How I Built This Website with AI

I Built This Website With AI. Here's What Actually Mattered.

Everyone builds websites with AI now. In 2026, saying you built a site with AI is like saying you typed it on a keyboard. New sites ship with AI-generated code by default, and AI website builders have grown into a huge market. The tools are everywhere, and most of them are cheap or free.

Most AI-built sites still look identical. They still fail to convert. They sit there collecting dust instead of leads. The tool was never the bottleneck. Your thinking was.

I built this site with AI after more than a decade building websites and scaling businesses online. I run product and growth for ComparePower, which pulls over 100,000 organic visits a month, and I founded AlbaniaVisit. The reason this site works has nothing to do with which model I picked. It has everything to do with how I aimed it.

Should you treat your website like a product?

Treat your website like a product, not a résumé. A product has a job: pull in the right people and move them to act. Most personal sites are static bios that exist without any job to do. Product thinking gives the site a measurable purpose and a real reason to keep improving.

Product managers never build anything without three answers first. Who is this for? What problem does it solve? How do we measure success? I answered those before I opened a code editor or a single AI tool.

That shift changes the whole thing. Your site stops being a page people land on and starts being a system that runs while you sleep. It filters good opportunities in and wrong ones out. And it earns trust before anyone fills out a form.

Who is your website actually for?

Define your audience before you pick fonts, colors, or frameworks. For this site I mapped three groups: business owners who want real growth, people who need SEO and content that produces measurable results, and collaborators sizing up how I think. Each group needs different questions answered in the first ten seconds.

For every group, I wrote down what they need to see immediately. What would make them reach out? What would make them bounce? Those answers shaped each page, each section, each line of copy. Skip this and you'll build a beautiful site that does nothing.

How do you position yourself on a personal website?

Position yourself with one sentence, not fifteen skills. My statement: I help small businesses get seen online without wasting money on tactics that don't work. That single line decided the headlines, the case studies I feature, the services I highlight, and the ones I leave off entirely.

Most people list every skill they own and hope one lands. That's a buffet, not positioning. Clarity wins. When someone hits your site, they should know what you do and who you do it for in seconds. Make them work for it and they leave.

How do you build a website with AI that converts?

Use AI for execution and keep the strategy yourself. AI drafts layouts, scaffolds components, and writes first-pass copy fast. It doesn't know your audience, your positioning, or which case study deserves top billing. You supply the direction. The model supplies the speed. That division is why an AI-built site converts instead of collecting dust.

Here's the concrete version. I built this site on Astro and React, with Claude and Claude Code doing most of the typing and Cursor handling quick in-editor edits. Claude Code read the whole codebase, scaffolded components, and refactored my layout in a single pass. What used to take me a weekend of hand-coding took an afternoon. The design held up. The copy didn't, so I rewrote it myself.

Here's how I stayed in control of the process:

• I gave strategic prompts. Not "build me a website" but "create a clean, confident layout with purposeful white space that leads visitors toward the contact section."
• I rejected anything off-positioning. AI will generate generic marketing copy all day. My job was to say no.
• I tested every output against one question. Would this convince me if I were the visitor?

AI now writes a large share of new code. That sounds impressive until you notice that code without direction is just fast mediocrity. The builders who win don't generate the most code. They know what to build and why.

Should you design a website for behavior or looks?

Design for behavior. A product evolves from real user data, while a portfolio sits unchanged for years. I set up analytics to track which pages hold attention, which case studies drive inquiries, and where people scroll past. The numbers tell you what to fix. Good looks alone never generate a single lead.

When visitors lingered on one case study, I expanded that section and built more content around that problem. When contact submissions dropped, I simplified the form and sharpened the call to action above it. That loop is what separates a site that works from one that just exists.

AI tools can now watch live interactions and suggest design tweaks. But the call to act on that data, to rank one change over another, stays with you.

How do you make your website findable by AI assistants?

Structure your site for AI assistants, not only Google. People now ask ChatGPT, Claude, and Gemini for recommendations like "who's good at SEO in Dallas?" Sites that surface in those answers use clean schema markup, specific service pages, and consistent language. Ignore this channel and you're invisible in a fast-growing one.

What I did:

• Added proper schema markup so AI systems understand what I do, who I serve, and what I solve
• Wrote FAQ content that maps to how assistants retrieve and cite information
• Built a specific page for each service instead of one vague "Services" page
• Kept language consistent across the site to reinforce positioning in both search results and LLM answers

More people lean on AI to research and shop every month. The sites that show up in those recommendations aren't the flashiest. They're the clearest and the most authoritative.

What did building with AI actually produce?

Product thinking turned this site from a passive credential list into a lead generator. It ranks for keywords that bring qualified prospects, filters out mismatched work by stating plainly who I serve, and produces steady inquiries without paid ads. Visitors get real value before they ever reach out.

That last part matters most. Trust comes from useful content, not claims. By the time someone contacts me, they already know how I think, so the conversation starts higher up. Want to see if we're a fit? Tell me what you're building.

Does AI make you a better website builder?

AI makes fast builders faster and sloppy builders sloppier. It genuinely speeds up the work: developers report real productivity gains, and plenty of designers lean on it for ideation. But speed without direction just ships problems to production quicker. AI won't fix weak positioning or a missing strategy. It executes the plan you give it.

Plenty of teams have been burned by AI too, usually by inaccuracy. That's the real risk. The builders winning right now pair clear thinking with fast execution. They know their audience, positioning, and goals before they ever open Cursor or Claude Code.

What questions should you answer before building with AI?

Answer four questions before you touch an AI tool. Who are your three most important visitor types, and what must they learn in the first ten seconds? What's your one-sentence positioning? What single action should visitors take? How will you measure whether the site is working? Get those right, then build.

Sort out those answers first. Then use whatever AI tools you want. Use all of them. They're genuinely good and getting better every month. But they accelerate a plan. They don't replace one.

Your website is a product. Build it like one, and it'll keep working long after you close the laptop.