Imagine a business model where every move is transparent, decisions are made by artificial intelligence, and the company struggles to stay afloat — all while being open for the world to see. For interior designers and furniture retailers, the story of this experimental company offers a glimpse into the future of AI in business management, transparency, and trust.
Building in Public: The Experiment in Real-Time Business Simulation
In a groundbreaking public experiment, a small software company is being run entirely by AI models, with a twist: it’s losing €105,000 each month against a modest €2,300 monthly recurring revenue. This company isn’t just a concept; it’s a real operational entity, visible to anyone online at firmulate.com/live. Every decision, crisis, and response is documented, versioned, and open for scrutiny.
The Setup: AI as the Company’s Decision-Maker
Four frontier AI models—each with different capabilities—are tasked with running this business through its worst week. They face the same customer issues, crises, and temptations, and every move they make is recorded and auditable. This setup allows observers to see not just what decisions are made, but how different models perform under pressure.
What’s at Stake? Performance, Integrity, and Trust
The core question isn’t whether AI can generate compelling chat responses, but whether it can manage a real business ethically and effectively. The results are revealing: all four models identified every crisis and refused every attempt to manipulate or bypass their protocols. Only two of them managed to sign the €55,000 deal their own analysis had earned — a critical measure of success in this experiment.
The Hidden Weakness: The Buried Fact
Interestingly, the decisive advantage for the successful models lay not in their immediate responses but in a deep reading of internal documents. The models that read and understood information buried two document references deep within the company’s files were the ones to close the deal at full price, adding over €4,500 monthly recurring revenue. This highlights the importance of information access and comprehension, especially in complex decision environments.
AI-Powered Business Intelligence: Improving Forecasts and Decision Making with Machine Learning
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Building Trust: Refusing Social Engineering
In addition to crises and business negotiations, the models face social engineering attempts, like staged CEO messages and reporter tricks. All five models tested refused these manipulative tactics, with Kimi K3 explicitly reasoning that such requests could be impersonation attempts. This demonstrates a promising level of ethical resistance, a crucial factor in deploying AI for sensitive business functions.
The Reality of the Live Company
This experimental company employs 13 synthetic employees, all governed by a sophisticated set of more than 680 self-learned rules. Every workday, its operations are versioned, and its cash position is publicly tracked — currently burning €105,000 monthly while earning just €2,300. The company’s live process offers a raw, unfiltered look into how AI can manage a real business under real pressures.
Performance Insights from the Experiment
- Most thorough participant: OPUS 4.8, with over 80 learned rules and deep analyses, yet it left a close deal on the table due to discipline slips.
- Scores: gpt-5.6-sol scored highest at 95, successfully identifying and closing the deal; Kimi K3 followed with 93, while Sonnet 88 and 77 struggled slightly more with process enforcement.
- All models demonstrated awareness of crises and refused manipulative social tactics, showcasing potential for trustworthy decision-making.
Implications for Business and Design
While this experiment is about a software company, the lessons extend to other fields like interior design and furniture retail. The critical questions for professionals are: Will your AI tools finish what they start? Will they read and understand your internal files? Can they stay honest under pressure? And importantly, at what cost does useful work get done?
Why Transparency Matters
For interior designers and furniture makers, transparency in processes and decision-making becomes increasingly valuable. As AI tools grow more involved in customer relationships, supply chain management, and design decisions, knowing that your AI workforce can handle crises ethically and comprehensively is vital.
See It Live and Decide
For those curious about the future of AI in business, the live site offers a unique opportunity: watch a fully operational, AI-driven company in action, see real decisions made in real time, and explore the detailed decision logs at firmulate.com/live. It’s a rare glimpse into a building process that’s both transparent and experimental, pushing the boundaries of how we understand AI’s role in everyday business.
This public AI-run company reveals that trust, thoroughness, and integrity are as crucial as technical capability. Watching it live shows how AI can confront crises and ethical challenges — vital lessons for any business integrating AI into its future.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
