PepsiCo's Lay's gave its 27,000 potato farmers a 10-day blight early warning.

Lay's sources all its Indian potatoes from about 27,000 small farmers, most of whom own under a hectare. A single blight outbreak can cut a potato crop by up to 80% if it is not caught early, and erratic monsoon shifts have made the old seasonal planting playbook unreliable.

Instead of trying to inspect every plot, Lay's teamed with agri-tech firm Cropin to build an AI model that reads satellite imagery, remote sensing and years of historical weather, then issues a crop-level forecast. The system can warn of disease, frost and water stress up to 10 days ahead.

Farmers get the forecast as simple color codes on a smartphone app like Cropin Grow or Plot Risk, while Lay's gets the same end-to-end visibility into its sourcing. The pilot covered 62 farms across Gujarat and Madhya Pradesh before scaling.

The model was shortlisted at Cannes Lions and won the Creative Data Grand Prix at Spikes Asia 2024, a signal that the jury read it as a genuine supply-chain mechanism rather than a marketing stunt.