Post by Dylan Anderson-Berens

A few years ago, I worked on research at Tufts asking a simple question: Which digital tools actually work for smallholder farmers—and why do so many fall short? That work didn’t just stay academic. It pulled me into the space—working with platforms like eKutir and BovControl, and eventually helping launch an early-stage venture (GleTech) to try to apply what we were learning in practice. We built, tested, iterated… and ultimately stepped away. But the questions never really went away. I’ve been revisiting that work recently—especially as the conversation shifts toward AI, digital advisory, and data-driven agriculture. Short version: The tools have evolved. The core challenges haven’t. I wrote a short piece reflecting on: 1) what we got right (and wrong) 2) what trying to build revealed and 3) where I still see risk—especially around overhyped impact claims. Would genuinely value perspectives from others working in this space. 👇 Article below