Generative AI in the Real World: Putting AI in the Hands of Farmers with Rikin Gandhi

Why It Matters

By delivering real‑time, locally relevant agronomic guidance through ubiquitous mobile channels, AI empowers smallholders to increase productivity, mitigate climate risks, and participate in modern value chains.

Key Takeaways

• AI-powered Farmer.Chat delivers localized video, weather, market advice via mobile.
• Multilingual voice, text, image inputs enable farmers without smartphones to interact.
• Retrieval‑augmented generation reduces hallucinations from 30% to ~1% using reranking.
• Dynamic content merges static videos with real‑time data for context‑specific guidance.
• Knowledge‑graph and routing architecture streamlines simple queries versus complex decisions.

Summary

The podcast features Ben Laura and Rikin Gandhi, CEO of Digital Green, discussing Farmer.Chat – a generative‑AI platform that puts localized agricultural knowledge into the hands of smallholder farmers across India, Ethiopia, Kenya and Nigeria.

Digital Green leverages a library of over 10,000 farmer‑produced videos in 40 languages, combined with real‑time weather, market data and decision‑support models for fertilizer, irrigation and dairy feed. The system accepts voice, text, image and even IVR inputs, translates them to English, queries a vector‑search knowledge base and returns answers in the farmer’s native language.

During the pandemic the organization shifted from offline projector screenings to a deterministic bot, then to GPT‑3/4‑driven retrieval‑augmented generation. By applying reranking and filtering they cut hallucinations from >30 % to ~1 %. They have built a 300,000‑question benchmark, manually reviewed 10 %, and are integrating knowledge graphs and LLM routers to route simple weather queries directly to APIs.

The move to on‑demand, AI‑personalized advice enables farmers to diagnose pests, plan sowing or harvest dates, and optimize inputs, boosting yields while reducing emissions. Scaling to six million users demonstrates how low‑bandwidth, multimodal AI can accelerate climate‑resilient agriculture in the Global South.

Original Description

Rikin Gandhi, CTO of Digital Green, talks with Ben about using generative AI to help farmers in developing countries become more productive. Farmer.Chat integrates information from training videos, sources of weather and crop information, and other data sources in a multimodal app that farmers can use in real time.

Follow O'Reilly on:

LinkedIn: https://www.linkedin.com/company/oreilly/

Facebook: http://facebook.com/OReilly

Instagram: https://www.instagram.com/oreillymedia

BlueSky: https://bsky.app/profile/oreilly.bsky.social