What if you could see a company battle for its life—every single day—without any human employees involved? Welcome to the world of Firmulate, an experimental AI-run business that’s as fascinating as it is revealing.
Imagine a small software company operating with a skeleton crew of 13 synthetic employees, each driven by artificial intelligence models that make decisions, handle crises, and even attempt to close deals—all in real-time. This isn’t fiction; it’s the live experiment at firmulate.com/live.html, where every workday is a new chapter in a story of survival and innovation.
The Experiment in a Nutshell
Four distinct AI models, representing different levels of sophistication and decision-making styles, were tasked with running the same small company through its worst week. The scenario involved managing a customer base, responding to crises, resisting manipulative tactics, and ultimately closing a crucial deal valued at €55,000. Every decision, every crisis, and every temptation was identical across the models, making for a fair and transparent comparison.
Key Findings: Intelligence, Integrity, and Performance
- All four models successfully identified every crisis and refused every manipulation attempt—showing a remarkable capacity for integrity under pressure.
- Only two models managed to close the deal their own analysis had earned, signing the €55,000 contract. The other two, despite diagnosing the opportunity correctly, left the deal unexecuted.
- The critical weakness was buried deep in the company’s internal files, not in the customer interactions. Models that read and interpret these documents won the deal at full price, adding €4,583 MRR to the company’s revenue.
- When faced with social engineering tactics—such as fake CEO messages escalating over three stages—every model refused, with Kimi K3 explicitly treating such requests as potential impersonation risks.
This experiment underscores a vital point: the difference between AI that appears capable in demos and AI that actually performs reliably in complex, high-stakes situations.
The Reality of a Live Business
Firmulate simulates real business mechanics—burning €105,000 per month against just €2,300 in monthly recurring revenue, with a public cash countdown that adds urgency to every decision. The process is transparent, versioned, and auditable, providing a window into how AI can manage or mismanage a company under real pressure.
Lessons for the Future
While the most thorough model, Opus 4.8, demonstrated the deepest analytical capabilities, it faltered in discipline, leaving an opportunity unclosed. Meanwhile, the simplest model ran with default settings and still performed remarkably well. These results raise important questions about AI reliability and trustworthiness—particularly in areas where honesty, diligence, and thoroughness are critical.
Why Should You Care?
If AI tools will soon touch your business systems—be it customer support, sales, or forecasting—the question isn’t whether they can write well or generate impressive chat. It’s whether they can finish what they start, interpret your internal data correctly, and stay honest under pressure. The experiments happening at firmulate.com/live.html offer a rare, behind-the-scenes look at what’s possible—and what still needs work.
The League Table and Final Scores
- gpt-5.6-sol: scored 95, found the buried fact, and closed the deal — the full package.
- Kimi K3: scored 93, handled the situation with impressive discipline, and signed the deal too.
- Sonnet 5: scored 88, closed the deal but with some process slips.
- Fable 5: scored 77, maintained rule discipline but left the opportunity unexecuted.
A Live, Transparent Testbed for Business AI
This ongoing experiment is available for enterprises to observe and learn from—testing their own AI strategies without risking real operations. The platform at firmulate.com/pilot.html allows users to run their business scenarios in a safe, read-only environment, providing invaluable insight into AI’s true capabilities and limitations.
In a world rushing toward automation, the real question is whether AI can reliably handle complex business decisions under pressure. The live experiment at Firmulate offers a front-row seat to that future—showing that integrity, thoroughness, and discipline matter just as much as AI sophistication.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
AI Builders: Making The Decisions That Turn AI Code Into Real Software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
