⚡ How I Built and Deployed a Client’s Website in 90 Minutes — with AI as My Pair Programmer
TL;DR:
- 🧑💼 Friend’s uncle needed a quick landing page for ABC Logistics
- 🧠 Used ChatGPT for planning + Claude Sonnet (in Copilot) for rapid dev
- 🛠️ Built the page, added branding & feedback-driven tweaks
- 🌐 Deployed it for free using GitHub Pages + GoDaddy DNS
- ⚡ Done in ~90 minutes — AI-assisted dev is a cheat code!
The Setup:
A friend asked me to build a website for his uncle’s new business.
All I had was a PowerPoint, a domain name, and an open evening.
So I teamed up with some smart AI agents — and we got it done in one sitting.
📨 A Tiny Ask That Snowballed
It started like most side gigs do — a casual message from a friend:
“Hey, my uncle’s launching a company. He needs a simple landing page. Can you help?”
There wasn’t much to go on. Just a .pptx with some basic info, a logo file, and a domain name they’d already bought. No developer on their team. No fancy design specs. Just a blank slate and a simple ask:
“It should look professional… and we don’t want to spend on hosting.”
Perfect.
🧠 Planning with ChatGPT: The Invisible PM
Before touching any code, I fired up ChatGPT and shared everything I had — the context, constraints, and files.
It became my project manager in seconds.
We mapped out a high-level plan:
- Extract structured content from the PowerPoint
- Design a lightweight, single-page layout
- Incorporate branding from the logo
- Host it somewhere free
- Connect it to the client’s domain
What would’ve taken 30 minutes of thinking or scribbling got distilled into a clear action plan in 30 seconds.
🧑💻 My AI Pair Programmer Goes to Work
For implementation, I leaned on GitHub Copilot’s Chat (Agent Mode) — which, under the hood, was using Claude Sonnet 4. That model is surprisingly good at front-end work.
With a single prompt:
I want to build a simple, responsive landing page for a logistics company called ABC LOGISTICS. The page should be clean, professional, and mobile-friendly. Use only HTML, CSS, and optionally TailwindCSS or a lightweight framework—no backend logic is needed.
The structure should include:
- A header with the company name and tagline
- An "About Us" section
- A Services section (split into Logistics and Employee Transport Services)
- A Sustainability or Green Initiatives section
- A Contact section (with address, phone, and email)
- An optional Call-to-Action at the bottom
- A footer with minimal legal/disclaimer text
Here's the {{Extracted content}} that should be used to fill in the sections.
💡 The result should be a single HTML file (or include an external CSS file if needed) that I can easily host. Optimize for clarity and visual balance.
Boom — HTML and CSS, styled and ready to go.
I didn’t write a single line of layout code from scratch. Just reviewed, tweaked a few labels, and it looked clean on the first try.
Adding the logo? Copilot handled that too.
Aligning elements, changing backgrounds, cleaning up sections? All handled via short follow-up prompts.
The only thing I had to do manually was clean the background of the .jpg logo using an online tool. Everything else — structure, styling, polish — came from prompting.
🔄 The Feedback Loop
As with any real-world project, a few edits came in later.
My friend shared a couple of requests:
- Remove some extra company details
- Make the layout more minimal
- Adjust mobile view for cleaner contact info
I just passed these as prompts to Copilot.
It iterated like a good junior dev — fast, consistent, and tireless.
🌐 From Prompt to Production (Without Spending a Rupee)
Once the page looked good, I set up a GitHub repository and used the platform’s free static site hosting. For connecting the domain, I had to do a bit of DNS wizardry — just following chatgpt's lead.
Within 90 minutes of starting, the website was up and running on the client’s domain.
No infrastructure. No deployment scripts.
Just AI-stitched web magic — no scripts, no stress.
📌 Why This Felt Different
This wasn’t just a “build a page” story.
It was a real taste of how development is evolving:
- ChatGPT acted as strategist, spec-writer, and reviewer
- Copilot (Claude Sonnet under the hood) handled 90% of the implementation
- I became the orchestrator, not the factory
For once, I wasn’t chasing edge cases or fiddling with flexbox. I was just telling the tools what I wanted — and they built it.
🧭 The Bigger Picture
We often talk about “AI replacing developers.” But the real power is in augmenting us.
This wasn’t about speed — though finishing it in 90 minutes felt good.
It was about how seamless and fun it felt:
- No tutorial rabbit holes
- No blank-canvas paralysis
- Just clear ideas, fast execution, and space for creativity
If anything, it left me more excited for the future of development.
💬 Curious About the Stack?
If you're wondering what helped me pull this off:
- ChatGPT (for planning, extraction, strategy)
- GitHub Copilot w/ Claude Sonnet 4 (for generation + feedback loops)
- GitHub Pages (for zero-cost hosting)
- A friend’s trust (for creative freedom)
🧩 Final Thought
We’re entering a time where a solo dev with AI tools can ship professional results — fast.
This wasn’t a hackathon or startup sprint.
Just a quiet evening, a small ask, and the right tools.
And sometimes, that’s all it takes.
🔮 What’s Next?
This isn’t the end of my AI-assisted dev adventures. I’ve even used similar workflows during a company hackathon (more on that soon 👀).
Top comments (8)
Fun, right? I had the exact same setup for one of my projects and it worked out better than I gave it credit for in the beginning.
You're absolutely right about it not taking jobs, but it will shift them. Planning. Design. Architecture. Systems. These are going to be the new baseline. It won't matter if you can write Java or Python if you can't design and build the system for it, too.
Yeah, I think we're going to see a big shift toward system design and business logic being the real differentiators. The tools are getting better and better at writing code.
Incredible speed! AI really is reshaping how fast we can go from idea to deployment. I had a similar experience recently while reviewing sites like [inatboxsapkindir.com.tr](inatboxsapkindir.com.tr — clean structure, lightweight design, and quick delivery reminded me how much good planning (and smart tools) can accelerate the process.
That site you mentioned is a great example, clean and efficient
Love this breakdown, that's how modern web dev feels with the right AI tools! Have you tried pushing this workflow to even more complex apps?
Not yet on really complex apps, but I’ve tried building a shopping cart and a ToDo list with some specific requirements—part of a recent company hackathon challenge. I’m planning to write an article about it this Sunday, would love for you to check it out
Interesting, great to see AI leveraged for efficient building!
This is extremely impressive, especially getting everything live in 90 minutes with AI handling most of the heavy lifting