By Ankur Aggarwal, Chairman of CropLife India and Managing Director of Crystal Crop Protection Ltd
In 2025, the Ministry of Agriculture and Farmers Welfare rolled out an AI-enabled monsoon forecasting programme that delivered localised weather alerts via SMS to 38 million farmers across thirteen states, at a cost of less than one paisa per farmer. The scale of the initiative was unprecedented. More importantly, it marked a shift in how artificial intelligence is being deployed in Indian agriculture: not as a pilot or proof of concept, but as a decision-support tool operating at population scale.
That shift matters because Indian agriculture is facing a convergence of risks. Climate volatility has increased the frequency of extreme weather events. Pest and disease pressures are intensifying and becoming less predictable. Resistance to existing crop protection chemistries is rising across major crops. For most farmers, especially smallholders, farm income is now shaped less by effort alone and more by timing, probability, and risk management.
The scale of crop loss illustrates the challenge. Studies across Indian agriculture estimate that pests, diseases, and weeds together account for roughly 15 to 25 percent yield losses across major crops, translating into annual economic losses running into several lakh crore rupees. Climate-linked shocks have compounded this burden, increasing the likelihood that a single misjudgement or delayed intervention can wipe out a season’s earnings.
Crop protection exists to reduce this uncertainty. Yet the scientific system that underpins it is under strain. Discovering new active ingredients has become slower, more expensive, and less predictable. Globally, developing a new molecule can take more than a decade, with the vast majority of candidates failing in early discovery or optimisation stages. As resistance builds against older chemistries, this slow pace of innovation leaves farmers increasingly exposed.
Artificial intelligence is beginning to change how this problem is approached.
At its core, molecule discovery is a problem of probability. Researchers are working across vast chemical spaces, knowing that only a small fraction of compounds will ultimately deliver efficacy against target pests while meeting safety, environmental, and regulatory thresholds. For decades, this has meant broad trial-and-error screening, long timelines, and high early-stage failure.
AI allows this search to become more directed. By analysing historical chemistry data, bioassay outcomes, pest biology, and known resistance mechanisms, machine-learning models can prioritise candidates with a higher likelihood of success before they enter expensive laboratory testing. Across global crop science R&D, this approach is already being used to shorten early discovery cycles and focus effort on genuinely differentiated modes of action.
For Indian agriculture, the relevance is particularly acute. Pest behaviour, resistance pathways, and climatic conditions in India differ sharply from those in many Western markets where legacy chemistries were originally developed. AI-enabled discovery makes it possible for Indian field data to inform molecule selection much earlier, improving the chances that new solutions are not only effective in controlled trials, but relevant under Indian farming conditions.
Discovery, however, is only part of the story. Many crop losses occur not because solutions are unavailable, but because interventions are mistimed or misapplied. This is where AI is also reshaping how crop protection is used on the ground.
The shift is already visible in how crop risk is being managed at scale. Under the Digital Agriculture Mission, the Ministry of Agriculture and Farmers Welfare has begun embedding AI directly into national crop protection systems. The National Pest Surveillance System now uses machine learning and image-based diagnostics to monitor pest incidence across 66 crops and more than 430 pest species, enabling early warnings and targeted advisories before infestations escalate.
Platforms such as Krishi 24/7 complement this by scanning agricultural media and field reports in multiple Indian languages to surface emerging pest and disease risks in near real time. At the foundation of these efforts lies AgriStack, which is building a unified farmer and land database to enable personalised, crop- and location-specific advisories. Together, these systems are shifting crop protection from reactive response to anticipatory risk management, with reported outcomes including significant changes in farmers’ planting and intervention decisions and meaningful reductions in avoidable input costs in precision-led deployments.
AI-driven advisories that integrate hyperlocal weather forecasts, crop-stage information, and pest surveillance can guide farmers on whether intervention is necessary at all, and if so, when it is most effective. This improves efficacy while reducing unnecessary spraying, lowering input costs, and slowing the development of resistance. In crops such as cotton, AI-supported pest monitoring has already enabled a shift away from calendar-based spraying towards threshold-based decisions, improving net returns and yield stability.
What allows these advances to reach scale in India is the presence of digital public infrastructure. Just as the Unified Payments Interface transformed financial transactions by making them simple, universal, and low-cost, India is now building similar public rails for artificial intelligence. As the Minister for Electronics and IT, Ashwini Vaishnaw, has often observed, just as UPI democratized payments, the IndiaAI Mission seeks to democratize intelligence. The implication is straightforward: once intelligence is delivered through accessible, trusted channels, its marginal cost falls sharply while its impact multiplies.
This approach will be central to the India AI Impact Summit 2026, where India’s experience of deploying AI as a tool for developmental impact will be presented on a global stage. The relevance of this model extends beyond national borders. Many countries across the Global South face similar constraints: small and fragmented landholdings, limited extension capacity, high exposure to climate risk, and rising biological threats to crops. India’s experience suggests that when AI is paired with public digital infrastructure and grounded in local agronomy, it can materially improve how agricultural risk is managed.
Artificial intelligence will not remove uncertainty from farming. Weather shocks, pest outbreaks, and biological limits will remain part of agriculture. But AI can reduce avoidable loss. By accelerating molecule discovery, improving precision in application, and embedding intelligence into everyday decisions, it strengthens farm resilience at a time when resilience is becoming the defining requirement of agricultural productivity.
For farmers, the benefit is concrete. Better information, better tools, and better timing improve the odds that a season’s effort results in a stable harvest. That is the promise of AI in crop protection, and why India’s approach, if sustained and scaled thoughtfully, matters well beyond its borders.
. Views expressed are personal