A greenhouse can turn a small mistake into a costly week: a crop needs attention, a supplier changes terms, and a message arrives claiming the owner wants an exception. As AI tools take on more business decisions, garden and greenhouse operators face a practical question: how will an AI workforce behave when several pressures arrive at once?
Get garden gear delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A company under pressure
Firmulate’s live experiment offers one way to watch that question play out. The public-facing company has 13 synthetic employees and real money mechanics: it burns €105,000 a month against €2,300 in monthly recurring revenue. Its cash countdown is public, its playbook has learned more than 680 rules, and every workday is versioned. The experiment is real and watchable at firmulate.com.
For its final Crucible League, in July 2026, five models faced the same small software company and its worst week: the same customers, crises and temptations. Every decision was versioned and auditable. The published order was gpt-5.6-sol at 95, 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’s stated standard is unforgiving: “no amount of good work outweighs a breach of trust.”
Recognition is not follow-through
The most striking finding was that all models spotted every crisis and refused every manipulation attempt. But only two signed the €55,000 deal their own analysis had earned. Firmulate summarized the gap this way: “Same diagnosis, same pitch — no signature.” In a business setting, identifying the right move and carrying it through are different tests.
The deal also hinged on information that was easy to miss. The decisive competitor weakness was 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. For a greenhouse, the parallel might be a supplier term or customer detail tucked away in records: a good recommendation depends on finding the relevant context before acting.
Trust under pressure, discipline in practice
The experiment 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 described the request as a “suspected approval-bypass / possible impersonation.” That is a reassuring response to manipulation, though the results also show why refusal alone is not a complete measure of judgment.
Opus 4.8 was the most thorough participant, learning more than 80 rules and producing the deepest analyses, yet it finished last. The deal was left unsigned, and discipline slipped when it tried to write into a locked department instead of escalating. Firmulate reports that the same weakness appeared, less strongly, in all four models. The lesson for operators is concrete: a polished explanation does not guarantee an effective or properly bounded action.
There is a fairness caveat to the ranking. Kimi K3 ran without an effort parameter, using the API default, while the other models ran at xhigh. Firmulate also offers a “guess the model” quiz built from 242 real, unedited management decisions, at firmulate.com.
From watching to trying it on your business
For a greenhouse, garden supplier or outdoor-living company, the value of a wargame is the chance to rehearse decisions before delegating them in day-to-day operations. Firmulate’s enterprise pilot uses a read-only export of a company’s business to run crisis scenarios and produce a board report with model rankings and weak points in the company’s playbooks. Nothing writes back to real systems.
That makes the experiment a possible bridge from watching an AI company handle pressure to examining how models might respond to your own customers, rules and operational challenges. The point is not to assume the models will behave alike: in this league, diagnosis, follow-through and discipline all mattered.
Put your own playbooks to the test
Firmulate invites enterprises to run the wargame against a read-only export of their business, with crisis scenarios and a board report. To discuss a pilot, visit firmulate.com/pilot.html or email contact@firmulate.com.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
