AgriSmart
Client
- AgriSmart
Project type
- AI Co-Pilot for Sustainable Farming
What I did
- Web App
- UX
- AI
Year
- 2025
A mobile-first AI co-pilot for small and medium farmers in developing regions. It takes satellite imagery, weather data and AI analysis — the instruments large corporate farms pay for — and turns them into plain, actionable advice on an ordinary smartphone, in the farmer's own language.
- 01
Inherited farming knowledge is losing its reliability against climate volatility: the patterns it encodes describe a climate that is no longer the one outside.
- 02
The technology that answers those questions — satellite analysis, agronomy modelling — is priced for corporate agriculture and out of reach for smallholders.
- 03
Advice that does exist arrives as reports and figures, which is not a form a decision can be made from in a field.
- 04
Language and literacy stand between the farmer and the guidance, so tools written for one audience simply do not reach the other.
- 01
Satellite NDVI analysis through Google Earth Engine, so crop health is measured from orbit rather than estimated by eye.
- 02
Pest and disease diagnosis from a phone camera via the Gemini API — the farmer photographs the leaf and gets an identification and a response.
- 03
A contextual assistant that answers in the local language, so the guidance arrives in the form the decision is actually made in.
- 04
A sustainability score for water management that makes the long-term cost of a choice visible at the moment the choice is made.
- 05
Mobile-first throughout, designed for the device and the connection farmers actually have rather than the one the data assumes.
Reading a field from orbit
NDVI over the season shows which part of a plot is under stress before it is visible walking it, and says what to do about it.
Identifying a pest
The farmer photographs an affected leaf; the diagnosis and the treatment come back in their own language, in minutes rather than a visit.
Deciding when to irrigate
Weather, soil condition and the sustainability score combine into a recommendation that accounts for the water it costs.
LeaLab
