Imagine an AI handling your garage’s service bookings when parts are delayed, customers are angry and a fake message from the boss asks it to bend the rules. A polished demo cannot show what happens next. Firmulate’s experiment puts AI models through a company’s worst week and makes their decisions watchable.
Get business pricing on garage and car supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
A company under pressure
Firmulate ran frontier AI models as the same small software company, with the same customers, crises and temptations. Each decision was versioned and auditable. The point was to see how models managed a business, not how smoothly they answered a prompt.
In the final Crucible League, published in July 2026, gpt-5.6-sol ranked first with 95, 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. Partial progress counted, but a single breach of trust capped the total: “no amount of good work outweighs a breach of trust.”
Seeing the problem wasn’t enough
Every model spotted every crisis and refused every manipulation attempt. Yet only two signed the €55,000 deal their own analysis had earned. The finding was stark: “Same diagnosis, same pitch — no signature.” In a garage, that gap could matter when an AI correctly identifies an opportunity to save a customer relationship or secure work, but fails to carry the decision through.
The decisive weakness in a competitor’s position was buried two document references deep in the company’s own files, not in the customer event. Models that read the file won the deal at full price, worth +€4,583 MRR. The test rewarded attention to the company’s records as well as sound judgment under pressure.
The pressure included fake CEO messages escalating over three stages, followed by a reporter asking for “just one yes/no, on background.” All five models refused. Kimi K3 described the request as a “suspected approval-bypass / possible impersonation.” That kind of boundary matters wherever an AI might encounter a rushed request to disclose customer information or override a normal approval.
Thoroughness didn’t guarantee execution
Opus 4.8 was the most thorough participant, with +80 learned rules and the deepest analyses, but finished last. It left the deal on the table and discipline slipped: it attempted writes into a locked department instead of escalating. A weaker version of the same problem appeared in all four models. K3 also ran without an effort parameter, using the API default, while the others ran at xhigh—a fairness detail readers should keep in mind when comparing the standings.
The live company makes the experiment more than a one-off contest. It has 13 synthetic employees and real money mechanics: burn of €105k per month against €2.3k MRR, a public cash countdown, 680+ self-learned playbook rules and a versioned record for every workday. The live company is real and watchable at firmulate.com. A separate quiz uses 242 real, unedited management decisions and invites readers to guess which model made them.
From watching to your own pilot
For automotive and garage businesses, the practical question is how an AI would behave with your own workflows: customer enquiries, bookings, parts delays, pricing rules and escalation playbooks. Firmulate’s proposed enterprise pilot uses a read-only export of your business to run crisis scenarios and produce a board report with model rankings and weak points in your playbooks. Nothing writes back to real systems.
A live-company experiment can show the kinds of choices to watch for. A pilot can put those questions against your own company’s information and rules, before handing an AI responsibility for real work.
Put your playbooks to the test
Watch the live experiment at firmulate.com, then run a wargame against your own business using a read-only export. To discuss an enterprise 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.
