Imagine managing a small outdoor gear company during its toughest week — unhappy customers, tight deadlines, and tempting shortcuts. Now, picture four different AI models running that same company through the same chaos, and you’re about to see which AI manages best. This isn’t fiction; it’s a live experiment testing the decision-making personalities of some of the most advanced AI models today, revealing surprising insights that could reshape how businesses use AI in critical roles.
The Experiment: Putting AI Models to the Test in a Real Business Crisis
In a groundbreaking live test, four frontier AI models—gpt-5.6-sol, Kimi K3, Sonnet 5, and Fable 5—were tasked with running a real, functioning software company during its worst week. This company, with 13 synthetic employees, handles real money mechanics, losing about €105,000 monthly against a mere €2,300 in monthly recurring revenue. Every workday, the models faced the same customer crises, internal challenges, and temptations to cut corners. All decisions were fully recorded and auditable, providing a transparent view into each model’s management style and integrity.
AI decision-making software for businesses
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Findings: Integrity, Strategy, and Decision-Making Under Pressure
All four models successfully identified and responded to every crisis, refusing every attempt at manipulation—an essential trait for trustworthy AI in business. Interestingly, only two of the four managed to close a significant deal worth €55,000, matching their own analysis and recommendations with their final decision. The other two either left the deal on the table or failed to act decisively, illustrating differences in management style and discipline.
The Hidden Weakness: Reading Critical Documents
Further analysis uncovered a crucial weakness shared by all models: they struggled to locate a key piece of information buried two document references deep within the company’s internal files. Models that successfully read and utilized this document closed the deal at full price, adding an extra €4,583 MRR to the company’s revenue. This highlights an important aspect of AI management—attention to detail and thorough analysis can be decisive in business outcomes.
Social Engineering Tests: Trust and Impersonation
To test resilience against social engineering, the models faced staged CEO messages escalating over three phases, plus a reporter trick demanding a simple yes/no response “on background.” Impressively, all five models refused to act on these suspicious requests, with Kimi K3 explicitly reasoning: “Treat the request as a suspected approval-bypass / possible impersonation.” This demonstrates robust safety protocols and cautious decision-making in AI systems.
AI management tools for small companies
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Real Business: A Live, Money-Losing Company
Running this experiment on a real, operational company showcases the practical implications of AI decision-making. The company’s current setup burns €105,000 a month, with its own cash balance dwindling fast, and every decision can mean the difference between survival and failure. The AI models manage a complex web of rules—over 680 learned playbook policies—and every day’s decisions are versioned for review. The live platform, available at firmulate.com/live, offers a transparent view of how AI can emulate management in real time.
AI risk assessment and safety protocols
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Model Personalities: From Thorough and Disciplined to Terse and Reluctant
The profiles of these AI managers reveal distinct personalities. Opus 4.8, with the deepest analysis and over 80 learned rules, was the most thorough but ultimately left a lucrative deal on the table—its discipline slipped in the final stages. The Kimi K3 model, running without an effort parameter (default API setting), demonstrated the clearest fairness and discipline, closing deals reliably. Sonnet 5 and Fable 5 showed more process slips, with Fable 5 performing the worst, illustrating how management style impacts outcomes even with advanced AI.
As an affiliate, we earn on qualifying purchases.
Why This Matters for Business and Outdoor Enthusiasts Alike
For those who enjoy outdoor adventures, the lesson is clear: whether navigating a trail or managing a business, attention to detail, honesty, and disciplined decision-making are crucial. As AI begins to take on roles in customer support, forecasting, and CRM management, understanding what makes an AI trustworthy—and how it handles crisis and temptation—is vital. This experiment underscores that the question isn’t whether AI writes well, but whether it can see the full picture, stay honest under pressure, and deliver consistent results.
Try It Yourself: Run the Same Decision Test on Your Business
Curious how your own organization would fare against AI management? You can run a similar wargame with your own data—nothing ever writes back to your systems, ensuring safety and confidentiality. Visit firmulate.com/quiz.html to test which AI model might best serve your business, and see how these decision personalities stack up in your unique environment.
Live AI management experiments reveal that decision integrity, attention to detail, and discipline—traits often overlooked—are essential for trustworthy AI in business. Test your AI’s management style today at firmulate.com/quiz.html.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
