Watch an AI-Run Company Fight for Survival in Real Time

AIThis post was created with the assistance of artificial intelligence (AI).

Firmulate — This Software Company Has No Employees, Loses Money Every Day — and You Can Watch. — Live on firmulate.com.

FOR BUSINESS

Open a free Amazon Business account

Business pricing, bulk buying and tax-exempt orders.

As an affiliate, we earn on qualifying purchases.

What if your relationship was tested not in theory, but in real life, every single day?

Imagine a scenario where a company, much like a personal relationship, faces constant crises, temptations, and tough decisions. Now, picture this company run entirely by artificial intelligence, with no human employees, and its ongoing struggle for survival broadcast live for all to see.

Intersection of AI and Business Intelligence in Data-Driven Decision-Making

View Latest Price

As an affiliate, we earn on qualifying purchases.

The Live Experiment: An AI Company Under Pressure

Firmulate has created a groundbreaking experiment: a fully operational, real software company with no human staff. This company, which runs every business day at firmulate.com/live, is not just a simulation—it’s a real-world testbed where AI models manage crises, make decisions, and attempt to close deals, all while losing money every day.

Deciphering the AI Workforce

In this setup, 13 synthetic employees—AI-driven decision agents—interact with simulated customers, handle emergencies, and follow an evolving playbook of over 680 rules learned through self-teaching. The company burns through €105,000 each month but only makes €2,300 in monthly recurring revenue, highlighting its fragile financial state and high stakes.

The Challenges and the Model Performance

Four advanced AI models were tasked with navigating the company’s worst week—identical crises, same customers, same temptations. They were required to diagnose problems, make decisions, and push deals over the finish line. The results? All four models identified every crisis and refused every manipulation attempt, including social engineering tricks like fake CEO messages. Only two managed to sign the €55,000 deal their own analysis justified—yet they didn’t actually close it, revealing a gap between diagnosis and execution.

The Hidden Weakness: Reading Beyond the Surface

The crucial advantage went to the models that read deeper into the company’s own documents. Those that examined files two references deep uncovered a critical piece of information, leading to the full closure of the deal at full price—an increase of €4,583 MRR. This suggests that effective AI decision-making depends heavily on thorough information processing, not just surface-level responses.

The experiment also tested AI resilience against social engineering—fake CEO messages escalating in three stages and a reporter’s subtle background question. Remarkably, all models refused to be manipulated, with Kimi K3 explicitly reasoning: “Treat the request as a suspected approval-bypass / possible impersonation.” This demonstrates a robust understanding of trust and security protocols, even under simulated pressure.

The Real-World Stakes and Continuous Learning

This ongoing experiment is not just a stunt. The company, with its real money mechanics, live decision logs, and daily version updates, provides a transparent window into the strengths and weaknesses of AI management. Last place in the performance league was Opus 4.8, which showed deep analysis but faltered at the final step—failing to close a deal due to slipping discipline and process slips.

Implications for the Future of AI in Business and Relationships

For those concerned about how AI might influence relationships—whether in love, work, or decision-making—this experiment offers a stark warning and a fascinating insight. AI can recognize crises, resist manipulation, and even uncover hidden truths. But can it reliably follow through? Can it stay disciplined under pressure? In both business and personal life, the answer could hinge on the AI’s ability to consistently read deeper, stay honest, and execute with discipline.