|
|
|
You can get e-magazine links on WhatsApp. Click here
|
|
|
|
|
|
Digital Green's AI farming advisory assistant crosses 10L users
|
|
Thursday, 06 August, 2026, 16 : 00 PM [IST]
|
|
Our Bureau, Mumbai
|
Digital Green India has announced that FarmerChat, the multilingual, AI-powered farming advisory assistant it deploys, has crossed 10 lakh users in India since its launch in October 2024, a milestone spanning farmers and frontline extension workers across the country. The announcement comes alongside the launch of FarmerChat 2.0, a significantly upgraded version built to make trusted, hyperlocal farming advice easier to access for millions more.
Available in five languages, FarmerChat offers timely, context-aware guidance across a farmer's decisions, from crop and pest management to livestock care and weather, even in low-connectivity environments.
Smallholder farmers face a range of challenges that make their incomes highly volatile from one year to the next. Among the biggest is access to timely, contextual advice: the knowledge exists, but reaching it at the moment a decision must be made is hard. To help farmers navigate these everyday choices, it built FarmerChat, an AI-powered app that provides tailored advice for on-farm decisions. A conversational interface lets farmers ask questions by voice, photo or text — accommodating both low literacy and the reality that many field problems are easier to photograph than to name.
FarmerChat answers field-level questions across the farming cycle: crop planning, pest and disease management, livestock care, weather forecasts, and inputs. Rather than waiting for a farmer to think of a question, it surfaces the right guidance at the right time, aligned to crop stages, seasons and livestock growth.
Powering the assistant are fine-tuned language models, combined with Retrieval-Augmented Generation (RAG) and expert-validated datasets, and refined through reinforcement learning from human feedback (RLHF). Crucially, its outputs are not left unchecked: local agronomists and veterinarians regularly validate the responses farmers receive, feeding corrections back into the models — so answers stay grounded in verified agronomy rather than generic AI output. The result is guidance that is technically accurate, locally relevant, and actionable in the field.
FarmerChat marks a fundamental shift in the economics of agricultural extension. Traditional, in-person advisory can cost Rs 3,300 per farmer each year. Its video-based model brought that down sharply, and FarmerChat has reduced it further to Rs 33 per farmer a year, making it possible to reach far more farmers with quality advice than conventional systems ever could.
In India, FarmerChat has already answered over 30 lakh queries, with women making up around 45 percent of its users. Farmers are not just using the tool — they are acting on it: around 60 percent active users act on the advice they receive, and 91 percent report greater confidence in their farming decisions, according to a third-party evaluation by 60 Decibels.
The rollout of FarmerChat's first version saw strong early adoption, with 67 percent of active users applying advice from the application, according to a third-party evaluation by IDinsight.
But farmers themselves pointed to the next step. Over the year, they moved away from typing their questions, choosing instead to speak to the assistant or photograph the problem in their field, and, notably, those who used the image feature proved significantly more likely to act on the advice. FarmerChat 2.0 locks these preferences into the design and moves the app away from waiting for a farmer to type a question. Suggested starter prompts — tailored to a farmer's crop, location and season — now surface the questions farmers should be asking based on where they are in the growing calendar.
Version 2.0 also introduces a simpler, clearer interface and advice that adapts to each farmer, through new advisory cards and an evolving farmer profile shaped by their crop, location, season and past questions.
Building on this momentum, the company aims to add another 35 lakh farmers to the platform by 2028 — deepening women farmers' participation, driving adoption of climate-smart practices, and improving household incomes.
Nidhi Bhasin, CEO, Digital Green India, said, "Everything in FarmerChat 2.0 came from listening to farmers. Through millions of interactions, they shared that they wanted advice suited to their crop, their season and their language, delivered the way they naturally communicate. The assistant was rebuilt around exactly that. But it is still evolving. The next step is the Agricultural AI Harness, an ecosystem foundation where trusted services for advisory, weather, markets and schemes plug in, powered by reasoning and memory that knows each farmer's context. One farmer, one trusted place, every service they need. And benchmarking quality rigorously, combining human-led feedback with automated evaluation, will continue so the advice a farmer receives only grows more accurate, relevant and trustworthy."
|
|
|
|
|
|
|
|
|
|