Building a Business with AI Mentors: The Steve Jobs GPT

Scaling a business without a traditional team is a challenge many entrepreneurs face. For Yesim Saydan, a branding and communication expert in her early 50s, the solution wasn't hiring more freelancers—it was building a digital workforce.

Based in the Netherlands, Saydan runs a solo consultancy helping senior executives and established entrepreneurs grow their authority. To manage the workload, she turned to OpenAI's custom GPTs, creating a virtual team of over 17 specialized agents. Her most innovative creation, however, is a mentor modeled after one of tech's most iconic figures: Steve Jobs.

Saydan's journey to AI integration began with a traditional corporate career. After moving from Turkey to the United States for her MBA, she spent 14 years working in high-stakes environments, including a role as a project manager at Citibank on Wall Street. Her career took her to New York, Paris, and eventually the Netherlands.

She launched her own business a decade ago seeking flexibility as social media began to rise. However, scaling her consultancy presented a hurdle. Hiring freelancers often meant spending excessive time training them on her specific methodology, frequently without the same level of investment she felt.

I spent a lot of unnecessary time training the freelancers on my specific framework, and it often felt like they didn't care as much as I did.

When OpenAI launched the custom GPT feature, Saydan realized she could build the ideal team she had been looking for.

Saydan initially envisioned a four-person AI team but quickly learned that AI produces subpar results when overloaded with too many tasks. The solution was specialization. She created a distinct custom GPT for every critical function in her business, resulting in a team of more than 17 agents.

Her digital workforce includes:

• A market researcher
• A sales call analyst
• A proposal writer
• A video scriptwriter
• A LinkedIn profile evaluator using six specific pillars

The power of these agents lies not just in their creation, but in their training. Saydan creates standard operating procedure documents for each task and client, serving as the training material for her agents. She maintains client-specific agents trained on tone, goals, and past conversations, ensuring she never starts a task from scratch.

After building her employee agents, Saydan turned her attention to mentorship. She asked herself which historical or living figures she would want as mentors, landing on Steve Jobs for his renowned creativity and innovation.

She spent roughly 40 hours researching and building training assets for the Jobs GPT. The training process involved two distinct approaches:

1. Strategic Transcripts: Uploading videos where Jobs explained his product strategies and philosophies.
2. Execution Examples: Using transcripts from product launches like the iPhone and iPad to teach the AI his thought process and execution style.

The instructions began with a persona definition: "You are Steve Jobs, you have decades of experience... your most important skill is creativity, innovation, thinking out of the box." She has since expanded this library to include custom GPTs for Dan Kennedy and Elon Musk.

One of the most significant challenges with AI is its tendency to be agreeable. Saydan noted that asking a direct question like "What do you think of this idea?" usually results in the AI simply pleasing the user.

To bypass this limitation and access the trained "superpowers" of her mentors, she uses a specific prompting technique. Instead of asking for an opinion, she asks for a rating.

On a scale from one to 10, how good is this idea?

If the AI rates the idea a five, she follows up with, "OK, what would make it a 10?" This forces the AI to draw on the specific experience and frameworks she trained it with. For strategic outputs, this back-and-forth refinement process typically takes three to five rounds.

Despite her success, Saydan admits that the rapid advancement of AI is frightening. She recalls the introduction of products like NotebookLM sparking a fear that "this is going to make the entire human race obsolete."

She experiences moments of paralysis regarding the potential societal impacts, questioning if everyone will end up homeless. However, she grounds herself by acknowledging the limits of human foresight.

Her perspective has shifted from fear to integration. She believes that while AI is powerful on its own, it becomes truly magical when combined with human expertise. By using custom agents as an extension of her brain rather than a replacement, she achieves outputs that feel polished and strategic, freeing her to focus on the bigger picture. As she concludes, there is simply no turning back from this technological evolution.

Yesim Saydan's experiment demonstrates that specialization is key to effective AI integration. Rather than relying on a single, generalized tool, she built a modular system where each agent excels at a specific task.

Her approach to mentorship—training AI on specific methodologies and prompting for constructive criticism—offers a blueprint for how professionals can leverage historical expertise in the digital age. Ultimately, her experience suggests that the future of work isn't about humans versus AI, but rather humans plus AI.