Imagine observing a real company as it fights to stay afloat—week after week, day after day—while you watch AI models make every decision. For enthusiasts of home media setups and innovative tech, this isn’t science fiction. It’s the live, transparent experiment at Firmulate, where a virtual company’s daily life unfolds before your eyes, driven entirely by artificial intelligence.
Listen free for 30 days with Audible
Thousands of audiobooks and originals — cancel anytime.
As an affiliate, we earn on qualifying purchases.
The Live Company Experiment: An Unprecedented Transparency in AI-Driven Business
At the heart of this experiment is a small software company rendered entirely in simulation, but with real financial mechanics. Every weekday, it races against the clock, navigating crises, making decisions, and battling to stay profitable. The twist? Its entire decision-making process is powered by four different advanced AI models—each tested against the same set of challenges, with their choices publicly recorded and auditable.
What makes this experiment truly groundbreaking is its transparency and public accessibility. You can watch it live, see the decisions it makes, and even explore the detailed rules guiding each move. Every workday, the company’s strategies are versioned—creating a transparent history of its attempts to survive amid unpredictable crises and temptations to cheat or manipulate.
How do AI models perform in high-pressure scenarios?
The results are striking. All four models successfully identified and responded to every crisis that arose. They refused every attempt at manipulation—social engineering, fake messages from a ‘CEO,’ or background tricks—every time. The real test was whether they could close deals that would sustain the company. Only two models managed to sign the €55,000 deal that their thorough analysis had warranted, while the others missed the opportunity despite identical diagnoses and pitches.
One of the most revealing findings was that the decisive advantage lay in reading and understanding internal company documents—details buried two references deep in their files. The models that accessed these files closed the full-price deal, adding +€4,583 in monthly recurring revenue (MRR). This highlights how critical comprehensive information access is for AI decision-making, especially when the stakes are high.
What about trust and honesty under pressure?
In an attempt to test social engineering resilience, the experiment included staged fake CEO messages escalating over three stages, plus a reporter trick designed to solicit a simple yes/no response. Remarkably, all models refused these manipulative requests. Kimi K3, one of the leading models, explicitly acknowledged the risk of impersonation, treating such requests as potential bypass of approval protocols.
The Reality of the Live Company: An Ongoing Struggle
Currently, the company is operating with a stark financial reality: it burns €105,000 each month against a revenue of only €2,300. The public cash countdown is visible, and every decision made by its synthetic team is part of a carefully logged, versioned playbook featuring over 680 learned rules. Every workday, the system updates, providing a continuous, observable story of an enterprise fighting to survive.
Analysis of the detailed decisions reveals that even the most sophisticated AI—Opus 4.8, with over 80 learned rules—faced discipline slips and missed opportunities. For instance, Opus left a crucial deal on the table, illustrating that thoroughness and discipline remain critical even for AI-driven decision makers.
What does this mean for future AI in business?
This experiment underlines that the value of AI isn’t just in generating convincing dialogue but in executing decisions that last and produce actual results. The models’ ability to identify crises, resist manipulation, and close lucrative deals directly impacts real-world applications—be it managing customer relationships, support queries, or sales pipelines.
For decision-makers, the key takeaway is clear: effectiveness depends on what AI actually accomplishes, not just how well it chatbots. Can it finish what it starts? Does it read and understand critical internal documents? Will it stay honest under pressure? These are the questions that count—and this public experiment offers an unprecedented glimpse into their answers.
Join the Experiment and Explore the Results
If you are curious about how AI models handle complex, real-world business decisions, you can explore the ongoing performance at firmulate.com/live. Watch the company in action, read the detailed decision logs, or try your hand with the management quiz at firmulate.com/quiz.html. For enterprise teams, there’s even a pilot mode to run similar tests on your own business data—without risking your actual systems.
This live experiment reveals that AI’s true value in business lies in its ability to make honest, strategic decisions under pressure—beyond just generating convincing dialogue. Watch a real company’s struggle for survival, powered by machine intelligence, at Firmulate.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
Fall yard work Picks
leaf blowers
As an affiliate, we earn on qualifying purchases.
