Paul Windemuller runs a 201-cow dairy farm in Michigan. Dairy farming operates on some of the tightest margins in agriculture, and those margins depend on data that used to sit scattered across milking robot exports, feed logs, and a shoebox of paper receipts.

Google published Windemuller's story this week as an example of Gemini 3.6 Flash, its latest model built for agentic reasoning and tool use. Windemuller used it to build a local, multi-agent system that reads his farm's data and tells him, every morning, what is driving his profitability up or down.

The system has four parts. An orchestrator manages the daily workflow. Ingestion agents standardize raw files from his milking robots and feed logs. An analysis agent evaluates the biological and weather variables that move a dairy farm's numbers day to day. A reporting agent turns all of it into a plain-language morning briefing.

Windemuller built the system around Daily Static Variable Margin, a number that holds market prices constant so he can see whether a swing in his margin came from something he controls, like feed efficiency or herd health, rather than a broader move in prices he cannot do anything about. When that number drops, the system tells him why.

The system does not touch APIs or scrape data from anywhere. It watches a monitored folder. When Windemuller saves a CSV export or a photo of a receipt into that folder, the system reads it, merges it with everything else, and keeps his data under his own control.

Google frames this as leveling the playing field for small businesses. Daily agentic workflows that require constant reasoning and tool execution have historically been too expensive and technical for an independent farmer to build. A cheaper model narrows that gap, but the more interesting part is not the model. It is that one farmer, with no engineering team, built a working multi-agent system to run a core function of his business.

That is the real signal for any small business owner reading this. You do not need to write code or hire a developer to get a version of what Windemuller built. Viktor lives inside Slack and Microsoft Teams. You @mention it in a thread the same way you would ask a colleague. The output, a PDF, a report, a task created in your CRM, an email drafted in Gmail, lands where it should land.

Three uses map onto what Windemuller built. Viktor can watch a folder or inbox for recurring documents, invoices, receipts, supplier reports, and turn them into a standing summary instead of a monthly pile. It can build a daily briefing pulled from the numbers that matter to your business, the way Windemuller's reporting agent does for his margin. And it can flag the one number that moved for a reason you need to know, rather than making you scan a spreadsheet yourself.

The lesson from a dairy farm in Michigan is not really about agriculture. It is that the tools to run a business like a much larger one, with real-time reporting and automated data work, no longer require a technical team to build them.

Any tool that touches client data and business systems has to earn trust before it earns access. Viktor is SOC 2 compliant, GDPR and CCPA compliant, and CASA Tier 3 certified. Credentials sit in an encrypted vault, your data is never used to train models, and every action Viktor takes in Slack requires your approval before it happens. Full detail is at viktor.com/security.

You get $100 of free credits to begin. No time limit, no commitment. That's enough to do real work and see what Viktor can actually do before you spend a penny. There's also $50 off your first bill. You must use this exact link to receive both benefits.

See what it can do for your own numbers: https://AIThatDelivers.com.

Disclosure: Some links in this article are affiliate links. If you choose to get started with Viktor using the links provided, I may receive a commission, at no additional cost to you. I only recommend tools I use and believe in.