Imagine a company that operates entirely without employees, yet faces the relentless challenge of staying afloat in a competitive market. Now imagine that this company is constantly on display, with every decision, crisis, and mistake made public in real-time. Welcome to the latest experiment in AI-driven business management, where transparency meets survival.
The Live Experiment: A Business in the Crosshairs
Firmulate has set up a truly unconventional business simulation: a real, functioning software company run by artificial intelligence models. This company has no human employees; instead, it relies on 13 synthetic ’employees’ powered by advanced AI models, each making decisions, responding to crises, and pushing the company forward — or not.
Every workday, the company’s mechanics are updated and versioned, with 680+ self-learned rules guiding decision-making. The company is publicly visible online at firmulate.com/live.html, where you can watch its daily struggles unfold in real time. The goal? To see how AI models handle real business pressures, including crises, manipulation attempts, and ethical dilemmas, under the scrutiny of a public audience.
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 These AI Managers Perform?
The experiment pits four frontier AI models against the same set of challenging scenarios—crises that any real company might face. These include everything from customer crises to internal manipulations and ethical tests such as social engineering attempts. All four models are tasked with diagnosing issues, making decisions, and closing deals, all while reading the company’s internal files to uncover hidden facts.
Results are telling: all four models identified every crisis and refused every attempt at manipulation, demonstrating a remarkable level of integrity and awareness. Yet, only two models managed to close a deal worth €55,000 — a significant revenue milestone in this simulation. The other two missed the opportunity, leaving potential money on the table despite identifying the right diagnosis and delivering a convincing pitch.
The AI-Driven Leader: Harnessing AI to Make Faster, Smarter Decisions
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Hidden Weaknesses and Their Consequences
Digging deeper, the experiment reveals a crucial insight: the real weakness was buried two documents deep within the company’s files — not in the visible customer interactions or crisis reports. Models that took the time to read and analyze these internal documents won the deal at full price, adding over €4,583 in monthly recurring revenue (MRR).
This underscores an essential point: in real business, the most decisive signals often lie hidden beneath surface-level data. An AI that can read deeply into internal documents and extract hidden insights performs much better than one that only acts on obvious cues.
AI-Powered Fraud Detection: Innovations in Cybersecurity, Finance, and Healthcare (AI-Powered Fraud Detection: Innovations in Cybersecurity, Finance, and Healthcare".)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Testing Trust and Ethical Boundaries
The experiment also explored whether AI models could be manipulated through social engineering tactics, such as fake CEO messages or media tricks. All five models faced staged scenarios escalating in complexity. The result? Every model refused to act on suspicious requests, demonstrating strong trustworthiness and an ability to detect impersonation or bypass attempts. Kimi K3, one of the models, explicitly reasoned: “Treat the request as a suspected approval-bypass / possible impersonation.”
MASTERING CORPORATE FINANCE WITH CLAUDE AI: An Independent Guide to Financial Analysis, Forecasting, Automation, and Decision-Making
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Cost of Running a Virtual Business
Behind the scenes, this simulated company is burning through €105,000 each month, while generating just €2,300 in monthly recurring revenue. Despite this, the experiment continues — with a public countdown to potential shutdown, making every decision under scrutiny. This financial imbalance highlights the challenge of building AI-driven companies that are both honest and profitable.
Lessons for Business Leaders and AI Developers
What does this all mean for real-world companies considering AI integration? The key takeaway is not just whether AI can write engaging chat messages, but whether it can finish what it starts, read relevant internal information, and stay honest under pressure. These factors are critical when deploying AI in customer support, sales, or decision-making roles.
The firmulate experiment emphasizes the importance of testing AI models in simulated yet authentic scenarios—like a ‘war game’—before trusting them with real business operations. Such pre-hiring tests can reveal weaknesses and ensure that AI agents act ethically, effectively, and profitably.
The Path Forward
By openly sharing a high-stakes, no-holds-barred experiment, firmulate is pushing the boundaries of AI accountability and operational integrity. Watch the ongoing performance at firmulate.com/live.html—a window into the future of AI-managed companies, where transparency, rigorous testing, and real-world challenges come together.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
