What if your next business partner was an AI that you could watch in action, every day?
Imagine a company that operates entirely without human employees, losing money every month, yet still fighting to stay alive — and you can see it all unfold online. This isn’t science fiction; it’s the live experiment conducted by Firmulate, where four advanced AI models are running a small software business through its worst week, and the results could change how you think about investing and automation.
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 Live Company That’s Watching Itself
At the heart of this experiment is a real, functioning software company with a twist: it has no human employees. Instead, it is managed daily by AI models, each making decisions, responding to crises, and attempting to close deals — all while being transparent and auditable. You can watch this company live at firmulate.com/live.html.
What makes this setup extraordinary is its transparency and openness. Every workday, the company’s decisions are versioned and logged, creating a detailed record of how AI handles real-world business challenges. The goal? Measure whether AI can not only identify problems but also pursue solutions honestly, without manipulation or shortcuts.
As an affiliate, we earn on qualifying purchases.
How Do These AI Managers Perform?
The experiment tested four frontier AI models — GPT-5.6-sol, Kimi K3, Sonnet 5, and Opus 4.8 — against the same set of crises, temptations, and customer interactions during the company’s worst week. Despite the challenges, all four models recognized every crisis and refused every unethical manipulation attempt. But when it came to closing a crucial deal, results diverged.
Only two models succeeded in signing a €55,000 deal, which their analysis had identified as the correct move. The other two failed to follow through and left potential revenue on the table, despite having diagnosed the opportunity correctly. This gap was rooted in a buried piece of information in the company’s files — a detail that the models that read deeper into the documents were able to leverage to secure the deal.
As an affiliate, we earn on qualifying purchases.
What Does This Mean for Business and Investment?
For investors and business leaders, the implications are profound. The experiment demonstrates that AI systems can accurately diagnose crises and resist manipulation — skills critical for managing real companies — but also reveal that thoroughness in reading and understanding internal data can be decisive in closing deals and generating revenue.
In other words, the quality of AI decision-making depends heavily on how deeply it can analyze internal information. The model that performed best, GPT-5.6-sol, successfully identified and capitalized on hidden details, winning the deal at full price. Meanwhile, the less thorough models missed these opportunities, leaving money on the table.
As an affiliate, we earn on qualifying purchases.
Lessons for the Future of Work and Investment
This experiment isn’t just a fun tech showcase; it’s a glimpse into the future of automated management and AI-powered enterprise. Companies deploying AI to handle customer support, sales, or decision-making need to ask: Will these systems stay honest under pressure? Will they read all relevant data before acting?
For investors, the takeaway is clear: an AI’s ability to deliver consistent, honest results in complex situations is a crucial measure of its value. The difference between signing a deal and losing it can be a buried document or a subtle nuance missed by a superficial analysis.
Built-in-Public Transparency
To make this experiment even more compelling, every decision the AI makes is openly recorded and auditable. You can see how each model responded to crises, how they handled social engineering attempts, and how disciplined they were in following internal protocols. The entire process is designed to show that AI decision-making can be transparent, reliable, and accountable.
Why Should You Care?
Whether you’re an investor, an entrepreneur, or simply interested in the future of work, this experiment underscores a vital point: the true value of AI in business isn’t just in generating slick words or quick answers. It’s in its ability to finish what it starts, stay honest under pressure, and leverage internal knowledge effectively — all crucial factors that will determine your next investment’s success.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
