By Faqroon Editorial Team
Editorial note: this article describes a practical workflow and the checks used during development. AI assisted the process, but decisions and verification remained manual.
How I Used AI to Build Tartivo.com From Scratch: My Real Experience
Building a web app used to mean spending weeks between developers, designers, database specialists and hosting providers.
For Tartivo.com, I decided to take a different approach: use AI as a development partner and build the project step by step myself.
This is my return of experience — from the first idea to the website going online.
From an idea to a real web app
The idea behind Tartivo was simple: create a marketplace where people can sell, buy and rent products and services, inspired by platforms such as classified-ad websites.
But having an idea is easy. Turning it into a working product is the difficult part.
I used AI throughout the process to help me think about:
- Website structure
- User experience and UI/UX
- Features
- Database structure
- Development
- SEO
- Content
- Deployment
- Marketing
The important thing was not asking AI to "build everything" at once. I worked progressively, validating each step before moving to the next.
Designing the website with AI
I started with the design.
I used AI to explore layouts, colors, components and user journeys before turning them into real web pages.
For Tartivo, the goal was to keep the experience simple: users should be able to quickly browse categories, create an advertisement and contact a seller.
AI was particularly useful for generating ideas and improving the interface without needing to create every design from zero.
Choosing the technology stack
Once the design was clearer, I moved to development.
The stack depends on the requirements of the project, but the important lesson I learned is this:
Don't choose technologies just because they are trending. Choose them because they solve your actual problem.
AI helped me compare different approaches, understand technical trade-offs and generate parts of the application.
GitHub Copilot became part of my development workflow
One of the tools I used heavily was GitHub Copilot.
Instead of simply asking:
"Build my website."
I gave Copilot smaller, more precise tasks.
For example:
"Create the registration page with validation and error handling."
Then:
"Add the forgot-password functionality."
Then:
"Review this component and improve its security."
This way, AI became more like a coding assistant than a magic button.
GitHub's current Copilot workflow also goes beyond simple autocomplete. Its Agent mode can work through multi-step tasks, modify files, run commands and iterate on problems. (GitHub Docs)
Database, authentication and backend
A marketplace needs more than a nice interface.
I also had to think about:
- Users
- Advertisements
- Categories
- Images
- Authentication
- Contact information
- Permissions
- Search
- Data relationships
AI helped me understand how the SGBD/database, backend and frontend should communicate.
But I still had to make the final decisions and verify the generated code.
That's one of the biggest lessons from this project:
AI can accelerate development, but you still need to understand what you're deploying.
From localhost to a real website
After development came another important part: deployment.
I had to deal with:
Domain → DNS → VPS → Server → Application → Database → Security → HTTPS
Buying a domain is only the beginning.
You also need to configure DNS records correctly, prepare the VPS, deploy the application, configure the server and make sure everything works correctly in production.
This part was less "fun" than coding, but it was essential to turn Tartivo from a local project into a real website.
AI also helped with SEO and marketing
The interesting part is that AI wasn't only useful for programming.
I also used it for:
- SEO content
- Article ideas
- Meta titles and descriptions
- Keywords
- Social media posts
- Advertising copy
- Flyer concepts
- Promotional visuals
For example, instead of creating a generic advertisement, I could ask AI to adapt the message to a Tunisian audience, while keeping Tartivo's brand identity.
This allowed me to move much faster from development → content → marketing.
What about hiring a company?
This was another question I had during the project:
Would it be better to build everything myself or pay a company?
My experience showed me that there isn't one answer.
If you have the technical knowledge and enough time, AI can significantly reduce the amount of repetitive development work.
But for a company, outsourcing can still make sense when you need:
- A production-ready architecture
- Advanced cybersecurity
- Professional UI/UX
- Complex integrations
- Infrastructure management
- Long-term maintenance
- A team that can move faster
A practical approach can therefore be hybrid:
Build the MVP yourself with AI → validate the business → outsource specialized or critical work when necessary.
What I learned from building Tartivo
The biggest lesson wasn't about a specific programming language or AI tool.
It was about the way software development is changing.
Today, one person with a good idea can use AI to move through many stages of a project much faster: planning, coding, testing, debugging, deployment and marketing.
AI coding agents are also becoming more capable of handling multi-step development workflows rather than only generating individual lines of code. (GitHub Docs)
But AI doesn't remove the need for developers.
It changes where their time is spent.
Instead of writing every piece of code manually, more time can go into architecture, validation, security, product decisions and quality control.
My Tartivo.com AI workflow
The process can be summarized simply:
Idea → Planning → UI/UX → Development → Database → Testing → VPS → Domain & DNS → Deployment → SEO → Marketing
And AI can assist at almost every stage.
That's what made the Tartivo project interesting for me: I wasn't simply using AI to generate code.
I was using AI as a product, development and marketing assistant to turn an idea into a real web application.
Practical checklist from the project
- Define the product scope and the pages users need before asking for implementation help.
- Review generated code line by line, then run syntax checks before deployment.
- Test the public HTTP responses, redirects, sitemap and mobile layout after each release.
- Keep real credentials outside source files and use a controlled production configuration.
- Document failures and limitations instead of presenting generated output as proof of success.