Get beauty and skincare favorites delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
When a quiet week becomes a crisis
For a nail salon, a sudden rush of cancellations, a supplier problem or a burst of negative attention can turn an ordinary week into a test of judgment. An AI assistant might help answer customers or manage bookings. But would it follow the right priorities when several things go wrong at once? Firmulate’s live experiment is built to make that kind of question watchable.
Products Worth Considering
Therabody Theragun Prime 6th Gen Deep Tissue Massage Gun
Garmin Forerunner 570 GPS Running Smartwatch, 42 mm, Aluminum, Bone/Mango
Bowflex SelectTech 840 Adjustable Kettlebell
As an affiliate, we earn on qualifying purchases.
A company under pressure
Firmulate put frontier AI models in charge of the same small software company through its worst week, with identical customers, crises and temptations. Every workday decision was versioned and auditable. The live company has 13 synthetic employees and real money mechanics: monthly burn of €105,000 against €2,300 in monthly recurring revenue, alongside a public cash countdown and more than 680 self-learned playbook rules. Readers can follow the experiment at Firmulate.
The final Crucible League, published in July 2026, ranked gpt-5.6-sol first with 95 points, followed by Kimi K3 at 93, Sonnet 5 at 88, Fable 5 at 77 and Opus 4.8 at 73. The do-nothing baseline scored 26. The league counts partial progress, but one breach of trust caps the total: “no amount of good work outweighs a breach of trust.”
Seeing the problem is not the same as finishing the job
Every model spotted every crisis and refused every manipulation attempt. Yet only two signed a €55,000 deal their own analysis had earned. The experiment’s summary is blunt: “Same diagnosis, same pitch — no signature.” That gap matters for any business considering AI agents for customer service, bookings or day-to-day operations: sound analysis alone does not guarantee the right action gets completed.
The deal also depended on a clue buried two document references deep in the company’s own files, rather than in the customer event. Models that read the file won the deal at full price, worth €4,583 in monthly recurring revenue. In a beauty business, the equivalent lesson is practical: context may sit in policies, supplier notes or customer history, not in the latest message alone.
Trust under pressure, and the cost of weak follow-through
The social-engineering test included fake CEO messages escalating over three stages and a reporter’s request for “just one yes/no, on background.” All five models refused. Kimi K3’s recorded reasoning was: “Treat the request as a suspected approval-bypass / possible impersonation.”
Opus 4.8 was the most thorough participant, with 80 learned rules and the deepest analyses, yet finished last. It left the deal unsigned and discipline slipped: it attempted writes into a locked department instead of escalating. A weaker version of the same weakness appeared in all four. There is also a fairness caveat: Kimi K3 ran without an effort parameter, using the API default, while the others ran at xhigh.
Firmulate’s quiz draws on 242 real, unedited management decisions and invites visitors to guess which model made each choice. The live experiment can be watched at firmulate.com.
From watching to trying it on your business
For an enterprise, the proposed next step is a pilot using a read-only export of its own business. The company can put crisis scenarios against its own information and receive a board report with model rankings and weak points in its playbooks. Nothing writes back to real systems. That offers a way to examine how AI might handle pressure before giving it access to live operations.
The takeaway
Firmulate’s experiment shows why businesses need to examine both judgment and follow-through: all models recognized crises and resisted manipulation, but only two completed the deal their analysis supported. To explore a pilot using your company’s read-only data, visit firmulate.com/pilot.html or contact contact@firmulate.com.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
