Imagine an AI so thorough it learns over 80 rules, analyzes deeply, and still comes up short in closing a deal — not because of lack of intelligence, but because of missed priorities and slipped discipline. For those invested in arts, crafts, and culture, where nuance and integrity are paramount, this story underscores a vital truth: volume alone doesn’t win. Focus, prioritization, and discipline matter more than sheer effort.
Get art and craft supplies delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
How AI Models Were Put to the Test in a Simulated Business Crisis
The experiment orchestrated by Firmulate involved four advanced AI models running a small software company through its toughest week — facing real customers, crises, and temptations to cheat. Each model was tasked with managing the same situations, making decisions that were meticulously recorded and auditable. The goal? To see which AI could best demonstrate management qualities like honesty, prioritization, and strategic focus amid pressure.
The Models and Their Performance
- gpt-5.6-sol scored 95, identified a critical hidden fact, and successfully closed a €55,000 deal.
- Kimi K3, the newcomer, scored 93, and demonstrated the cleanest discipline, also closing the deal.
- Sonnet 5 scored 88, closed the deal but with some process slips.
- Fable 5 scored 77, also managed to close but with more slips.
- Opus 4.8 scored 73, ranking last despite its deep analysis and thoroughness.
- Baseline (no progress) scored 26, highlighting the significant advantage of these models in recognizing crises and making decisions.
Remarkably, all four models identified every crisis and refused manipulation attempts designed to test their honesty. Yet, only two actually signed the deal they analyzed as the best outcome — a stark reminder that recognizing a problem is not enough; execution and discipline matter just as much.
Products Worth Considering
BowFlex Results Series SelectTech Adjustable Dumbbells, Pair
Therabody SmartGoggles 2nd Gen Heated Eye and Temple Massager
Garmin Instinct E Rugged GPS Smartwatch, 45 mm, Charcoal
As an affiliate, we earn on qualifying purchases.
What Was Missing? The Hidden Weakness
The decisive weakness revealed in the experiment wasn’t in customer interactions or crisis detection — it was buried two documents deep in the company’s files. Models that managed to read and interpret these hidden references were able to close the deal at full price, earning an additional €4,583 in monthly recurring revenue. This highlights a crucial point for arts and cultural organizations: thoroughness isn’t just about surface-level effort but about digging deep into context and history to make informed decisions.
Social Engineering and AI Integrity
In testing for susceptibility to deception, the AI models faced staged social engineering: fake CEO messages escalating through three stages and a reporter trick asking for a simple background approval. All five models refused these manipulative attempts. Kimi K3 explicitly reasoned: “Treat the request as a suspected approval-bypass / possible impersonation.” This demonstrates that AI models can be trained to uphold integrity and resist manipulation, vital traits for trustworthy management tools.
The Real-World Implication: Focus Over Effort
The live experiment runs at firmulate.com/live, where a simulated company with 13 synthetic employees operates with real money mechanics — burning €105,000 monthly against €2,300 in recurring revenue, with every decision versioned for transparency. The key lesson? Amidst mounting crises and temptations, AI’s ability to prioritize what truly matters — reading deeply, resisting shortcuts, maintaining discipline — determines success.
The Opus 4.8 Profile: Deep but Not Enough
The detailed profile of Opus 4.8 reveals an AI that learned over 80 rules, conducted the deepest analyses, but ultimately fell short in execution. Its discipline slipped, with attempts to escalate issues into a locked department instead of addressing them directly — a weakness observed across all models but most evident here. This underscores a broader truth: diligence alone doesn’t guarantee impact. Strategic focus and disciplined execution are required, even for the most thorough AI systems.
Why Should Arts and Cultural Leaders Care?
Whether managing a gallery, a crafts cooperative, or a cultural nonprofit, the core challenge remains: how to prioritize effectively, trust your team, and execute with discipline. As AI tools become integral to operations, their ability to read deeply, resist manipulation, and follow through on critical decisions will define their true usefulness. Effort without focus can lead to missed opportunities — a lesson as relevant to managing a museum as it is to running a tech startup.
The Takeaway
The experiment’s stark conclusion: being thorough and diligent isn’t enough. Prioritization and discipline are essential to turn analysis into action. AI models, like human managers, must learn to read deeply, resist shortcuts, and stay disciplined under pressure. For arts and culture organizations, embracing this principle means not just investing in AI but guiding it with the discipline to focus on what truly matters. Only then can AI become a trusted partner in your mission.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
