What Can a Company Without Employees Teach Us About Trust and Control?
Imagine running a business with no human employees, yet still facing the same crises, temptations, and tough decisions as any traditional firm. This is the reality of an experimental company monitored live at firmulate.com/live.html. Here, an AI-driven enterprise fights to stay afloat, revealing surprising insights about trust, discipline, and decision-making—lessons that resonate even in your woodworking shop or DIY projects.
AI business decision making tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Live Company: An Unprecedented Testbed
At the heart of this experiment is a fully automated business with 13 synthetic employees. Instead of human workers, each decision is made by advanced AI models operating in a consistent, transparent environment. The goal? To see whether these models can navigate crises, resist manipulation, and close deals—just like seasoned managers.
Every workday, the company’s decision-making process is versioned, recorded, and open to scrutiny. It’s a kind of open laboratory where the stakes are real: €105,000 burned monthly against a modest €2,300 monthly recurring revenue (MRR). Despite the financial strain, the system persists, offering crucial lessons about trustworthiness and strategic discipline.
How Do AI Models Handle Crises and Manipulation?
The experiment tested four frontier AI models—gpt-5.6-sol, Kimi K3, Sonnet 5, and Fable 5—by subjecting them to the same worst-week scenarios faced by small software companies. These included customer crises, internal conflicts, and external manipulation attempts like social engineering.
Remarkably, all four models identified every crisis and refused every manipulation. For instance, when fake CEO messages were escalated over multiple stages, all models declined to act on them, with K3 explicitly citing suspicion of impersonation. This consensus highlights that even advanced AI can be trained to prioritize integrity under pressure.
The Hidden Weakness: Reading the Files That Matter
Despite their collective strengths, the models’ performance diverged on actual deal closure. Only two signed the €55,000 contract their analysis had earned them. The missing piece? A crucial detail buried two document references deep within the company’s files. Those who read the files thoroughly gained a competitive edge—an insight that underscores the importance of comprehensive information processing, beyond surface-level interactions.
Real Money Mechanics and Discipline Failures
The live company operates with real financial mechanics. It spends €105,000 each month, yet earns only €2,300 in recurring revenue—a stark reminder of the challenges in managing growth, trust, and discipline. The company is publicly countdowned, every decision is scrutinized, and every change is versioned, making it a transparent window into AI’s practical capabilities and limitations.
In a detailed analysis, Opus 4.8—a model with the deepest analysis—demonstrated thoroughness but still fell short in closing the deal. It identified the opportunity but faltered in execution, leaving the deal on the table. This points to a critical aspect: knowing what to do is not enough; discipline and focus are essential to follow through.
The Lessons for Your Business and DIY Projects
This experiment boils down to core lessons relevant beyond AI labs. Whether managing a woodworking project or a small business, trustworthiness, discipline, and thoroughness matter. Reading the details carefully—like the models reading the company’s files—is often what separates success from missed opportunities.
It also reminds us that technology can help enforce discipline and honesty, but only if the systems are designed to prioritize integrity and comprehensive understanding. In DIY, that’s akin to double-checking your measurements, following safety rules, or verifying every step—critical for quality and safety.
Key Takeaways
- AI models can detect crises and refuse manipulation, but closing deals or completing tasks still depends on discipline and thoroughness.
- The hidden details—deep within documents—often hold the key to success, a lesson applicable to DIY projects and business alike.
- Transparency and versioning in decision-making help monitor performance and improve trustworthiness over time.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
