Next Generation AI Model for Tomatoes

More accurate forecasts, less manual work

Your Harvest Forecast for tomatoes now runs on a fundamentally new AI model. It learns from real cultivation data across the Source platform, which means it accounts for climate variability and greenhouse-specific conditions that no single grower could encounter alone.

What changes for you:

Several data processes that previously required regular manual registrations have been automated. You no longer need to maintain them yourself, and the forecast actually improves because of it. Less input from you, better numbers from Source.

At the three-week horizon, mean forecast accuracy has increased by 33% compared with the previous generation. The share of cultivations with forecasts more than 20% off target has dropped by 25%, and severe outliers have been cut in half.

Why this matters for your operation:

More reliable harvest predictions make it easier to hit commitments, negotiate a better price, and reduce waste. And because the model learns from every cultivation running through Source, its performance compounds as more growers use the platform. The forecasts you rely on will keep getting better over time.

Rollout is phased. An initial group of greenhouses is already live, with broader expansion planned in the coming weeks and full deployment to all tomato customers expected in the coming months.