Inspiration

Small holder farmers around the globe are losing billions of dollars each year due to declining yields caused by climate change. Meanwhile, agriculture is responsible for a staggering 24% of global carbon emissions. But what does this really mean? It means smallholder farmers are missing out on revenue that could transform their livelihoods. It means they’re unable to contribute as much as they could to solving food shortages, especially in developing countries. Regenerative agriculture has the power to change this. It can decarbonise farms, boost soil health, increase yields, and capture more carbon, paving the way to net-zero emissions for the entire agriculture value chains. But the challenge? Farmers struggle to learn and apply these techniques.

What it does

Agrify is a platform decarbonising the agricultural value chain with regenerative intelligence. We built tools for farmers to use to decarbonize the agricultural value chain and earn carbon income through regenerative intelligence.

Our platform is a game-changer. All a farmer needs to do is, register with proof of land ownership and enter their farm location on our app. We then remotely monitor their farm’s health by identifying their unique soil profile, powered by our cutting-edge machine learning models, and track how much carbon they’re capturing.

But we don’t stop there. We use this data to generate actionable recommendations through our regenerative intelligence model, powered by Gemini. Our platform offers step-by-step guidance in regenerative farming practices tailored to their farm's unique soil profile—delivered in the farmer’s native language.

How we built it

We developed advanced machine learning models in Python to estimate soil properties for farms using only satellite data. We achieved this by collecting historical soil data for each property under consideration and pairing it with corresponding satellite data. This dataset was then used to train our models, enabling them to accurately estimate soil properties based on satellite inputs. These estimates are then processed through our Regenerative Intelligence Model, a custom foundational model built on Google Gemini’s generative AI, to generate tailored recommendations for enhancing soil health and increasing soil organic carbon (SOC). To make this accessible, we created an intuitive, mobile-friendly web app where farmers can easily register and receive their personalized recommendations.

Challenges we ran into

Obtaining historical soil data and satellite data to train our models was particularly challenging, especially in African countries where soil data collection and management practices are less robust. Additionally, securing programmatic access to satellite data posed its own set of difficulties.

Accomplishments that we're proud of

Having working models for most soil properties. Having a platform that is extremely user-friendly and easy to use and only requires you to provide an address in order for it to work.

What we learned

We discovered the critical importance of companies partnering with farmers in their supply chain to decarbonize their value chains. We also recognized the profitability of the carbon marketplace for companies that generate and sell carbon credits.

What's next for Agrify

We are seeking a partnership with a company interested in adding a meaningful project to their ESG, CSR or Net Zero portfolio. We are looking for a sponsor to support 100 smallholder farmers in transitioning to regenerative agriculture from subsistence farming, enhancing yield and income.

Share this project:

Updates