How Gemini Flash agents are helping a Michigan dairy farmer
Google Blog 2026-07-28 16:15:00
Context: Michigan dairy farmer Paul Windemuller is leveraging AI agents built with Gemini 3.6 Flash to streamline his farm's operations, automating complex data analysis and saving hours of manual work. Windemuller, a 2024 Nuffield International Farming Scholar, uses a custom multi-agent system to integrate siloed data and provide actionable daily briefings. His farm, Dream Winds Dairy, has grown from 30 leased cows in 2014 to a highly automated facility milking 260 Holsteins.
Key Facts
- Paul Windemuller's custom multi-agent system, built using Gemini 3.6 Flash, integrates siloed farm data, such as milk quality and feed logs, to calculate biological efficiency and provide actionable daily briefings.
- The system uses a local, directory-based file-interface to extract and merge visual and numeric metrics from CSV data exports, photos of paper receipts, PDFs, and invoices.
- Gemini 3.6 Flash features a 1 million token context window and a 64,000 token maximum output, achieving a 17 percent reduction in output tokens compared to Gemini 3.5 Flash at a lower cost per output token.
- The system is designed to optimize for Paul's primary metric: Daily Static Variable Margin (SVM), which isolates the true biological and operational efficiency from market noise.
- The reporting agent translates results into a daily Farm CEO Briefing that isolates daily margin drivers and provides actionable recommendations, such as adjusting ventilation for heat stress, allowing Paul to focus on growing his business.