Technology & Management
Behind the Build: How I Used AI to Develop My Personal Website
When I decided to rebuild my personal website as a single home for my advisory work, books, articles, and knowledge output, the conventional route was available. Yet my long-standing interest in generative AI, together with my academic and professional background in Management Information Systems, led me to a different experiment: to make the website itself a practical example of the purposeful and responsible use of these tools that I advocate.
I am writing this article to document the experience as it happened: what the tools contributed, what remained my responsibility, and what I learned from the mistakes along the way.
Compass and Helm: AI Accelerates, Humans Lead
Many assume that building software with AI is automatic, reduced to “the push of a button.” This experience showed me the opposite:
- AI offers remarkable speed, suggests alternatives, and produces solid first drafts.
- But it does not set the vision, define priorities, or carry responsibility for values, privacy, and security.
- Without someone who knows what they want, how systems are built, and when to reject an unsuitable solution, the result is a fragile product lacking coherence and character.
Stages of the Build: How the Journey Unfolded
I divided the work into four stages—sequential, yet overlapping in practice:
- First: Analysis and requirements engineering. I defined precisely what the website needed: Arabic and English, a right-to-left experience designed natively rather than translated, a poster library with organized image upload, display, and management, a trustworthy testimonials system, and an admin panel for content, testimonials, and posters.
- Second: Direction and prototyping. Working with the tools was never random. It relied on clearly scoped instructions that named the technologies, required accessibility and permission standards, and stated what must not change at each step.
- Third: Iterative testing and performance tuning. Each change went through a recurring series of automated checks and manual tests: loading speed, screen-size compatibility, accessibility, and the correctness of links and text direction.
- Fourth: Human oversight and final review. Key components and changes were reviewed and tested, and I personally reviewed the texts and visual identity before approving them.
Division of Roles: What the Tools Did, and What Remained Mine
Throughout this journey, I used modern AI-assisted development tools and methodologies, including established Vibe Coding environments accessed through their official and trusted sources. These tools supported requirements analysis, the generation of parts of the codebase, the exploration of alternatives, and faster testing, while clear controls were applied to security, privacy, and access management. Scope definition, decision approval, content review, and final-output evaluation remained under my direct supervision.
For professional accuracy, Vibe Coding should not be portrayed as a “magic lamp” that builds a complete digital product at the press of a button. It is an AI-assisted development approach that still requires sound requirements engineering, systems thinking, change management, code review, functional and security testing, and user-experience evaluation. It is conceptually distinct from Infrastructure as Code (IaC), which focuses on provisioning and managing infrastructure through traceable configuration files. A single platform may benefit from both approaches, but the two concepts are not interchangeable.
What Did I Learn from the Setbacks?
The journey was not free of stumbles, and I consider mentioning them part of professional honesty:
- Some early outputs were correct in form but unsuitable in function, and needed to be redirected.
- Small edits sometimes caused side effects in other parts of the website.
- Links, text direction, permissions, and the mobile experience required repeated checks—not a single pass.
- A successful build alone is not enough to judge the quality of the user experience; it has to be tried as a visitor would see it.
Red Lines: Privacy and Security
In family and educational counseling, privacy is a red line, so clear rules were set:
- Separating consultations: Direct consultations are not handled automatically on the website; they are directed to their accredited channels.
- Testimonials: Human review before approval and publication, with contact details never shown to visitors.
- Responsibility: AI is a tool for enabling and building; professional and advisory decisions remain a human responsibility.
Three Recommendations for Anyone Using AI
- Understand the fundamentals before delegating to the tool: The deeper your knowledge of systems engineering and project management, the better you can steer the tools toward robust solutions.
- Don’t accept first outputs as they are: The tool gives you a good draft; the real skill lies in critique, refinement, and continuous improvement.
- Protect your project’s identity and message: Don’t let the tools impose a generic template; shape them to serve your own voice and mission.
Conclusion
The development of this website has extended over more than four continuous months, beginning with planning and analysis and progressing through design, development, testing, and review to reach its current form. I thank God Almighty for what has been achieved, and I ask Him to make this work a source of benefit, growth, and positive value for people everywhere.
The journey, however, is far from over. We are still at its earliest threshold, and the future holds broad opportunities for learning, development, and shared success. I believe that every beneficial success in the world adds value to us all, and that every worthy undertaking deserves the care and effort required to present it in a form that honors its purpose.
I welcome communication, thoughtful feedback, ideas, and opportunities for continuous improvement. We share one world—and it has room for everyone.
This article documents an ongoing practical experience and will be updated as the website’s structure and working methodology evolve.