Can AI help India’s struggling farmers double their income?

Sanjeev Mane, a sugarcane farmer in Maharashtra’s Sangli district, gets hourly updates about soil moisture and nutrients in his field, crop disease vulnerabilities, weather forecasts and more from an AI-powered app. He asks the bot how much water or fertiliser is needed. This has made his farming more scientific and helped him save money and his soil health by avoiding excessive inputs. He doesn’t need to rely on experience and intuition like before.“It’s like having a crop doctor with you... I don’t have to travel 450km to a university now to speak to an expert if I need help,” said Mane, who has been a farmer for more than 40 years and has his farm in Ashta village.The 69-year-old farmer was introduced to this technology by the Baramati Agriculture Development Trust, which partners Microsoft for AI-enabled agriculture. Sensor fusion, geospatial data and drone data are parsed by AI to deliver actionable knowledge to the farmers. In the pilot, farmers saw over 40% rise in crop yield, 30% reduction in chemicals and 40% drop in water usage.“Indian farmers have historically faced challenges due to lack of access to technology and knowledge,” said Venkat Krishnan, executive director, public sector, healthcare, education at Microsoft. “The success of AI in India’s farm sector sets a global benchmark, showcasing AI’s potential to improve efficiency and sustainability, while reducing reliance on harmful pesticides and fertilisers.”Agriculture employs about 43% of India’s population while contributing less than 20% to gross domestic product (GDP). Boosting farmer incomes has been a priority as the sector battles productivity and sustainability challenges amid unpredictable climate.Microsoft and Google along with startups like Cropin and KissanAI as well as non-profits like Wadhwani AI are developing and deploying AI to solve local challenges. The global AI in agriculture market was valued at $1.8bn in 2023 and is expected to reach $12.4bn by 2034 at a CAGR (compound annual growth rate) of 19%, according to market intelligence firm BIS Research.“The need of a farmer in general — whether there is AI or not — is basically information which is real-time, contextual and relevant to their conditions,” said JP Tripathi, director, agriculture programmes, Wadhwani AI, adding that this is so far only available in pockets.The stress on the government’s agriculture extension system is so high that its ability to provide the right kind of information to farmers on even a seasonal basis is limited, he said, noting that on average, one extension worker in India caters to 750-5,000 people, depending on the conditions.It may not even be humanly possible at the country’s scale. AI can play a useful role in this context, said Tripathi, adding that Wadhwani AI has created computer vision-based disease detection models for cotton integrated into the government’s National Pest Surveillance System and a grievance redressal bot for the PM-Kisan scheme. medium119851434To transform agriculture, accurate local-level data collection is key, experts said, adding that the government is making efforts towards this. Further, multilingual, speech-enabled AI models can help in data collection on the ground while accurate information can improve governance.Milind Sohoni, head of the Centre for Technology Alternatives for Rural Areas, IIT Bombay, for instance, leverages Google’s Agricultural Landscape Understanding API, which uses satellite data and AI to produce insights on farm fields across India, to create accurate village-level maps for the Maharashtra land records department.“AI is not a silver bullet but it is an accelerator… Otherwise, I would have to go out and survey the land,” Sohoni said. “Because of AI, now land records have a chance of being improved, because it can be done at scale. But it still needs the experts and it needs the land records department to do the right thing.”Whether it is land management, weather forecasting or irrigation, uncertainty in Indian agriculture needs to be addressed with better tools in order to see better yields, he said.Agriculture is one of the key focus areas for the IndiaAI Mission’s application development pillar for socio-economic transformation. Meanwhile, the agriculture ministry too is looking to increase digitisation and the use of emerging technology with an agri stack to enable services.“AI holds significant promise for India’s agricultural sector, especially in building a base layer for the digital agri stack,” said Alok Talekar, lead, agriculture and sustainability research, Google DeepMind. “AI can help analyse vast datasets, including satellite imagery and climate data, enabling precise, actionable insights at the individual farm level.”Google Cloud is collaborating with the Uttar Pradesh government for a Gemini-powered and Beckn-enabled Open Network for Agriculture that will provide farmers one-stop access to essential services, including advisory services, credit, mechanisation and market linkages to sell their produce.Bengaluru-based startup Cropin has built a small open-source agri AI model and an agri-intelligence platform for climate-resilient agriculture. Its AI-powered disease early warning model, deployed in tie-up with PepsiCo in India, reduced pest and disease threats by 80%, said Krishna Kumar, founder and CEO.It may take a decade or so for AI-based agricultural solutions across a crop’s life cycle to mature and reach about 50-60% of farmers, said Tripathi. medium nocaptionhttp://youtu.be/KyOKIMvYKQs?si=xJ1SRYfUL__ecduL medium119851260Global farmers say aye to AI toolsGlobally, agriculture is seen as a key sector that can be transformed with emerging technologies such as AI, as it is a segment most impacted by climate change. Climate change uncertainties apart, multitude of other factors that determine crop yields and distribution — from quantum of inputs to soil condition, rainfall patterns and market linkages — make farm sector a prime candidate to leverage AI for more efficient practices, predictability, identification of optimal inputs and timing, as well as advisory to farmers on the ground. Here’s a look at how different countries are making moves to apply AI to agriculture.The USAI Institute for Next Generation Food Systems launched in 2020 with a $20mn government grant to optimise food production, processing, distributionFocus areas are efficient plant breeding, water/nitrogen use, crop forecasting and reducing food wasteAs much as 87% of US agriculture industry using AI as of late 2021, estimates suggestGermany10% of farms use AI solutions, with 38% planning to adopt itCrop protection, forecasting seen as promising areas to deploy AIGerman Research Centre for AI promotes AI solutions for efficient and sustainable farmingChinaAI used in genomic analysis and breeding for high-yield cropsDeepSeek models help farm robots with pest detection and real-time monitoring, used for local insights for disease prevention, meteorologyChina’s AI in agriculture market estimated at over $12bn in 2024African nationsRwanda’s agriculture strategy includes AI and IoT for yield improvement and sustainabilityMkulima GPT, a ChatGPT-based Swahili chatbot, aids smallholder farmers with disease detection and adviceHello Tractor in Kenya connects farmers to tractor owners, streamline ops, digitised 3.5mn acres since 2014Kenya’s Agricultural Observatory Platform provides real-time weather data to 1.1mn farmersAI soil testing kits in Ghana offer personalised fertiliser recommendations to farmers