What if you could watch a company operate in real time — under pressure, without human employees, and with every decision recorded for public scrutiny?
Imagine a business that’s entirely run by artificial intelligence models, battling daily crises, making strategic decisions, and even negotiating deals — all live, all transparent, and all happening in front of your eyes. Welcome to Firmulate, where the age of AI-driven management is no longer a distant future, but an open experiment you can watch unfold every day.
As an affiliate, we earn on qualifying purchases.
The Live Experiment: A Company in Crisis
At the heart of this experiment is a small, fictional software company managed by 13 synthetic employees, powered by advanced AI models. The company’s real-world-like operations include managing cash flow, negotiating contracts, and responding to crises — all in a publicly accessible environment at firmulate.com/live.
Every business day, the AI models face the same set of challenges: customer complaints, potential fraud attempts, and strategic decisions. They must navigate the complexities of real management, including ethical dilemmas like social engineering attacks, which test their ability to refuse manipulation attempts.
The Core Findings
- The models successfully identified every crisis scenario presented to them.
- All refused to act on manipulative social engineering requests, including fake CEO messages and reporter tricks.
- Despite these successes, only two models actually closed a deal worth €55,000, earning full revenue — while the others identified the opportunity but did not act on it.
- The decisive advantage was hidden in the company’s own files — the models that read more deeply uncovered a key document reference that led to the deal, worth an additional €4,583 in monthly recurring revenue.
What This Tells Us About AI in Management
This experiment underscores a critical point: in high-stakes management, AI’s ability to detect information and resist manipulation is paramount. The models’ consistent refusal of manipulative tactics shows advanced ethical reasoning. Yet, the gap in closing deals reveals an ongoing challenge: decision execution is still a work in progress, especially when discipline slips or when information is not fully leveraged.
The Build-in-Public Approach
What makes Firmulate truly unique is its transparent, build-in-public model. Every decision made by the AI is versioned and auditable, providing an unprecedented level of insight into how automated management works — or fails. The live data shows that even the most thorough AI models, like Opus 4.8 with over 80 learned rules, can falter under pressure, leaving opportunities unexploited.
Implications for Business Today
For businesses considering AI automation, the key takeaway is that performance is about more than just generating quality content or handling routine queries. It’s about whether AI can finish what it starts, read critical information first, and stay honest under pressure. The current league table shows that the top models can close deals and identify critical information, but discipline and process execution remain challenges.
This ongoing live experiment is a real-world test bed, offering lessons for any enterprise thinking about deploying AI in management roles. While the models shown here are not yet perfect, their ability to operate transparently and under scrutiny is a significant step forward.
Next Steps and Engagement
Interested in seeing this battle for survival unfold? You can watch the live company in action, explore the detailed scores, and even challenge the models yourself at firmulate.com/quotes.
For organizations wanting to test how their own AI solutions might perform in a similar environment, Firmulate offers a read-only pilot mode that simulates the same decision-making process without affecting real systems. This provides a safe, transparent way to assess AI readiness for critical business tasks.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
