ILRI — Advanced Crop Advisory
Agriculture & AgriTech
Agro-met advisory that generates the bulletin for one farmer, not one district. An advanced scientific crop model built by agri and AI experts resolves IMD forecasts against each farmer's crop and growth stage — delivered in their local language, with every advisory traceable to its agronomic rationale.
An advanced crop model-based prediction and agro-meteorological advisory system built by Kenpath with ILRI (International Livestock Research Institute, CGIAR). Conventional agro-met advisory ends at the district bulletin — one document covering many crops at many growth stages, useful to an agronomist and rarely decisive for the farmer holding it. This system generates the bulletin at farmer scale instead. An advanced scientific crop model, built jointly by agricultural experts and AI experts, simulates how the farmer's actual crop at its actual growth stage will respond to the weather IMD is forecasting, and the advisory reaches the farmer in their own language by voice or text — with the reasoning behind every recommendation open to inspection.
Key capabilities
Advanced scientific crop model
Built jointly by agricultural experts and AI experts, the model simulates how a specific crop at a specific growth stage and soil responds to the incoming forecast — so advisory rests on crop science, not statistical correlation.
Agro-met bulletin generation at farmer scale
The system generates the advisory bulletin per farmer rather than per district, resolving the IMD forecast and AMFU bulletin science against that farmer's crop, stage and location at the moment the decision has to be made.
Curated agro-met knowledge corpus
AMFU bulletins, crop calendars, soil and rainfall data and IMD references, curated and schema-designed for the focus districts and crops — structured from the outset for scale-out to new geographies.
Local-language voice and text delivery
Advisory reaches the farmer in the language they actually speak, by voice or by text, so neither literacy nor a smartphone stands between the forecast and the decision.
Explainability by design
Every advisory traces back to the forecast and field conditions it used, its decision path and its agronomic rationale, so an AMFU scientist can audit any recommendation the system issues.
Expert-in-the-loop validation
AMFU agro-met experts review and validate outputs throughout, keeping agronomic judgment in the loop as the crop model and the advisory improve.
When the monsoon stops being predictable, advisory has to get precise
El Niño and La Niña have made monsoon variability the defining agricultural risk of the decade. An El Niño year can move the monsoon's onset by weeks, break it mid-season, and compress the window in which a sowing, irrigation or spraying decision is still recoverable — and the same rainfall that rescues a crop at flowering will ruin one at harvest. That is precisely the distinction a district-level agro-met bulletin cannot draw. The science needed to draw it already exists: IMD's forecasts, the agro-met expertise accumulated in AMFU bulletins, crop calendars, and soil and rainfall records going back decades. What has been missing is the mechanism to resolve all of it against one farmer's field. That mechanism is the crop model — an advanced scientific model built by agricultural experts and AI experts together, which simulates the crop's physiological response to the forecast at its actual growth stage in its actual soil. AI does the work around the model: grounding it in the curated knowledge corpus, generating the bulletin in the farmer's own language, and keeping an audit trail so every recommendation can be traced to its inputs and its agronomic reasoning. Crop science sets the answer; AI makes it reach the farmer while the decision is still open.
Kenpath's role
Kenpath is the technology partner building the agro-meteorological advisory system with ILRI. Kenpath builds the advanced scientific crop model together with agricultural experts, curates the agro-met knowledge corpus and designs its schema, builds the forecast and farmer profile pipelines against AMFU and IMD outputs, develops the local-language voice and text channel, and implements the bulletin generation, explainability and evaluation layers. The team works directly with AMFU agro-met experts so that agronomic judgment shapes the system continuously, rather than being consulted once at the end.