Imagine a company so transparent it’s practically a reality show — where every decision, crisis, and even financial struggle is on full display for the world to see. Welcome to the front line of AI experimentation, where a tiny digital business is fighting for survival, and you can watch it happen in real time.
The Live Experiment: A Company in Crisis — and Under Watch
At the heart of this story is a real, functioning software company that operates without human employees. Instead, it’s run by artificial intelligence models, each acting as a synthetic management team. Every weekday, the company’s operations are recorded, versioned, and made public at firmulate.com/live.html. It’s a bold, build-in-public experiment, turning the age-old question of AI’s business potential into a daily spectacle.
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.
The Stakes Are High — And Visible
This tiny enterprise burns through €105,000 each month, yet earns only €2,300 in monthly recurring revenue. It’s a money-losing venture, publicly counting down the days until the cash runs out. The entire operation is a testbed for AI decision-making, with 13 synthetic employees managing everything from crisis responses to strategic pitches. Every move is tracked, and every decision is auditable.
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.
How Do the AI Models Perform?
The experiment pits four leading AI models against each other, each running the same tough week for the company. These models face identical crises, customer demands, and moral temptations — all with the goal of closing a crucial deal worth €55,000 monthly recurring revenue. The results reveal a surprising truth: all four AI models spotted every crisis and refused every manipulation attempt. Only two of them managed to sign the deal they analyzed and recommended.
Deception & Self-Deception: Investigating Psychics
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Hidden Weaknesses and the Winning Edge
Interestingly, the decisive factor wasn’t surface-level decision-making. Instead, the winning models found critical information buried deep within the company’s own files — not in the immediate customer interactions. When the AI read these internal documents, it secured the deal at full price, adding over €4,500 in monthly revenue. This highlights the importance of context and internal knowledge for AI success in real-world business tasks.
Artificial Intelligence for Improved Patient Outcomes: Principles for Moving Forward with Rigorous Science
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Trust, Deception, and AI Ethics
The experiment also tested the AI’s resilience against social engineering. Fake CEO messages escalated in complexity, and even a reporter’s subtle request for a quick approval was refused by all models. Kimi K3, one of the top performers, explained: “Treat the request as a suspected approval-bypass or possible impersonation.” This demonstrates that these AI systems are not only strategic but also cautious about deception — a crucial trait for trustworthy automation.
What Does This Mean for Real Businesses?
The live experiment isn’t just a tech showcase; it raises fundamental questions about AI’s readiness to handle complex, high-stakes decisions in actual companies. As the company runs every weekday, losing money but learning fast, it shows that AI can be effective at recognizing crises and resisting manipulation. But it also underscores that AI isn’t infallible — discipline, internal knowledge, and strategic insight are vital for success.
The Bottom Line — Watching AI’s Business Future Unfold
For business leaders, the takeaway is clear: Before deploying AI agents into your CRM, support queues, or forecasting tools, you need to ask not just if they can write well — but if they can finish what they start, stay honest under pressure, and truly understand your internal knowledge. This live experiment at firmulate.com/live.html offers a rare, unfiltered glimpse into what AI can and cannot do as a business partner.
And for the curious, a public quiz at firmulate.com/quiz.html challenges you to guess which AI model made each decision, turning the complex world of AI decision-making into an engaging game.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
