TL;DR
Get office and shipping supplies delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
A simulated AI team running a startup app demonstrates daily decision-making, milestones, and stalls over 44 days. This offers insights into AI management of real business processes.
An emulated AI team has been running a real startup app, GewerkTon, in a simulated environment over 44 days, revealing how AI manages daily decisions, milestones, and setbacks. This demonstration offers a rare look at AI’s capacity to run core business functions and adapt to challenges in real time, making it a significant development in AI management.
The simulation, powered by the AI Company Emulator from Thorsten Meyer AI, begins with GewerkTon’s actual initial state — one founder, an experienced site manager testing the app, and no customers. From there, the AI team, consisting of six agents across five roles—product, engineering, pilot success, business development, and finance—operates in a fully emulated environment, making decisions daily based on simulated inputs and outputs.
Over the 44 days, the AI team achieves notable milestones: winning its first pilot on day 6, shipping its first requested feature on day 16 after overcoming engineering blockages, and converting a pilot into the first paid license by day 44. The simulation also reveals periods of stagnation, such as rejected reviews and unrecorded offers, which the AI team struggles to resolve without human intervention. The founder’s directives, issued via the feed, act as short-term unblocking measures, illustrating how human oversight influences AI decision-making in practice.
The replay is a detailed, transparent record of each decision, with a live feed showing team activity, an office map indicating who is working, and a timeline allowing review of specific days. For more details, see the full simulation overview.
Implications of AI-Managed Startup Decisions
This demonstration offers valuable insights into how AI can manage core business functions, including product development, customer engagement, and financial decisions. It highlights both the potential efficiencies and current limitations of AI in handling complex, unpredictable environments. For entrepreneurs and investors, understanding AI’s decision-making patterns and bottlenecks can inform future integration of AI tools into real-world startups, potentially transforming management practices and accelerating innovation. However, it also underscores the ongoing need for human oversight, especially during periods of stagnation or crisis.
As an affiliate, we earn on qualifying purchases.
Background on AI Emulation of Business Operations
The AI Company Emulator, developed by Thorsten Meyer AI, simulates complete companies—including crises, financial mechanics, and decision points—using advanced AI models from Firmulate. This approach allows researchers and developers to observe AI behavior in a controlled yet realistic environment, providing insights into management quality and decision-making processes. The GewerkTon simulation is part of a broader effort to explore AI’s capacity to run or support real businesses, with the current run starting from the app’s actual initial state but progressing entirely through emulated decisions and outcomes.
Previous discussions around AI in management have focused on chatbots or narrow automation, but this simulation demonstrates AI’s potential to handle multi-faceted operational roles over an extended period, offering a glimpse into future possibilities for AI-driven startups or management systems.
Unresolved Questions About AI Management Capabilities
It is still unclear how well these emulated decision patterns translate to real-world startup environments, where unpredictability and human factors play larger roles. The simulation’s environment, while detailed, simplifies many aspects of actual business dynamics, such as customer behavior and market volatility. Additionally, the long-term sustainability of AI-led management remains untested beyond this 44-day simulation, and the impact of human oversight during critical moments requires further exploration.
Future Steps for AI-Driven Business Simulations
Researchers plan to extend the simulation to include more complex scenarios, longer time frames, and integration with real-world data. There is also interest in testing how AI teams perform with different startup models or in competitive environments. The goal is to better understand AI’s limitations and strengths in managing business operations, potentially leading to hybrid human-AI management systems or fully autonomous AI startups in the future. Meanwhile, observers will watch for real-world applications and pilot projects that build on these insights.
Key Questions
Can AI fully replace human startup founders?
Currently, the simulation suggests AI can manage many operational decisions, but human oversight remains essential, especially during crises or stagnation. Full replacement is not yet feasible.
What are the main challenges AI faces in managing startups?
Challenges include handling unpredictable customer responses, navigating complex decision-making under uncertainty, and responding to unforeseen crises—areas where human judgment still outperforms AI.
How reliable are the simulation results for real-world applications?
The results provide valuable insights but are based on a controlled, emulated environment. Real-world conditions may introduce variables that the simulation does not capture.
Will AI management become common in startups?
While promising, widespread adoption depends on further research, technological advancements, and evidence that AI can reliably handle complex, unpredictable environments over the long term.
What role will human founders play as AI takes on more management tasks?
Humans are likely to continue overseeing strategic decisions, managing crises, and providing creative input, with AI supporting operational efficiency and data-driven decision-making.
Source: Thorsten Meyer AI
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
