Imagine watching a startup operate live, making real decisions, facing crises, and even risking bankruptcy — all without a single human employee. This isn’t science fiction; it’s the latest experiment in AI management, and it’s happening now at firmulate.com/live. For entertainment, curiosity, or a glimpse into the future of work, this is the real story of an AI-managed company fighting for its financial life, visible to anyone willing to tune in.
The Live Experiment: An AI-Run Company in Action
At the heart of this ground-breaking experiment is a small software company run by an AI simulation named Firmulate. Unlike familiar chatbots or assistants, this company has 13 synthetic employees, each driven by sophisticated algorithms that mimic management decisions, finance mechanics, and crisis handling. The company’s daily life is fully public, with every decision versioned and publicly accessible, providing a unique transparency rarely seen in the corporate world.
Each day, the AI models—powered by frontier language systems—must navigate the company’s worst week: confronting customer crises, internal dilemmas, and ethical temptations. This isn’t just a game; it’s a real-time demonstration of how AI can perform in a complex, unpredictable environment. The performance is scored on a scale from 26 to 95, with the highest score achieved by the GPT-5.6 model. This AI identified crucial buried information in company files, enabling it to close a deal worth over €4,500 in monthly recurring revenue, illustrating that reading and understanding internal documents can be a competitive edge.
Decisiveness Under Pressure: The AI’s True Test
One might think that AI is only as good as its ability to generate convincing chat responses. Not here. Every decision—whether to sign a €55,000 deal, refuse a manipulative sales tactic, or escalate a crisis—is rigorously recorded and auditable. In a notable test involving social engineering, all four models refused to be duped by staged manager messages or reporter tricks, demonstrating a high level of trustworthiness and ethical restraint.
Interestingly, the models that read internal documents made the difference in securing a lucrative deal. The buried fact within the company’s files proved decisive, revealing that access to internal knowledge can tip the scales in a competitive environment. This underscores the importance of information access in AI decision-making — a vital consideration for real companies contemplating AI integration.
As an affiliate, we earn on qualifying purchases.
The Harsh Reality of an AI Company Losing Money
The company operates with a monthly cash burn of €105,000 against a meager €2,300 in monthly revenue. It’s a startup in financial distress, with a public countdown to bankruptcy. Each day’s decisions are carefully monitored, and every code update, known as a version, reflects the latest learning and strategies. This rigorous versioning environment allows observers to see exactly how different AI models—some more disciplined than others—perform in managing real money and real crises.
The most thorough participant, Opus 4.8, ran over 80 learned rules and conducted deep analysis but still finished last among the models. Its discipline slipped during critical moments, such as failing to escalate issues properly. This highlights an essential insight: even sophisticated AI systems can falter without disciplined processes or effort parameters.
internal document analysis tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Does This Mean for the Future?
While the company is a controlled experiment, the implications are clear. As AI models become more capable of handling complex, real-world tasks, their ability to remain honest, read internal data, and finish what they start will be crucial. This experiment shows that AI can identify crises, resist manipulation, and even close deals without human intervention — but only if it is designed and tested for these exact circumstances.
For businesses, the takeaway is simple: before deploying AI into critical roles—whether in customer support, sales, or management—it’s vital to test how well it performs under pressure. The question isn’t just about how well an AI writes or speaks; it’s whether it can complete tasks ethically, thoroughly, and reliably when it matters most.
As an affiliate, we earn on qualifying purchases.
Watch the Company Live
Curious about this high-stakes AI experiment? You can watch the company’s daily struggles unfold in real time at firmulate.com/live. See how decisions are made, how models respond to crises, and whether a machine can truly manage a business — all in plain sight, every workday. For insights into decision-making, read the actual quotes from the models at firmulate.com/quotes.
This live experiment demonstrates that AI can identify crises, resist manipulation, and close deals, but only if rigorously tested. Watching this company fight for survival offers a rare glimpse into AI’s potential and its limitations in real-world management.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
As an affiliate, we earn on qualifying purchases.
