Science
From static soil snapshots to dynamic & emergent soil functions.
Biology is at the heart of soil function but knowing who is in the soil doesn’t tell you what they are doing. Instead of creating static databases of microbes we infer metabolic activity and complex biogeochemical mechanisms probabilistically, using AI.
Foundational models enhanced with real-world agronomic metadata.
Our AI models don’t start from scratch. They build on proprietary real-world agronomic metadata, giving them the context to infer complex biogeochemical interactions, identify emergent functions and spot metabolic reactions that correlational-based models completely miss.
Systems level biogeochemical interactions simulated in the Dry Lab.
The Dry Lab turns proprietary data into globally representative soil archetypes, then uses first-principles physics to simulate systems level biogeochemical interactions at scale, helping teams test likely soil behaviour before field work.
Ground-truth soil analysis keeps each model real world tested.
Computation is only the starting point. We continuously fine-tune our enhanced AI models with ground-truth soil analysis from our UK lab, feeding diverse new soil data back into the system so outputs stay real world tested.
Soil is a dynamically constrained system - context is key.
Once we understand a soil's functional potential, we can simulate its likely state in our computational Dry Lab, dynamically introducing physical and environmental constraints into virtual soil models, built and validated using proprietary datasets.
- Configure custom scenarios
- Understand real-world metabolic reactions
- Rapidly simulate product & practice interaction effects
Simulated with AI. Validated in the lab.
Simulation shows what is possible, but agronomy demands reality. Our Dry Lab explores potential, while our Wet Lab anchors predictions in ground truth.
- Physical testing feeds our understanding
- Measure real time soil function in the field
- Combine soil biology & physicochemistry with environmental & farming context
Closing the loop with continuous agronomic learning.
We combine the ground truth functional data generated in our Wet Lab with critical farming context including historic weather patterns, crop types and real-world farm management data.
