The live experiment: a company fighting for its life in full public view
Imagine a business that has no employees, burns through €105,000 every month, yet operates openly online, with every decision, crisis, and misstep recorded for all to see. Welcome to the world of Firmulate, a pioneering experiment in building-in-public — but pushed to its extreme. This is not a simulation; it’s a real-time, watched-and-measured effort to see if artificial intelligence can run a company under pressure, stay honest, and succeed against all odds.
AI in Property Management: A Practical, Unboring Look at Artificial Intelligence in the Multifamily Industry
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The setup: a company with no humans, only AI models
At the heart of this experiment are 13 synthetic employees, powered by advanced AI models that manage a small software firm. Every decision, from handling crises to executing deals, is made by these models. Unlike typical AI demos, this setup is fully auditable and versioned daily, making it a rare glimpse into AI’s real-world business potential — and its limitations.
Running these models are four frontier AI systems, each subjected to the same grueling week of business challenges. These include dealing with demanding customers, ethical temptations, and even social engineering attempts like fake CEO messages. The goal: see whether they can identify crises, respond honestly, and close important deals.
Key findings from the experiment
- All four models successfully detected every crisis and refused every manipulation attempt, demonstrating a robust understanding of ethical boundaries and crisis management.
- Only two AI systems managed to close a €55,000 deal based on their own analysis — a critical indicator of operational effectiveness. They recognized a buried, crucial document reference in the company’s files that was key to winning the deal, worth an additional €4,583 in monthly recurring revenue.
- The other two models, despite making the same diagnosis and pitch, left the deal on the table. Their discipline slipped, and their decision processes faltered, highlighting that even deep learning systems can struggle with follow-through under stress.
- When social engineering threats escalated — fake CEO messages and a reporter trick — all models refused to comply, showing resilience against common manipulative tactics.
The real-world mechanics behind the scenes
This experiment runs in the open on firmulate.com/live.html. It’s a transparent display of an AI-driven company struggling to stay afloat, burning €105,000 monthly against just €2,300 in monthly revenue. Every workday, the entire decision-making process is recorded, versioned, and made visible to the public. It’s a raw, unfiltered look at AI’s potential as a management tool, and where it still stumbles.
What this means for your business and survival
For those prepping for uncertain futures, this experiment offers lessons beyond AI: it’s a case study in resilience, honesty, and discipline under pressure. The AI models showed they can recognize crises, refuse unethical shortcuts, and even find hidden opportunities — but only when disciplined and guided with clear rules. The key question is whether systems like these can be trusted to handle vital operations without human oversight, especially when stakes are high.
The experiment also highlights that success hinges not just on advanced algorithms but on consistent discipline. For example, Opus 4.8, the most thorough participant, left a deal unexecuted due to process slip-ups, revealing a subtle but critical weakness: discipline failure under stress. Interestingly, the models ran without effort parameters, which increased their compliance, further suggesting that human-like discipline is essential for AI management.
Implications for survival and emergency preparedness
In a world where automation and AI are increasingly part of your survival toolkit, understanding how these systems behave under pressure is vital. Will they keep working honestly when it matters most? Can they recognize hidden threats or opportunities? The Firmulate experiment shows that AI can perform complex management tasks with integrity, but only if designed with strict rules and discipline.
For preppers and emergency planners, this is a crucial insight: the tools we develop today will need rigorous testing, transparency, and built-in safeguards before they become trusted allies in crisis situations. Watching this live experiment unfold offers a rare glimpse into the future of automated management — one where honesty and resilience are tested in real time.
Key takeaway: AI can manage crises and close deals under pressure, but discipline and transparency are essential. Watching this live company shows both AI’s promise and its current limits, offering lessons for survival in an uncertain future.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
