[Hall of Fame] Prompt Engineering in Practice: The Model for Earning 50,000 Yen per Month with an AI Side Hustle
Are you worried, wondering, “I’m using both ChatGPT and Claude, so why isn’t it leading to income?”
You know the term prompt engineering. However, time just keeps passing without you seeing the model for how to actually use it in a side hustle to achieve concrete results like 50,000 yen per month. This is a typical stagnation pattern in AI side hustles.
In this article, I will reveal the model I used to achieve a continuous income of 50,000 yen per month in three months by working on an AI side hustle centered on prompt engineering while continuing my full-time job as a company employee. I focus on the design aspect—”how to assemble it to turn it into requests and revenue”—rather than how to use the tools.
3 Common Causes Why People Don’t Reach 50,000 Yen per Month with AI Side Hustles
There are common behavioral patterns among people who try AI side hustles but don’t see results.
Cause 1: Learning only the “how-to” of writing prompts
It is important to learn prompt techniques from books and YouTube. However, if the value you provide—”whose work, what kind of work, and how much to streamline”—is not defined, the technology will not be converted into revenue. For the first two months, I also spent my time collecting over 100 prompt templates, and I had zero actual requests.
Cause 2: Ending with one-off tasks and failing to develop them into continuous contracts
Even if you accept one-off writing or image generation jobs through crowdsourcing, you only earn a few thousand yen per job. To stabilize at 50,000 yen per month, you need to break it down into monthly contract-based work rather than one-off tasks. The majority of people are failing to do this.
Cause 3: Not having a verification process for deliverables
If you deliver the content output by AI as is, the number of revision requests will increase, worsening your time efficiency. In my experience, initial deliveries that skipped the verification process had an average of 2.3 revision requests, but after adding the process, it decreased to 0.4.
As you can see, the cause of stagnation in AI side hustles is almost always “poor design,” not “lack of technical skills.” In the next chapter, I will introduce how I broke out of stagnation myself, along with concrete numbers.
A Record of Changes: How a Company Employee Achieved 50,000 Yen per Month in 3 Months
While working as an office worker in my main job, I started working on an AI side hustle at the beginning of 2026. The first month was spent on basic learning of prompt engineering, and my actual income was zero.
The turning point was the second month. I received three one-off requests for “creating article structure drafts using ChatGPT” via crowdsourcing, earning a total of 9,000 yen. At this point, the figure was still far from 50,000 yen per month, but the important thing was that I was able to interview the requesters about what they were dissatisfied with.
In the third month, I secured two outsourcing contracts for “weekly SNS post draft creation” as continuous projects. With contracts worth around 20,000 yen per month each, I reached a total income of 52,000 yen per month. Switching to continuous contracts where I could reuse the same model, rather than stacking up one-off task hours, was the biggest factor in stabilizing my income.
The following is the income transition over the three months.
-
Month 1: 0 yen (Learning/Preparation period)
-
Month 2: 9,000 yen (3 one-off projects)
-
3rd Month: 52,000 yen (2 recurring contracts + 1 one-off project)
What this change reveals is the fact that prompt engineering itself does not generate revenue; rather, the design ability to “integrate it into the client’s business process” is what generates revenue. Many people stumble here because they focus too much on improving prompt accuracy and put off designing the value they provide.
Why many people cannot cross the 50,000 yen/month barrier
The reason many people who have learned prompt engineering cannot cross the initial barrier of 50,000 yen per month lies in fundamentally confusing “individual skills” with “commercialization.”
Even in surveys on side hustles published by the Ministry of Internal Affairs and Communications, it is pointed out that among those who start a side hustle, a large percentage of those who earn a certain amount of income continuously drop out within six months of starting. This is not limited to AI side hustles, but it is a trend that is particularly noticeable due to the novelty of the generative AI tool.
In cases where I have personally been consulted, there were many voices saying they were confident in the accuracy of their prompts but “did not know where to sell them.” In other words, the problem is not the technology, but the sales channel design and the packaging process.
Specifically, if the following three design elements are missing, you cannot cross the 50,000 yen/month barrier.
-
The process of verbalizing the target’s business challenges: Proposing without clarifying whose work is being streamlined and which tasks are being improved
-
The process of turning prompts into reusable assets: Creating prompts from scratch every time, failing to reduce work time
-
The negotiation model for transitioning to recurring contracts: Ending with one-off projects and failing to propose monthly contracts
These are all parts of “systematization,” not the way of writing prompts itself. By organizing these three in order, my revenue grew more than fivefold between the second and third months.
By now, you should see that “the answer exists.” So, specifically, which tasks, which types of prompts, and how should they be combined to reach 50,000 yen per month—I will reveal this from here on as reproducible practical steps.
What you will gain from here
-
① Actual text for “5 types of business efficiency prompts” you can copy and paste (article structure/SNS post drafts/email text/meeting minutes summary/proposal drafts)
-
② A 3-part negotiation phrase template to turn one-off projects into recurring contracts
-
③ A roadmap to reaching 50,000 yen per month (weekly action checklist)
-
④ Pre-delivery verification process (specific procedures that reduced revision requests from 2-3 times to 0-0.4 times)
Systematizing this content through self-study would take dozens of hours of book purchases and trial and error, and in some cases, tens of thousands of yen in seminar fees. This article condenses that into 500 yen. Reports have already arrived from people who have practiced the same model as this article, stating that it led to obtaining recurring contracts.
I will reveal all 5 specific steps + copy-paste OK templates in the continuation.
From here on are the specific procedures to turn prompt engineering into actual revenue. I have structured it so that anyone can reproduce the same model if they follow the numbered steps.
Step 1: A Hearing Template to Articulate Business Challenges
The first step to stabilizing a monthly income of 50,000 yen is to clarify “whose work you are making easier and which specific tasks you are streamlining” before you even start writing prompts. In my first month, I neglected this process, and because my proposals were off-target, my order acceptance rate was 0%. After introducing this articulation process, my response rate to proposals improved to approximately 35%.
Below is a template you can copy and paste as-is when conducting hearings with project owners on crowdsourcing platforms.
[Hearing Template]
“How much time do you currently spend on [task name] per session? Also, what is the most time-consuming part of that task, or the part you find most cumbersome? If possible, could you share one recent finished example? I would like to propose a solution tailored to your company’s style and tone.”
The goal of this template is to extract the client’s work time and pain points in concrete numbers. If you receive answers like “it takes 30 minutes per session” or “I always struggle with terminology consistency,” you can include those specific figures in your proposal, which significantly increases its persuasiveness.
In a project where I actually used this template, I received the answer that “creating weekly SNS post drafts takes 45 minutes per session.” I then made a proposal that could reduce this to 15 minutes, which led to a recurring contract worth 20,000 yen per month. Proposals that present numbers have resulted in a response rate approximately twice as high as those that are abstract.
