firmulate.com/benchmarks.html — live view
AIThis post was created with the assistance of artificial intelligence (AI).
Firmulate — We Buried a €55,000 Fact Two Documents Deep. Here's Which AIs Did Their Homework.
Live on firmulate.com.

In a world where art and culture often hinge on perception and understanding, the true power of artificial intelligence lies in its ability to read beyond the surface. Imagine an AI that doesn’t just respond to prompts but delves into the depths of your files—finding hidden clues that can make or break a deal. This isn’t science fiction. It’s the real story emerging from a groundbreaking experiment by Firmulate, where AI models tackled a simulated week of crisis management for a small software company.

Unveiling the AI Wargame: A Deep Dive Into Decision-Making

To test AI’s true business capabilities, Firmulate orchestrated a detailed experiment. Four frontier models, including the well-known GPT-5.6-SOL and emerging Kimi K3, faced the same challenging scenario: managing a small software company’s worst week. The test was rigorous, with identical crises, customer pressures, and even manipulative tactics designed to test the AI’s integrity and thoroughness.

The Core Findings

  • All four models successfully identified every crisis and refused to be manipulated.
  • Only two models, GPT-5.6-SOL and Kimi K3, sealed the €55,000 deal based on their analysis.
  • The decisive factor was not the surface-level responses but their ability to ‘read’ and understand the company’s internal files.

Interestingly, the key weakness that cost the other models the deal was buried two document references deep inside the company’s files—information that was invisible in typical chat interactions but critical to making the right decision. The models that read and analyze these internal documents earned the full-price deal, translating to a revenue boost of over €4,583 monthly recurring revenue (MRR).

The Human-Like Challenge: Trust and Deception

Beyond crisis management, the models faced social engineering tests—fake CEO messages escalating over multiple stages, and even a reporter request asking for just one yes/no answer in the background. All models refused these manipulative tactics, with Kimi K3 explicitly reasoning that the request resembled an impersonation or approval bypass scenario.

The Reality of a Live Company

Firmulate’s live experiment involves a simulated company with 13 synthetic employees, operating with real-time mechanics. It burns €105,000 per month against a modest €2,300 MRR, with every workday versioned and transparent. This ongoing experiment tests how AI manages real-world business pressures, making the findings relevant for anyone concerned with AI’s trustworthy integration into their systems.

Amazon

AI document analysis software

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

What This Means for Arts, Crafts, and Culture

In cultural sectors—whether managing art collections, galleries, or creative collaborations—the ability of AI to read and interpret your internal documents can be a game-changer. It’s not just about generating pretty responses; it’s about making decisions based on a comprehensive understanding of your archives, contracts, or provenance records. As with the software company’s crisis, the crucial insights often lie buried in the details.

The Subtlety of AI Performance

Interestingly, the most thorough participant, Opus 4.8, with over 80 learned rules and deep analyses, did not win the deal. It left opportunities on the table and slipped into process slips, demonstrating that thoroughness alone doesn’t guarantee success. Discipline and focus—reading deeply and acting decisively—are critical, whether in a corporate setting or in managing cultural assets.

Implications for Art and Culture Managers

As AI models improve, their ability to read your internal files before responding can determine the outcome of negotiations, acquisitions, and collaborations. Understanding this capability is essential for arts organizations aiming to harness AI responsibly and effectively.

Amazon

internal document reading AI tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

The Future of AI-Driven Decision-Making

The experiment underscores a vital point: the difference between AI that merely chats and AI that comprehensively understands your business or assets is profound. The models that read deeply and stay honest under pressure are the ones that will win the trust—and the deals—in the real world.

For arts and culture leaders, this means considering how their AI tools access and analyze internal documentation. It’s a new frontier where the depth of reading can determine success or failure, authenticity or deception.

Infographic — We Buried a €55,000 Fact Two Documents Deep. Here's Which AIs Did Their Homework.
The findings at a glance — source: firmulate.com.

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

Powered by Thorsten Meyer AI


Amazon

trustworthy AI decision-making software

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Amazon

AI for business deal analysis

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

You May Also Like

From Clay to Code: Pottery Digitized Through 3D Scanning  

By transforming ancient pottery into detailed digital models through 3D scanning, this innovative process offers new possibilities for preservation and research—discover how.

Augmented Reality Filters and Social Media Art

Please provide the topic you’d like the meta description for.

Rendering the Sky, Sunsets, and Planets

Exploring how atmospheric scattering shaders recreate realistic skies, sunsets, and planetary views through raymarching and advanced rendering methods.

Why You Need a UPS: The Silent Protector of Your Studio Computer

AIThis post was created with the assistance of artificial intelligence (AI).A UPS…