Dmytro Strelchyn

Sense instead of rules

My story of launching E-shop business with AI

(in 2 months and 3 steps: Seed→Feed→Grow)

In 2024, after six years in the corporate world, I left my secure job to start my own business. This time I took a fresh approach, using a lot of AI and automation to make things faster, easier and more fun. Here’s what I learned along the way! 👇

So. My last day in Decathlon was emotional. I hugged my colleagues, signed the papers, and stepped into my first day without corporate obligations. It felt surreal. For the first time in years, no meetings, no messages, no reports, nothing…

Honestly, at that time, I even had no idea what was next. For me, it was a ticket to a long ride into uncertainty.

But it was impossible to ignore how quickly the world was changing. Even with my digital background, I felt my skills ageing. So I decided to dive into AI, automation, and no-code tools.

After spending time learning and researching, one idea kept coming back to me: What if a single person, using AI and automation, could outpace big companies in the market?

It was worth testing the idea in real life, so I rolled up my sleeves to see if it had potential or was just a fantasy.

What is Seed.Feed.Grow.

I chose to use the process taught in startup schools. With some adaptations to fit my needs, it became a 3-step process Seed.Feed.Grow.

Here’s an algorithm:

Step 1: Seed.
Run Micro-Experiments.

The goal at this early stage was to quickly and cheaply test a lot of ideas with real customers… But! Without making any sales (yes, it’s possible).

Why? Because launching any sales takes energy, time, and money. But when you’re on your own, with limited resources and uncertain outcomes, micro-experiments are key. They cost next to nothing but offer first important market insights.

I love online sales, so I researched and listed 10 product ideas that could potentially turn into a business.

I knew most ideas would fail, so I tested them all and gathered real customer data.

For each product idea, I created a simple web page with a buy button. However, clicking the button didn’t lead to a payment page. Instead, a message informed customers that the product was unavailable and they’d be notified later. I ran small online advertising campaigns to see if people would click the ‘buy’ button.

At this moment, reality hit me in the face 🤜😘 . The idea that was my favourite and fueled my passion—selling frisbees—showed zero results. I was upset and had to let it go.

But then, this data-driven approach highlighted an unexpected idea. It attracted significantly more traffic to the product page and more clicks on the buy button. For me, that was real market validation and a strong signal to move forward with building an MVP…

✨ AI & Tools I used:

Step 2: Feed.
Build a Minimum Viable Product (MVP)

Alright, now that one idea is validated, it’s time to build an MVP—a simple business setup to start selling.

On this stage I kept the effort minimal since the product could still fail.

My minimum E-commerce business setup included:

  • A website
  • An online payment system
  • Localized product content
  • One supplier
  • One logistics partner
  • A small warehouse (at home)
  • Minimal legal & compliance: local laws research and simple sole proprietorship registration.
  • Basic customer service (AI chat & email)
  • A simple brand concept and communication
  • Minimal online advertising

After two months of building, my rough website started making sales. 💸

This isn’t the main point of the post, but here is the link to my e-shop Essetia.com so you can see what one person can build in eight weeks with AI.

✨ AI & Tools I used in the MVP stage:

  • Zapier for automatic data flow between apps and tools
  • A GPT bot that I trained on local Czech laws and documents to understand legal requirements (later confirmed with a legal specialist—never fully trust the robots).
  • Shopify to build the website
  • Claude AI to adapt the website code
  • Aya AI for content translation and localization
  • NotebookLM to review and negotiate contracts
  • Cursor and Replit AI to write code for logistics and warehouse management apps
  • Shopify Inbox for AI-powered chat
  • ComfyUI for product photo
  • ChatGPT, Perplexity, Gemini to build simple brand concept and communication
  • ElevenLabs and CapCut AI for creating marketing videos

Step 3: Grow.
Go All-in and invest.

Right now, I’m in the MVP stage. But if the website data, unit economics, and customer feedback look good, I’ll go all in to capture the market. The risks of investments will be much lower thanks to the knowledge, data, and experience gained in the first two stages.

Conclusions

  1. Business launch speed has increased exponentially compared to previous years.
    In the past, testing an e-commerce idea took six months with a team. This time, I did it alone in two months (including learning time). Next time, I could do it in one month—six times faster and at a much lower cost.
  2. The Seed.Feed.Grow approach proved incredibly useful for solopreneurship.
  3. Have I proven that one person can compete with big companies?
    I’ll put it this way: my belief in the idea has grown significantly. But there’s still so much I need to learn and accomplish on this journey. I will share more.

Next steps

The strategic reason I chose to do everything alone this time is to fully rethink each process before automating it with AI in future. If I don’t understand the details, I can’t explain them to AI agents, and I certainly can’t expect them to perform well for me in the next phase.

And my next step is building a fully AI-native, agent-driven business.


In any case, it’s an exciting and challenging problem for me to solve. I’ll have fun and learn a lot along the way. The new, profitable business model could be a fantastic bonus. 😏

Posted

in

, , , , , , ,

by

strelchyn