AI-Powered App Creation: Is Prompt-Engineering All Hokum?

Introduction

---

Hey folks! The topic of today’s blog is a bit different from my previous ones. In this post, I’ll take you through my journey of building a Paper Trading App using only Gen AI tools—without any prior programming knowledge. I’ll be sharing the efficient and smart methods I discovered along the way during the development process.

My ultimate goal here is to show how having a clear vision is far more important and valuable than execution, a concept I’ve touched on in earlier discussions.

Unleashing AI's Potential: Effective Usage Strategies

---

Paper Trading App interface for simulating stock trades, used as an example in the blog on AI-powered app creation.

For all those who think using AI is a piece of cake—spoiler alert! It’s not. And I don’t blame you for thinking that way; I felt the same before diving into this project. Before building this app, I thought, “How hard could it be, right? Just input a prompt and let ChatGPT do the rest.”

But it’s not that simple, especially for complex tasks that require more than just one response, and particularly when you’re starting from scratch without prior knowledge of the subject. That said, completing these complex tasks isn’t impossible—you just need to know how to maximize the potential of your AI tools.

After finishing the project, I realized that I could have avoided some mistakes and worked more efficiently if I had followed certain strategies from the beginning.

Here are some of the effective strategies I discovered along the way:

Discuss your idea with an AI tool:

Once an idea strikes, gather your thoughts and start a conversation with an AI tool to discuss your idea. The goal is to engage in a debate-style conversation, where there's a back-and-forth discussion about the project. This process not only offers further insights and suggestions but also helps you understand the pros and cons. I find this to be one of the most basic steps to follow when beginning any project.

As individuals, one might overlook certain ideas, but having a conversation can spark new ways to approach a task. Another benefit is gaining clarity about what you're working on; in my experience, it helps reduce confusion, keeps me focused, and provides a clearer vision for the idea.

Preparing a Vision and Workflow:

Flowchart for Project Workflow and Prompts

Write down your final idea on a piece of paper, clearly stating your vision, the issues you're tackling, and your end goal. Once you're completely sure of what you aim to achieve, the first step is to create a workflow. Consider the different features of your project, how you want to execute them, and in what order. Create a flowchart outlining your course of action. This approach will make the entire process smoother and more efficient.

Create rough prompts for different steps:

Once the flow of the project is established, it’s helpful to write a rough prompt for each step in your flowchart. The key idea here is to avoid cramming the entire project into a single prompt, which can lead to missed details. By breaking it down, the AI tool can focus on specific tasks more accurately.

I usually prefer the first prompt to be an overview of the project and the end goal, rather than diving straight into the tasks. Here's an example of what I mean:

“In this project, we’re going to create a paper trading app with essential features to simulate real-time stock trading. My goal is to guide you through each step of the development process to ensure no detail is overlooked. From the initial concept to the final execution, this guide will provide step-by-step prompts based on my workflow. As we progress, I’ll provide each prompt in detail for you to implement, ensuring all previous steps are considered.”

This method ensures clarity, structure, and accuracy in completing each task and improves the final outcome by keeping everything well-organized and methodical.

Let AI generate your prompts:

Whenever I chat with my dad about the latest advancements in AI, he’s all about, “Prompt-engineering—that's what's booming, and that’s what you ought to do!” But in my opinion, prompt-engineering is just hokum. It’s a fancy term people use to sell themselves. I’m sure you’ve all seen the flood of courses being sold under this flashy name. I agree that quality prompts are important, but why not let AI handle that as well? Isn't the question a person asks themselves better than the one asked by anyone else, because only they truly understand how they function? What I’m getting at is that AI is your Prompt-Engineer!

Choosing a Master tool for your task:

While making my prompts, one of the practices I find useful is to first choose which AI I am going to use as my master tool—the one that will receive the prompts and give me the required output. Once this selection is out of the way, I write a rough prompt and then refine it to create my final version using AI tools other than my master tool. For example, if my master tool is ChatGPT, I might craft my prompts using Gemini or Claude. My rough prompts typically look something like: “The task I want to get done is…(details). Create a good and suitable prompt for ChatGPT 4 so that it covers all the points and performs the task efficiently, just the way I intend it to.”

While you can create prompts using the same master AI, I have found this technique to be more efficient. In my opinion, the key to a good prompt is always creativity—be as creative as possible.

Conclusion Prompt Engineering

In my journey of building a Paper Trading App using only Gen AI tools, I’ve learned that while AI can be incredibly powerful, the process isn’t as simple as inputting a single prompt; rather, it involves careful planning, clear vision, and strategic use of AI tools. You might be thinking, didn’t the title suggest that prompt-engineering is just a buzzword, yet the entire article discusses prompt-engineering? Well no, the essence of this blog isn’t about mastering prompt-engineering. It’s about efficiently approaching a project by harnessing the power of Gen AI. Ultimately, it’s your creativity and vision that will guide the AI to deliver the best outcomes, proving that AI is your true partner in innovation.