresses over time. Think of it as creating the knowledge base that allows technology to reason about learning. We’re excited about the Magpie Literacy partnership because its platform encodes core reading skills—like phonemic awareness, decoding, and fluency—and maps relationships between them. By incorporating those insights into shared infrastructure, the entire field can benefit from that work. Auditi: Incredible. That kind of leverage can help shift the entire ecosystem. What advice would you give an edtech developer that wants to build products that truly support learning? Sandra: Start by connecting your work to the existing infrastructure. Shared datasets and evaluation tools can help developers ground their products in learning science from the start. We welcome feedback and feature requests as we continue to map out roadmaps that can unlock chronic challenges for the field in getting to better, more effective tools. Auditi: I’d also add to your advice: Start with the research and focus on learning impact, not just product-market fit. And involve educators early in the RD process. Sandra: Yes, definitely; we collaborate early and often with educators to shape our products. Auditi: Looking ahead, what will success for the field look like in three years? Sandra: Success would mean we’re aligned around building high-quality tools grounded in learning science and designed to meet real teachers’ needs. Ideally, it would also mean a different kind of edtech marketplace—where tools work together, align with academic standards, and reflect strong research. Educators need to be confident that the technology they choose will support learning. FINAL WORDS Advancing learning science is essential, but research isn’t enough. We need infrastructure that allows insights to move beyond journals and into the tools educators use daily. When research, infrastructure, and product development come together, we have a real opportunity to reshape education innovation—and ensure tools reaching classrooms are grounded in how students learn best. Auditi Chakravarty is CEO of the Advanced Education Research and Development Fund
Reliability Insights
P
Technique: Bandwagon
Our AI detected use of specific narrative techniques in this piece.Fast Company
Coverage and analysis from United States of America. All insights are generated by our AI narrative analysis engine.