You are looking at a website built almost entirely with AI. Not by a developer. By me — a marketer who can write CSS if absolutely necessary but would rather not. Here is how I did it, what I learned, and why I believe this changes everything for creative professionals.
The Vision
I wanted a personal website that felt like me — not a generic template, not an overdesigned portfolio, but something with warmth, personality, and professional polish. Something that could house my marketing work, my travel writing, my devotions, and my photography. A digital home that grows with me.
In the past, this would have required hiring a developer or spending weeks wrestling with WordPress themes. With AI, I built the initial version in days.
The Tools I Used
v0 by Vercel
v0 was my primary building tool. It is an AI-powered interface that generates React components from natural language descriptions. I would describe what I wanted — "a hero section with a full-bleed photo, text overlay on the left, and my signature color palette" — and v0 would generate working code.
What makes v0 special is that it does not just create isolated components; it understands context. It maintains consistent styling, uses proper accessibility patterns, and integrates with the existing codebase.
Claude for Content and Strategy
While v0 handled the interface, I used Claude for longer content pieces and strategic planning. Claude helped me structure my site architecture, draft initial copy for pages, and think through the user journey.
The Tech Stack
The site runs on:
- Next.js — React framework for fast, SEO-friendly pages
- Tailwind CSS — Utility-first styling that AI tools understand well
- Vercel — Hosting with automatic deployments
- TypeScript — Type safety that catches errors early
I did not choose this stack; v0 did. And it is a good choice — these technologies are well-documented, widely used, and AI tools are trained extensively on them.
The Process
Phase 1: Design Direction
I started by describing my vision in plain language: warm but professional, editorial in feel, with a color palette inspired by hyacinths (my namesake flower) — soft pinks, teals, and warm yellows against a deep navy.
v0 generated initial concepts that I could iterate on. "Make the hero bigger." "Add more white space." "The pink is too bright — can we make it more muted?" Each instruction refined the design.
Phase 2: Page by Page
I built each page sequentially: homepage, about, work, journal, travel, devotions, contact. For each, I would describe the purpose, show examples of layouts I liked, and specify any particular content needs.
The AI understood context. By the time I got to the fifth page, it had internalized my design language and was making consistent choices without being explicitly told.
Phase 3: Content Integration
This is where my actual expertise came in. AI generated placeholder content, but I replaced it with my real writing, my actual photos, my genuine stories. The soul of the site is human; the construction is AI-assisted.
Phase 4: Refinement
No AI output is perfect on the first try. I spent significant time on refinement — adjusting spacing, fixing mobile responsiveness issues, ensuring accessibility, and polishing interactions. This phase was iterative: describe the problem, receive a fix, test, repeat.
What I Learned
1. Prompting is a Skill
The quality of AI output depends heavily on the quality of your input. Vague requests produce vague results. Specific, detailed prompts with examples and constraints produce dramatically better output.
I got better at prompting over time. By the end, I knew exactly how to describe what I wanted in terms the AI would understand.
2. Domain Knowledge Still Matters
Even though I was not writing code manually, my understanding of web design principles, user experience, and marketing strategy made me a better AI collaborator. I knew what to ask for because I understood what good looks like.
3. The 80/20 Rule Applies
AI got me 80% of the way incredibly fast. The last 20% — the polish, the personality, the perfect details — took disproportionately more effort. But that is still a massive acceleration compared to building from scratch.
4. Iteration is Everything
I did not get what I wanted on the first try. Or the fifth. Good results came from continuous iteration — generating, evaluating, refining, regenerating. Patience with the process pays off.
Implications for Marketers
This experience convinced me that every marketer should be experimenting with AI development tools. Not to become developers, but to expand what is possible without developer dependencies.
- Need a landing page for a campaign? Build it yourself in hours.
- Want to test a new microsite concept? Prototype it before asking for dev resources.
- Have ideas for improving your site? Try implementing them to prove the concept.
The traditional bottleneck of "I have the idea but need a developer to build it" is dissolving. Marketers who embrace these tools will move faster and have more control over their digital presence.
The Future
This site will continue to evolve. I will add features, write new content, and refine the design. And AI will help with all of it. Not as a replacement for my judgment and creativity, but as a tool that amplifies what I can accomplish.
If you are reading this and thinking "I could never build a website" — you are wrong. The barriers are lower than they have ever been. All you need is a vision, patience, and willingness to collaborate with AI.
The website you are looking at is proof.