What if artificial intelligence was given only AED 5,000 as capital and asked to establish a project that would generate income within a month? Would it take risks in search of quick returns or preserve the capital? And does it think like an entrepreneur or do its ideas flash on the screen but falter at the first test in the market?

Al Bayan put three of the most famous AI models to the same challenge: ChatGPT, Gemini and Claude without specifying the type of project or how to spend the amount. How did they respond? The answers were completely different: content for companies, an online store for a trending product, and mobile car washing. Not only was the difference in the project but also in dealing with money and risk; the first preferred to keep the liquidity, the second spent the capital in search of quick growth, and the third bought equipment that could continue generating income.

ChatGPT: The client before spending

ChatGPT chose to create a small studio for producing content for small businesses. It suggested a monthly package priced at AED 1,250, including eight short video clips, eight ideas and scripts for the content, and a publishing schedule. It decided not to spend anything initially and set a ceiling of AED 2,500 when needed: AED 600 for tools, AED 500 for models and landing page, and AED 700 for marketing, and AED 700 for transportation and expenses while keeping the other half of the capital.

His plan is to reach out to 100 projects in order to get four clients, with revenues of AED 5,000, which could rise to AED 10,000 with eight clients. However, he did not consider these figures as guaranteed and preferred to test the existence of a client willing to pay before spending.

“Gemini”: Half the money for advertising

While Gemini chose to create a small online store for a popular product in the UAE market, using artificial intelligence to analyze trends and find a product that combines high demand with less competition.

He spent almost all of his capital (AED 5,000): AED 2,500 on advertising on TikTok and Instagram, AED 1,500 for the product and delivery, and AED 1,000 to create a store or landing page. He also used artificial intelligence to set up descriptions, advertisements, and marketing videos.

Gemini expected sales of between AED 9,000 and AED 12,000 within a month, then reinvest the return in expanding the project. However, these figures represent total sales rather than profits; product costs, advertising, delivery, and operating expenses must be deducted.

"Cloud": Work communication equipment

"Cloud" has moved away from the digital world and chosen mobile car washing, considering that it can generate income since the first week without needing inventory. Customers can be converted into recurring subscriptions. He allocated AED 3100 for the project: AED 1800 for equipment, AED 800 for advertising, and AED 500 for cleaning supplies and operation, leaving AED 1900. He suggested starting with 5 cars at a discounted price, using "before and after" images in marketing, targeting a specific area with advertisements, and at AED 50 per wash, he estimated revenues of around AED 3600 per month. Adding this to the remaining AED 1900 brings the total available amount to AED 5500, along with equipment that can be used later. He also bet on an improved return in the second month and warned about the need to check the permits regulating the activity.

The crucial link

The previous answers confirm that artificial intelligence can propose a project, calculate costs, and map out the path to profit. However, humans remain the critical link; they are the ones who decide, take risks, execute, and interact with the market. It is humans who turn these plans into real money, especially since these plans are merely suggestions that may succeed or fail when tested in the market. Human expertise and ability to make decisions and adapt to changes remain the decisive factor in turning artificial intelligence ideas into a real project that generates returns.

Real-World Experiments

Real-world experiments reveal that relying on artificial intelligence does not guarantee success. Last July, the Wall Street Journal reported that entrepreneur Ben Brooka took his company—which he manages without employees and uses AI for various tasks—to around 10,000 paying customers and was on track to generate $10 million in revenue within the year.

Conversely, a study published in Management Science this year involving 640 entrepreneurs in Kenya, some of whom used a business consultant based on the GPT-4 model, found that high-performing entrepreneurs benefited by more than 15%, while low-performing entrepreneurs saw their performance decline by 8% to 10%. This confirms that AI can enhance project performance but does not guarantee success. The way humans utilize its advice remains a critical factor.