Ask, and they receive

Story & Pics: Sowmya Rajaram | Editor: Sowmya Rajaram

Farmers at the Bellamkonda Horticulture Farmers Producer Company Ltd office show off the FarmerChat application on their phones, that they use diligently.
FarmerChat, an AI assistant, is helping farmers across India and abroad get quick, relevant answers to questions about better farming practices, loans, market prices, subsidies, and much more. The result? Improved agricultural productivity.
Jyothi (extreme right) harvests onions and radishes in her farm. With her are Tirupathamma (centre), who is a unit in-charge at Papayapalem, and Anantlakshmi (extreme left), who is a Master Trainer for natural farming. They all use the FarmerChat application.

Kothapalli Jyothi (32) has hitched up her bright green-and-pink saree, and is plucking onions, brinjals, tomatoes and radishes as she walks nimbly through rows of vegetable crops. Behind her is a sea of fiery red chillies, glistening sharply in the afternoon sun. This healthy produce, and bountiful harvest in her five-acre farm (on lease) in Bellamkonda in Palnadu district of Andhra Pradesh is the product of good farming practices, many of which she has been learning about on her own through the FarmerChat application developed by Digital Green.

 

FarmerChat, developed and deployed in 2023, is an AI-driven chatbot (that can be queried in natural language) and Android application that helps farmers and agricultural extension workers in the remotest regions of the country get reliable, customised information and advisories to help increase their agricultural productivity.

 It’s been trained on a variety of data points – from government data from ministries, to agricultural practices from research institutions and universities, private sector partnership data such as market and weather information, and Digital Green’s own repository of peer-to-peer and other video advisories. The result? Information on demand, cutting down the farmers’ need to travel to Farmer Producer Organisations (FPOs) and meeting with extension workers every time they want answers on best prices, pesticide treatments, and the like.

Information at their fingertips

 Access to information was always a basic need for farmers – whether it was on soil health, organic techniques, or pesticide management, believes Gautam Mandewalker, Senior Director, Product Management, Digital Green. He explains: “Farmers who were more digitally literate would check YouTube or Google. Others would read the newspaper, ask their peers, the horticulture department, or extension workers.”

Jyothi harvests onions and radishes from her farm.

Digital Green had offered different iterations of information dissemination in the past, too. Many, such as videos, IVR-based systems, SMS-based systems, and deterministic chatbots through WhatsApp, received a good response from farmers during the cropping season. A Telegram chatbot launched in October 2023 was used by approximately 12,000 users across 3 countries (Ethiopia, Kenya and India) and 10 states in India (Andhra Pradesh, Telangana, Odisha, Gujarat, Rajasthan, Uttar Pradesh, Bihar, Madhya Pradesh, Assam, and Karnataka). “But it was very expensive to scale these systems, and it didn’t allow us to build innovative features and drive engagement in ways that farmers were demanding,” says Mandewalker.

The solution presented itself with the advent of conversational AI, which is powering the iteration of FarmerChat we see today. Built for low-literacy and low-connectivity users, it supports voice, text, and photo queries, and is designed to respond to seasonal, gender-specific, and climate-sensitive needs. “Now through FarmerChat, farmers and extension workers can simply search for that information in a chat interface, like they do on WhatsApp and Telegram,” Mandewalker says.

Powered by technology 

SaiPrasad Chirivirala, Head of Products at Digital Green, explains that FarmerChat uses Retrieval Augmented Generation (RAG) to pull answers from knowledge repositories. These repositories include government advisories and content from departments such as Horticulture, Krishi Vigyan Kendras, and others, as well as international journals.

“With FarmerChat, we have complemented our RAG system with direct inference. This allows farmers to access information that is also outside the scope of these content repositories – say, from other broad-based sources of agricultural information. For example, if a farmer has used a particular type of pesticide and wants another recommendation, FarmerChat can still provide those recommendations even if the answer isn’t contained in the curated content,” he says.

Chirivirala says they have seen a correlation between farmers with high acreage and the volume or queries they ask on the app. Small and marginal farmers are less likely to own smartphones, and therefore ask their questions to frontline extension workers, while farmers with farms bigger than 5–10 hectares tend to have smartphones, and therefore ask more questions.

Farmers also had queries beyond agricultural practices – they asked about subsidies, schemes, loans, etc. Because FarmerChat uses LLMs, it can offer nudges and notifications that supply this information or prompt the farmer to engage more with the app. “For every question, the app generates 3 follow-ups, prompting more questions,” Chirivirala explains.

For the farmers

The app is available in 20+ languages.

The app is available in 20+ languages. The Indian languages include Hindi, Kannada, Marathi, Odia and Telugu. Internationally, it is available in Swahili, Amharic, Oromo, French, Hausa, Kinyarwanda, Portuguese, Spanish, Shona and Zulu. Still, the number of queries asked via voice is five times that of text. There are more usage statistics, which tell their own story:

  • Queries mostly relate to pest & disease management (30%), seeds (15%), and fertilisers (10%).
  • Bihar, Odisha, Jharkhand, Karnataka, UP, AP, Telangana and MP top the list of states asking questions.
  • Over 50% of the queries come from male farmers, which could be because many women farmers do not own smartphones.

Expanding their reach

Vemula Padmavathi (55), is not one of them. She tells us that when she wanted to know how to tackle the leaf borers infesting her one-acre paddy field, she took a photo and uploaded it to the app. “I got suggestions to install sticky traps and pheromone traps to catch the borers, and install bird perches to attract birds that will eat the borers,” she says with a smile, showing us the app on her phone at the FPO office in Bellamkonda.

 

Like her, over 2,25,000 others now use FarmerChat across India, Ethiopia, Kenya and Nigeria. Of these, India accounts for over 1,60,000 users. Questions range from the best sowing techniques, to remedies for pests, information on schemes and subsidies, and best markets to buy and sell. Akula Sarswati (38), for example, says she got detailed advice about intercropping on her two-acre farm from FarmerChat. Based on that information, today, she plants coriander and onion in between her chilli plants, to manage pests.

Padmavati, a farmer from Nagireddypalem, shows the FarmerChat application on her phone.

Similarly, Jyothi, who was using marigold as an intercrop, now also plants beetroot, radish, coriander, and onion every six rows, “to control black thrips, and allow the chilli to extract nutrients from the soil”. She got this information from FarmerChat. It’s a double bonus — her income has increased as a result of the bounty from the chilli as well as the additional intercrops.

Over time, more features have been added, based on farmer inputs. The weather widget, for instance, is a big hit. “Farmers wanted to see the weather at a glance. They didn’t want to have to ask the app every day. So we created a widget that shows a seven-day forecast. And if they tap a button, they get a three-day forecast in natural language,” says Mandewalker. Text-to-speech was also introduced after they realised that farmers did not want to type and ask questions. Follow-up questions were introduced to make the app more proactive to use. The option to listen to the answers instead of reading them, was also a much-requested feature.

Coming up

Anantlakshmi, who is a Master Trainer for natural farming since 2016, helps Jyothi with her harvest.

Next on the anvil is enhanced video support, and support for more Indian languages. Digital Green also wants to test leaderboards and offer a gamification model where farmers who refer the app to others get incentives. Offline access and AI generated personalised crop calendars are also in the works.

 

The plan is also to make FarmerChat a unified, actionable solution that also addresses other requirements in the farmer’s journey, for eg: purchasing implements and livestock, seeking information on investing, or registering for government schemes. Mandewalker explains, “So instead of them just seeking information by asking a question,they can directly access that information because somebody else will be able to reach out to them through the app — say a scientist, or a service provider who is into drone services or seed supply, and so on”.

 

Jyothi, who has been growing paddy, vegetables, and chillies for the last 10 years, loves using the speech-to-voice feature on FarmerChat. “I had a lot of problems with pests. So I would take a photo on my phone, go to the frontline extension worker at the office, and show it to them. Or I would pluck a sapling, take it to the shop, and ask for remedies. With FarmerChat, getting advice has become much simpler. It saves me both time and money,” she says, before briskly making her through her field with a bunch of blood red chillies in her hand.

Want us to tell your data story? Write to us at cdssi@isdm.org.in with “Data Stories” in the subject line.

Author(s) :

Yes

Get in touch with authors

No ratings yet

Rate this article

Yes

Key topics

Data Science for social impact

Also found in

Share

Join Our Newsletter

Explore More Articles

Book

AI Systems for Social Good Playbook

This playbook serves as your step-by-step guide for designing responsible AI and data analytics solutions in the social sector. It guides both technical and non-technical teams through real-world projects, from understanding your context to monitoring your model post-deployment. Rooted in field-tested work, it aims to make AI more accessible, ethical, and impactful.
Case Study

Powering Asia’s Philanthropy Boom: Why Investing in Talent and Capacity Matters Now

In this blog written for WINGS, ISDM — a WINGS member and contributor to the APAC working group — makes the case that Asia’s philanthropy boom can only deliver lasting impact if investment in talent and institutional capacity keeps pace with growing capital. Drawing on India’s experience, it highlights gaps in governance, compensation, and leadership pipelines, and calls on donors and ecosystem enablers to treat capacity-building as a core strategy, not an overhead.
Blog

Code4Change a great platform to learn, collaborate, and understand how technical solutions can contribute to public systems like justice delivery

Participants from the third edition of Code4Change reflect on working with real-world justice system challenges and what it took to build meaningful solutions The Final Showcase of the third edition of Code4Change was held on 16 January 2026 at the International Centre Goa (ICG). The hackathon, organised by the Centre for Data Science and Social Impact (CDSSI) at ISDM, in partnership with DAKSH and the Centre for Social Sensitivity and Action (CSSA), Goa Institute of Management, centred on a single…
Case Study

The AI Platform Ending Pay Uncertainty in the Informal Sector

Kanak AI uses machine learning on real-world training data to predict income potential and recommend targeted upskilling courses, empowering job seekers to negotiate fair wages and build stable careers. Azad Ahmad (25), works as a Data Entry Operator in Gurgaon. He earns enough to support his family and sponsor his parent’s healthcare needs.. But it wasn’t always so. In 2021, when he graduated with a Bachelor’s degree, the COVID-19 pandemic was still raging, and he found himself navigating a tenuous…
We use essential and analytics cookies to operate this website and understand how visitors interact with it. As this site also functions as a login identity provider (IDP) for other ISDM portals, some cookies are necessary to enable secure authentication. By continuing to use this site, you consent to our use of cookies.