What AI Tells Us About Trust and Follow-Through in Business

Why Trust Matters More Than Words in Business

Imagine you’re running a woodworking shop, and an AI assistant suggests a new tool or technique. It’s quick, convincing, and seems authoritative. But can it follow through on what really matters — completing projects on time, sticking to budgets, and maintaining your reputation? In the world of AI, like in business, the true test isn’t just what a tool can say — it’s whether it can deliver under pressure.

The Crucible Experiment: Putting AI Models to the Test

Recently, four advanced AI models were challenged with the same real-world situation: managing a small software company during its most tumultuous week. This wasn’t a simulation — it was a live, watchable experiment where every decision was recorded and scrutinized. The goal? To see which AI could identify crises, resist manipulation, and ultimately close a crucial €55,000 deal.

The Models in the Race

• gpt-5.6-sol 95: The top performer, which not only found hidden information in company files but also successfully closed the deal, earning full marks for thoroughness and integrity.
• Kimi K3 93: A newcomer with a clean discipline record, which managed to close at full price, showing strong integrity and focus.
• Sonnet 5 88: Also closed the deal but with some lapses in process discipline.
• Fable 5 77: Demonstrated the best rule adherence but failed to execute the deal, leaving the company’s analysis unimplemented.

What These Results Reveal

While all models identified the crises and refused manipulative tactics (like fake CEO messages), only two completed the job and signed the contract. The decisive factor? In-depth reading of internal company files — buried two documents deep — proved critical. The models that read these files won the full deal, adding around €4,583 MRR in value.

The Hidden Weaknesses

Interestingly, even the most disciplined models showed vulnerabilities. For example, Opus 4.8, which was the most comprehensive participant, left the deal unexecuted, illustrating that thorough analysis doesn’t always translate into action. The experiment underscores a vital point: trustworthiness under pressure, and the ability to follow through, are invisible in typical chat demos but are what truly determine value.

The Real-World Implication for Businesses

For companies contemplating AI integration, especially in customer support, sales, or decision-making roles, this experiment offers a stark lesson. It’s not enough for an AI to produce convincing responses. The key question is whether it can finish what it starts, resist shortcuts, and act in accordance with the company’s best interests — even when manipulated or under stress.

Measuring What Matters

The current league table from the experiment ranks these models based on their performance:

• gpt-5.6-sol 95: Achieved the highest score, closing the deal after uncovering hidden company files.
• Kimi K3 93: Also closed the deal, showing the cleanest discipline of the field.
• Sonnet 88: Managed to close but with some process slips.
• Fable 77: Demonstrated excellent rule discipline but failed to execute the deal.

This ranking isn’t just about chat quality — it’s about execution, trust, and follow-through. An AI’s ability to read critical internal documents can be the difference between closing a deal and leaving money on the table.

Take Action: Run Your Own Business Wargame

Interested in testing your AI tools? Firmulate offers a way to simulate your own business scenarios, without risking real systems. Run a read-only export of your processes through the same rigorous challenge and see how your AI performs under real-world stress, with no risk to your operations. Find out whether your AI can stay honest, read deeply, and execute reliably — before you hire it for critical tasks.

Explore Further

Visit firmulate.com to watch the live experiment or try the interactive quiz at firmulate.com/quiz.html. Understand how AI models perform in actual management scenarios, and how that performance can impact your business outcomes.

Key Takeaway

In AI-driven business decision-making, the real measure of effectiveness isn’t just what the models say — it’s whether they can finish what they start, read critical information, and stay honest under pressure. Trustworthiness and follow-through matter more than ever, and they’re only visible through rigorous testing like the Firmulate live challenge.

Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html

Powered by Thorsten Meyer AI