The useful question is what changes for users, developers or buyers, and whether the announcement stays industry context or becomes something people can actually use.

Paul Windemuller, a Michigan dairy farmer, has implemented AI agents powered by Google’s Gemini 3.6 Flash to streamline farm operations. His custom multi-agent system consolidates siloed data—such as milk quality logs, feed records, and sensor readings—into a unified analysis, reducing manual spreadsheet work from hours to minutes each day. The system integrates data from isolated software sources, including cow-tracking collars, weather stations, and milk shipment portals, to calculate daily profitability metrics like Static Variable Margin (SVM). By automating these calculations, Windemuller reclaims time previously spent on administrative tasks, allowing him to focus on farm management and growth.

Windemuller’s AI system operates locally using Google Antigravity, avoiding external APIs or web scraping to maintain data control. The system processes exported CSV files, PDF invoices, and even handwritten receipts by saving them to a monitored folder, where Gemini 3.6 Flash’s multimodal capabilities extract and merge visual and numerical data. This approach ensures sensitive farm data remains on-site while transforming raw files into a cohesive business overview. The daily workflow includes automated calculations of SVM, which isolates biological and operational efficiency from market fluctuations, providing reliable insights for decision-making.

The cost-effectiveness of Gemini 3.6 Flash was critical for Windemuller’s small-scale operation. The model’s advanced reasoning and tool-use capabilities, combined with a 1 million token context window and reduced output token costs, made continuous agentic workflows feasible. Benchmark evaluations indicate a 17% reduction in output tokens compared to its predecessor, lowering operational expenses for small businesses. Windemuller’s system prioritizes SVM, a metric that remains stable despite volatile milk and feed prices, ensuring consistent performance tracking. This stability allows the AI to deliver actionable recommendations, such as adjusting ventilation for heat stress when SVM declines.

Windemuller’s initiative demonstrates how AI can be adapted for small agricultural businesses. The daily Farm CEO Briefing translates complex data into clear, actionable insights, such as identifying margin drivers or recommending operational adjustments. By eliminating manual spreadsheet work, the system enables Windemuller to spend less time on data analysis and more time on farm operations. His approach highlights a scalable model for independent farmers seeking to integrate AI into daily workflows, emphasizing cost efficiency and local data control as key factors for adoption.