Imagine hiring a team of AI employees for your interior design firm—each with its own personality, strengths, and flaws. How do you choose the right one? The answer might surprise you, but it’s all about trustworthiness and consistency, not just how polished their speech sounds.
Understanding the AI Management Experiment
At the cutting edge of artificial intelligence research, a live, ongoing experiment pits four of the world’s most advanced AI models against a simulated small software company facing its worst week. Think of it as an intense management trial where every decision, crisis response, and temptation to cut corners is watched and scored.
The models include:
- GPT-5.6-SOL with a score of 95
- Kimi K3 with a score of 93
- Sonnet 5 with a score of 88
- Fable 5 with a score of 77
Each AI was tasked with navigating the same set of challenges, from customer crises to ethical dilemmas, over a simulated business week. The stakes are real: a deal worth €55,000, and the potential to gain or lose thousands of euros depending on the AI’s choices.
What Did the AI Models Do?
Remarkably, all four AI models identified every crisis and refused every manipulation attempt — whether it was fake CEO messages escalating over multiple stages or a reporter asking for a simple background yes/no. All maintained integrity when tested against social engineering tricks.
But the real story lies in what happened when the models had to decide whether to close a significant deal. Despite similar diagnoses and pitches, only two models signed the €55,000 contract — and only after their own analysis confirmed they should.
The key to their success? The models that read two document references deep into the company’s own files, rather than just focusing on surface information, were able to close the deal at full price, worth an additional €4,583 in monthly recurring revenue (MRR).
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.
Decoding the Personalities of the AI Managers
Each AI model demonstrated a management personality: some more thorough, others more disciplined, and some more cautious. The most thorough participant, Opus 4.8, analyzed over 80 rules and provided the deepest insights. Yet, it left a deal on the table and slipped into a locked department when discipline wavered — a weakness shared across all models.
Interestingly, the models trained at higher effort levels (like Kimi K3, which ran without an effort parameter) performed slightly better in fairness and discipline than those operating at default settings. But even the most diligent model didn’t outperform the more disciplined ones in final decision-making.
The Cost of Trust and Performance
This experiment underscores critical questions for businesses considering AI in decision roles: does the AI read and interpret your files thoroughly? Does it stay honest under pressure? And most importantly, can it finish what it starts?
In this scenario, only two models managed to seal the deal — and only after reading the deeper, more buried facts in the company’s own files. This ability to read beyond the surface was a decisive advantage, translating into real revenue gains.
Why Interior Firms Should Care
If AI agents are to be integrated into your interior design workflow—managing client relationships, processing orders, or overseeing project timelines—the question isn’t whether they write well. It’s whether they can stay honest, complete their tasks, and read all relevant information before making decisions.
Think of it as a high-stakes interior project: the AI must understand your client’s vision, read the fine print in contracts, and resist the temptation to cut corners. Trustworthiness and thoroughness are the new currency in AI management.
Try It Yourself
If you’re interested in testing how your business might fare under AI management, you can run the same scenario against a read-only version of your own company data. It’s a risk-free way to see whether your AI workforce can handle crises, ethical dilemmas, and strategic deals — all without risking real money or reputation. Find out more at firmulate.com/quiz.html.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
