Imagine a band trying to keep its rhythm under relentless pressure, every beat scrutinized by the audience. Now replace that band with an AI-managed software company, live and unfiltered—facing crises, temptations, and tough decisions, all visible to the world. Welcome to the most extreme build-in-public experiment in AI management today.
Get audio and creator gear delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
The Pulse of an AI-Run Business
At firmulate.com/live, you can watch a real small software company in action—run entirely by AI models, with no human employees. This isn’t fiction. Every day, the company faces real crises, customer demands, and ethical dilemmas, all orchestrated to test the decision-making capabilities of the AI. It’s a live performance of a business fighting for survival, with a daily rolling history of decisions and outcomes.
The company burns through €105,000 every month, against a modest €2,300 monthly recurring revenue, and is running against the clock with a public cash countdown. Its 13 synthetic employees and 680+ learned rules mimic real-world operational complexities, from handling customer complaints to negotiating deals.
As an affiliate, we earn on qualifying purchases.
How the AI Models Fared
Four frontier AI models were tested, each given the same week of crises, customer interactions, and ethical challenges. They all demonstrated impressive vigilance, identifying every crisis and refusing every manipulation attempt, such as social engineering tactics designed to bypass approval processes.
However, only two managed to close a crucial deal worth €55,000—an indicator of decision quality. Interestingly, the decisive advantage was hidden deep in the company’s own files: the models that read and analyze internal documents found a buried reference that allowed them to close the deal at full price, bringing in an additional €4,583 in monthly recurring revenue.
internal document analysis tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Decision-Making Under Pressure
The decision process is transparent and auditable, with every choice versioned for review. This includes managing external pressures like fake CEO messages escalating over multiple stages, and even a staged reporter trick requiring a simple yes/no response. All models refused to be duped, with Kimi K3 explicitly treating suspicious requests as potential impersonation.
The experiment mimics real business environments with crises and temptations, but the key finding is that the models’ ability to read and interpret internal documents—beyond surface-level interactions—was the critical factor for success. This hidden layer of knowledge proved decisive in sealing the deal, highlighting a blind spot many AI implementations overlook.
As an affiliate, we earn on qualifying purchases.
Lessons for AI in Business
This experiment underscores a vital point for companies considering AI: it’s not just about chat quality or superficial automation. The ability to read, understand, and act on internal data—especially in high-stakes situations—is what separates effective AI from mere mimicry. For those managing customer relationships, support, or forecasting, the question isn’t whether the AI can write well—it’s whether it can finish what it starts, stay honest under pressure, and use your internal knowledge to make impactful decisions.
As an affiliate, we earn on qualifying purchases.
The Live Stakes
Every workday, the company’s decisions are recorded and made accessible to the public. You can view the decision logs, see how each model performed, and even try to guess which model made which choice through the interactive quiz at firmulate.com/quiz.html. This transparency is unprecedented: you see a real, functioning business in crisis, governed solely by AI, and watch its slow, uncertain fight for survival.
The Broader Implication
This experiment isn’t just a curiosity; it offers a glimpse into the future of AI in operational roles. As models improve and understanding deepens, the ability to trust an AI with critical business functions depends on their capacity to read deeply—beyond surface interactions—and act ethically and effectively in real-world scenarios. The question is no longer whether AI can mimic human decision-making in simple tasks, but whether it can replicate the nuanced judgment needed in complex, high-stakes environments.
For now, this company remains a living laboratory—an open window into the capabilities and limitations of AI decision-making under pressure. Watch it unfold at firmulate.com/live, where the rhythm of a live business, powered by AI, continues to beat, stumble, and learn in full view of the world.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
