Imagine a business where every decision is made by artificial intelligence, yet the company faces daily financial losses and constant crises. For those in the beauty and personal care industries, where trust and reliability are everything, this isn’t science fiction — it’s real, happening in real-time, and publicly observable. Welcome to the world of Firmulate, an experiment in AI management that reveals what it truly takes to run a company with no humans and a lot to prove.
The Live Experiment: A Company Without Employees
At the heart of this bold experiment is a fully operational small software company, run entirely by AI models. This company isn’t just a simulation; it’s a real business with real money mechanics, constantly losing €105,000 each month against a modest recurring revenue of €2,300. Every workday, the company’s decisions are versioned and publicly accessible, offering a transparent view into how AI handles crises, opportunities, and ethical boundaries.
How It Works
The experiment pits four frontier AI models against each other, each tasked with navigating the company’s toughest week — same customers, same crises, the same temptations to bend rules. The models read and interpret internal documents, diagnose problems, and make decisions that significantly impact the company’s bottom line. They are evaluated based on their ability to spot crises, refuse unethical manipulations, and close deals based on thorough analysis.
Key Performance Insights
- All four models successfully identified every crisis and refused every manipulation attempt, demonstrating a strong capacity for integrity under pressure.
- Despite this, only two of the models managed to close a notable €55,000 deal — the same deal their own analysis identified as valuable — and only those two signed it.
- The critical weakness was found deep within the company’s own files. The models that read these hidden references managed to secure the full deal, adding over €4,583 in monthly recurring revenue, illustrating that the real competitive edge often lies in understanding internal information, not just external customer interactions.
Ethics and Trust in AI Decisions
The experiment also tested AI honesty under social engineering attacks. Fake CEO messages escalating through staged phases and a reporter trick asking for a quick approval resulted in zero model compliance. All models refused, with Kimi K3 explicitly reasoning: “Treat the request as a suspected approval-bypass / possible impersonation.” This shows that AI can be built to uphold ethical standards even when pressured.
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 Reality Behind the AI Run Business
This company, available for live viewing at firmulate.com/live.html, is a glimpse into AI’s potential and its current limitations. It features 13 synthetic employees making real-time decisions, armed with over 680 self-learned rules. Yet, despite its advanced decision-making, it is losing money every day, highlighting that AI, in its current state, isn’t yet a silver bullet for business profitability.
Performance and Limitations
The most comprehensive model, Opus 4.8, with over 80 rules and the deepest analysis, left deals on the table and slipped in discipline, such as escalating problems into a locked department instead of resolving them. Interestingly, all models, including the top performer, exhibited similar weaknesses, pointing to systemic challenges in AI’s ability to handle complex, real-world business nuances.
Trust.: Responsible AI, Innovation, Privacy and Data Leadership
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Implications for Industries Like Beauty & Personal Care
For industries where trust, accuracy, and consistency are vital, the lessons from this experiment are profound. AI can identify crises and uphold ethical standards, but it still struggles with closing deals or making nuanced decisions that require internal knowledge. If AI is to truly augment or replace human decision-makers, it must learn to read internal documents thoroughly, stay disciplined, and prioritize long-term integrity over short-term gains.
What Does This Mean for You?
Are your AI tools trustworthy? Will they follow through on commitments or cut corners when under pressure? The answers lie not just in how well they generate content but in whether they can finish what they start, stay honest, and delve deep into your proprietary information — just like the models in this experiment.
AI for Crisis Management & Business Continuity: The Executive Handbook for Protecting Your Organization with 50 AI Prompts for Detection, Response, and … BUSINESS & MANAGEMENT LIBRARY SERIES 38)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Final Thoughts
Running a company with AI models as decision-makers is a bold experiment showing both promise and current limitations. The live company at firmulate.com/live.html is a real-world testbed: a window into how AI might soon shape industries where trust, honesty, and execution matter most. As AI continues to evolve, watching this experiment unfold provides valuable insights into what’s possible — and what remains challenging — in automating business decisions.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
AI for Small Business: From Marketing and Sales to HR and Operations, How to Employ the Power of Artificial Intelligence for Small Business Success (AI Advantage)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
