GPTBoost Case-Study: Blending AI with Human Creativity in Content Creation

In the world of content creation, opinions about AI’s role are sharply divided. Some creators dismiss AI, citing its inability to replicate the nuanced human touch, resulting in content that feels superficial. Others embrace AI wholeheartedly, often overlooking the necessity of human intervention to refine and enhance the automated output. However, amidst these polarized views, there exists a middle path — a balanced approach where the efficiency of AI is harmoniously integrated with human empathy and creativity. This approach, not widely discussed, represents a fusion of technology and human insight, offering a new paradigm in content creation.

Veselina Staneva, co-founder and head of product at GPTboost, exemplifies this method. Her innovative approach leverages AI as a tool, not a replacement, in the creative process. This article delves into Veselina’s unique strategy, offering insights into how AI can complement rather than substitute the human element in writing.

Using AI to Draft Content

The journey of content creation for Veselina begins with an idea. This could be a new feature or a concept she wants to explore in her articles. Here, AI steps in as a first-draft assistant. She uses ChatGPT to generate a basic structure of her thoughts. This initial draft, while rough around the edges, lays down the framework for the article.

The next stage is where Veselina’s expertise shines — iterative editing. AI-generated content is edited for tone, technical depth, and audience engagement. This process isn’t about making minor adjustments; it’s about sculpting the AI-generated rough draft into a piece that resonates with the target audience. Whether the draft needs to be more casual, more technical, or less sales-oriented, Veselina fine-tunes the content with precision.

Adding Human Touch to AI-Generated Content

A critical aspect of Veselina’s method is the human oversight. As she succinctly puts it, “AI won’t spit out the ideal text on its own. It’s like an intern who has to be micromanaged.” This analogy perfectly captures the essence of AI in content creation — a tool that requires guidance and refinement. It’s in this phase of manual review and customization that the content truly comes to life. Veselina ensures that the final product speaks the language of her audience, connecting with them on a level that AI alone cannot achieve.

The final step in Veselina’s process is a thorough manual review. This is where her deep understanding of her audience’s language and preferences comes into play. The AI might provide the structure, but the soul of the article — its tone, its nuance, its very essence — is shaped by Veselina’s own experience and understanding of her readers.

Veselina’s Creative Process in Action

Let’s dive into an example. Say we’re crafting this article using AI. First, we give the AI context, the target audience, and any vital quotes. We start by outlining the key events and lessons learned – the article’s backbone. We then guide ChatGPT on drafting: what to highlight, omit, and how to present the terms.

I prefer starting with a text plan rather than a full draft. It’s quicker to review and tweak a plan than to edit lengthy text.

I interviewed Veselina Staneva, a co-founder and head of product at GPTboost.
We talked about using AI to generate articles showcasing new features of their product.

Her writing process looks like this:
1. She has an idea (usually related to a new feature) and wants to write an article about the idea.
2. She uses ChatGPT to generate a draft. The prompt looks like this: "Act as a tech content writer. Write about feature X, which does ABC. The text is aimed at people XYZ". In the prompt, she puts the documentation snippets that need to be included and any relevant quote she wants in the text.
3. She sends subsequent prompts to ChatGPT to make it edit the draft. In the prompt, she asks AI to make the text shorter, more technical, casual, or less sales-oriented.
4. In a separate chat, she asks AI to generate the intro and the conclusion. At this step, she asks AI to include a call to action, links, or whatever seems relevant.
5. In the end, she reviews the article manually. The manual edit is crucial because she wants to use the target audience's language, and it's easier to make the final edit manually than explain to AI how to speak correctly. This skill comes from the experience of speaking with those people.

Act as a professional content writer who is inspired by William Zinsser.

Write a plan for the article about this topic. Plan where to put the quote: "AI won't spit out the ideal text on its own. It's like an intern who has to be micromanaged." said Veselina.
Plan to write an introduction paragraph, but not as a separate introduction section.
In the conclusion section, say, "If you want to learn about GPTBoost, the product built by Veselina Staneva, visit the page" Include that in your plan.
Remember that GPTBoost is not related to the content of the article. It's a product whose creator uses AI to generate text. GPTBoost doesn't help you write anything.
The content plan
The content plan

In the initial outline, I found the introduction overly long and the conclusion’s repetition unnecessary. The Analysis section didn’t align with my vision. And importantly, I had to craft the example myself. So, I instructed AI to refine the plan, removing these elements.

Drop Section 2 Analysis, Section 3 Example, Recapitulation, and Future Outlook.

In the Opening Statement, include: "Many creators who tried to use AI for writing articles end up in two camps. Those who claim it doesn't work and it won't ever work because the result lacks human touch, and it's too superfluous. And those who use AI for everything, but often forget to edit and improve the text later. But there is a third way. The way of enlightenment where the speed of AI gets combined with human empathy and creativity."


Revise the plan.

With the outline shaping up, I faced a choice: draft the full article or further refine the plan. Opting for brevity, I focused on planning each paragraph. This approach simplifies editing and ensures each section conveys its intended message precisely.

Make a more detailed plan for each section with a content plan for each paragraph.
A detailed plan for each paragraph
A detailed plan for each paragraph

After creating the draft, I noticed a minor issue – the article consistently used Veselina’s full name. For a more natural tone, I asked ChatGPT to switch to her first name after the initial mention. Now, it was time for manual editing, the crucial step where I would infuse the article with the subtleties and nuances that AI tools often overlook.


The approach adopted by Veselina in content creation is a testament to the potential of AI when used judiciously. It’s a reminder that while AI can significantly enhance efficiency, the human element remains irreplaceable in crafting content that truly resonates. For those interested in learning more about Veselina’s work and her product, GPTBoost - Tools and APIs for OpenAI Developers, visit

Older post

What mistakes do product managers make while using AI, and what to do instead

How product managers can use AI to make their work easier without suffering the consequences of AI misuse

Newer post

Debugging, controlling OpenAI usage cost, and monitoring AI applications using Langsmith and GPTBoost

Discover how to effectively monitor AI applications, track costs, and optimize API usage with Langsmith and GPTBoost. Get insights into managing OpenAI API interactions for improved efficiency and cost control.