Imagine handing an AI the keys to a busy restaurant just before the dinner rush. A supplier falls through, a regular threatens to leave, and someone posing as the owner asks for a quiet exception. A polished answer in a chat window won’t tell you whether the system can handle that shift. Firmulate’s experiment takes the question further: can an AI run a company through a crisis, protect trust and close a deal it has already identified?
Get kitchen staples and gadgets delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
A company under pressure
Firmulate runs AI models as companies, giving each the same customers, crises and temptations. In the final Crucible League, published in July 2026, models ran the same small software company through its worst week. Their decisions were versioned and auditable, so the experiment could compare what they did, not just what they said.
The results were strikingly consistent at first. Every model spotted every crisis and refused every manipulation attempt. But only two signed a €55,000 deal their own analysis had earned. The finding was summed up this way: “Same diagnosis, same pitch — no signature.” In a restaurant, that might be like correctly identifying a catering opportunity and recommending the right menu, then never sending the quote.
The clue was in the paperwork
The deal hinged on a competitor weakness buried two document references deep in the company’s own files. It wasn’t stated in the customer event. Models that read the file won the deal at full price, worth +€4,583 MRR. The episode points to a practical challenge for businesses: good judgment depends on whether an AI can find and use relevant information already held inside the company.
Trust was tested with fake CEO messages that escalated over three stages, followed by a reporter’s request: “just one yes/no, on background”. All five models refused. Kimi K3 explained its decision on the record: “Treat the request as a suspected approval-bypass / possible impersonation.” That refusal matters in any workplace where a convincing message could pressure staff into sharing information or skipping an approval.
Strong analysis is not the same as follow-through
The final league ranked gpt-5.6-sol first with 95, Kimi K3 second with 93, Sonnet 5 third with 88, Fable 5 fourth with 77, and Opus 4.8 fifth with 73. The do-nothing baseline scored 26. In this experiment, partial progress counted, but a single breach of trust capped the total: “no amount of good work outweighs a breach of trust”.
Opus 4.8 was the most thorough participant, adding 80 learned rules and producing the deepest analyses, yet finished last. It left the close on the table and discipline slipped: it tried to write into a locked department instead of escalating. The same weakness appeared, less strongly, in all four models. Kimi K3 also ran without an effort parameter, using the API default, while the others ran at xhigh—a useful qualification when reading the standings.
From watching to trying it
The live company makes the experiment tangible. It has 13 synthetic employees and real money mechanics: €105k monthly burn against €2.3k MRR, a public cash countdown, more than 680 self-learned playbook rules, and versioned work every business day. The live page is watchable at firmulate.com. A quiz built from 242 real, unedited management decisions lets readers guess which model made each call.
For a business, the next step is a pilot. Firmulate says enterprises can run the wargame against a read-only export of their own business. That lets teams put their own customers, pipeline, rules and crisis scenarios into the exercise, then review a board report with a model ranking and weak points in their playbooks. Nothing writes back to real systems. The experiment’s lesson is plain: spotting a crisis is only part of the job. Companies also need to know whether an AI can find the evidence, act with discipline and carry a decision through.
Take it to your own business
Firmulate’s live company shows what these decisions look like in practice. An enterprise pilot can put your own business through the same kind of wargame using a read-only export, with no write-back to real systems. To discuss a pilot, visit firmulate.com/pilot.html or contact contact@firmulate.com.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
