CASE STUDY
Grobiero Farms
AI-driven agronomy intelligence for a multi-property California tree-nut operator
THE SITUATION
A multi-property operation running on a dozen disconnected systems
Grobiero manages tree-nut and specialty-crop properties up and down California for owners and lenders who expect institutional-grade reporting. Almonds, pistachios, walnuts, organic. Twenty plus years of agronomic history sat across AgWorld field operations, Semios crop sensors, HortAu irrigation analytics, financial systems, and the field team's own notes.
Each owner question meant stitching pieces together by hand. The team wanted intelligence that worked the way they worked, not a generic agritech SaaS forcing their process into someone else's schema.
BEFORE
Each owner question meant stitching pieces together by hand from a dozen tools.
AFTER
One intelligence layer the whole operation runs on. Field judgment, amplified by AI.
WHAT I BUILT
An AI intelligence layer over the entire operation
I built a custom platform that unifies operational, agronomic, and financial data into one queryable surface. The team asks questions in plain English: yield trends across properties, irrigation anomalies, input cost outliers, supply and logistics exposure.
Underneath: ML pattern recognition on multi-season yield and input data, automated anomaly detection on sensor and finance streams, and reporting that property owners can rely on without a phone call.
Unified intelligence layer
AgWorld, Semios, HortAu, and their financial systems flowing into a single typed warehouse the team actually trusts.
Plain-English queries
An LLM-backed query layer over the entire operation. Field and office teams ask questions and get grounded answers, not a SQL prompt.
Pattern recognition at scale
ML on multi-season, multi-property data: yield trajectories, irrigation anomalies, input cost outliers worth a phone call to the field.
THE OUTCOMES
Scale without losing what made the operation good
1
platform across the full property portfolio
Live
through a full growing season, in daily use
Daily
plain-English queries by field and office teams, no engineer in the loop
STACK
- Next.js
- Postgres
- Gemini
- AgWorld API
- Railway
- Cloudflare
