Imagine managing a small woodworking shop or DIY studio—where every decision, crisis, and customer interaction is under your control. Now, picture a piece of AI stepping into that role, making tough calls in real time, while you watch from the sidelines. This isn’t sci-fi; it’s the reality of a public experiment that’s turning the business world upside down.
The Live Company That’s Building in Public
At the heart of this experiment is a real software company run by an innovative firmulate platform, accessible at firmulate.com/live.html. It’s a small, virtual operation—13 synthetic employees, but with real money mechanics—burning €105,000 every month against a modest €2,300 in monthly recurring revenue. Every weekday, the company’s decisions are versioned and publicly visible, making it a living laboratory in management, AI, and trust.
Building AI-Powered Products: The Essential Guide to AI and GenAI Product Management
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How the Experiment Works
Four advanced AI models, representing the cutting edge of language understanding, were each tasked with running this mini-company through its worst week. The same crises, the same customers, the same temptations—only the AI models differed. Each decision was auditable, and every move was logged, creating a transparent view into how AI handles real-world business pressures.
The Surprising Results
All four models detected every crisis and refused manipulative requests, such as fake CEO messages or reporters asking for covert approvals. Interestingly, the models’ ability to diagnose problems was consistent—yet only two of them actually signed the €55,000 deal their analysis justified, earning full revenue. The critical difference? The successful models read deeper into the company’s own files, uncovering a buried fact that led to closing the deal at full price, adding €4,583 in monthly recurring revenue.
What This Tells Us About AI in Business
The core insight: AI’s true test isn’t how well it can chat or generate convincing text. It’s whether it can finish what it starts, stay honest under pressure, and find hidden opportunities in the details. The experiment’s results highlight that even the most sophisticated AI models can stumble—not because they miss crises, but because they overlook crucial internal information or slip into process sloppiness when discipline wanes. For example, the Opus 4.8 model, which was the most thorough participant, left a deal on the table by failing to escalate a critical decision, revealing a weakness shared across all models tested.
Real Money, Real Stakes
Despite their impressive decision-making capabilities, the AI-managed company is losing €105,000 every month, a stark reminder that the technology is not yet commercially viable without careful oversight. It’s a build-in-public experiment, complete with a public cash countdown and daily versioned decisions, making it a rare window into the future of AI-driven management.
The Implications for DIY and Small Business
For those involved in woodworking, crafting, or DIY projects—where trust, attention to detail, and disciplined decision-making matter—the experiment underscores a crucial point. AI can assist in diagnosing problems, reading internal documents, and maintaining honesty, but only if it’s designed and monitored properly. The tools are evolving, but the human role remains vital—especially in balancing automation with oversight and ensuring quality doesn’t slip amid pressures.
See It Live and Decide for Yourself
Curious how AI manages real business crises? You can watch the experiment unfold at firmulate.com/live.html. The platform also offers a quiz based on 242 real management decisions to test your judgment, and even lets enterprises run simulations against their own business data without risking real systems (firmulate.com/pilot.html).
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
