TL;DR
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Firmulate is running a live experiment with a synthetic AI workforce managing a software company. Despite high analysis, only some decisions lead to revenue, revealing gaps between recognition and execution. The experiment underscores challenges in AI management and decision-making in real-world business contexts. Insights into AI’s role in business survival are detailed in the original analysis.
Firmulate has launched a live experiment where a synthetic AI workforce manages an entire software company, facing real financial pressures and customer crises. This setup aims to observe how AI handles not just diagnosis but also the execution of decisions, revealing challenges in automation’s role in business survival.
The experiment involves 13 synthetic employees working in a company with a monthly burn rate of €105,000 against €2,300 in recurring revenue. For more on how AI is transforming business operations, see the original analysis. Every workday is versioned, creating a public record of decisions, actions, successes, and failures. Despite the AI’s ability to identify crises and generate detailed analysis—over 680 self-learned rules—only a few decisions resulted in actual revenue gains.
During the July 2026 testing phase, five models competed in a league, with the top model, gpt-5.6-sol, securing a score of 95 out of 100, and the lowest, Opus 4.8, scoring 73 despite producing the most analysis. Notably, thorough analysis did not guarantee better outcomes; Opus 4.8, which produced an extra 80 rules, finished last because it failed to escalate or complete critical actions. The experiment revealed a persistent gap: AI can recognize problems but often fails to translate diagnosis into decisive action.
Trust emerged as a key factor. All models refused fake CEO approval requests, showing discipline, but only some managed to follow through on critical decisions that impacted revenue. This highlights the importance of reliable AI decision-making in business, as discussed in the original analysis. The experiment emphasizes that success depends on the AI’s ability to retrieve evidence, maintain discipline, and complete work, not just analyze or diagnose.
Implications of AI’s Limited Action in Business Contexts
This experiment demonstrates that current AI systems can identify issues convincingly but struggle with execution, exposing a challenge for deploying AI in real-world business operations. For companies considering AI automation, the findings highlight that analysis alone is insufficient; the ability to act on insights is essential for survival. The live, public nature of the experiment provides a transparent view of these challenges, emphasizing that successful AI management requires disciplined execution and organizational discipline, not just intelligence.
Products Worth Considering
Ledger Flex Crypto Wallet with 2.8-Inch E Ink Touchscreen
Ledger Nano X Bluetooth Crypto Wallet
Apple MacBook Air 15-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, and Liquid Retina Display
As an affiliate, we earn on qualifying purchases.
The Growing Role of AI in Business Management
As AI tools become more integrated into organizational workflows, questions about their practical effectiveness grow. Previous demonstrations focused on isolated tasks, but Firmulate’s live experiment pushes this further by managing an entire company in real time, exposing the gap between AI’s diagnostic capabilities and its ability to complete actions. The setup reflects broader industry concerns about AI’s readiness to handle complex, high-stakes decisions in live environments, especially where financial pressure and trust are involved.
“Recognition of problems does not automatically lead to successful action, especially in complex business contexts.”
— an anonymous researcher
Unresolved Questions About AI Decision-Action Gaps
It remains uncertain whether future iterations of AI models will better translate diagnosis into effective action or if fundamental limitations will persist. The experiment shows current models often fail to escalate or complete decisions that impact revenue, but whether this is a temporary technical issue or a core challenge of AI management is still to be determined. Additionally, how organizations can best structure AI workflows to bridge this gap remains an open question.
Next Steps for AI Management and Business Integration
Further experiments are expected to test improved AI models and management protocols that emphasize disciplined execution. Companies may also explore integrating AI decision support with human oversight to mitigate execution gaps. The ongoing publication of results aims to inform industry practices and guide future development of AI tools capable of not only diagnosing but also acting effectively in complex operational environments.
Key Questions
What does this experiment reveal about AI’s practical capabilities?
The experiment shows that while AI can recognize and analyze problems effectively, translating those insights into successful actions remains a significant challenge. Many models fail to escalate or complete decisions that impact business outcomes.
Why is the gap between diagnosis and action important?
This gap is critical because effective management depends not just on identifying issues but also on executing solutions. Without reliable action, AI cannot fully support business survival or growth.
Will future AI models overcome these execution challenges?
It is uncertain. While improvements are likely, current results suggest that disciplined processes and organizational discipline will remain essential for AI to be truly effective in operational roles.
How does this experiment affect companies considering AI automation?
It highlights that AI’s value depends on its ability to complete work, not just analyze. Companies should focus on systems that combine diagnosis with disciplined execution to ensure real-world impact.
Source: ThorstenMeyerAI.com
