August 28, 2026 | Andrew Lawlor

Last week we launched our new website. It reflects where the work has moved: Artificial Intelligence solutions for the enterprise.

What I want to talk about here isn’t the site itself. It’s how we built it, because the process turned out to be a better proof of our thesis than any page of marketing copy.

We used the tools we sell

AI did the strategy, the design, and the build. Claude played marketing consultant on go-to-market. Claude Design played web design specialist. Claude Code played WordPress developer. A human sat in the loop at every step, checking alignment and approving what shipped.

That last part isn’t a disclaimer. It’s the operating model. The AI moved fast, and we decided what was right.

Strategy first, because the AI already knew us

We iterated extensively with Claude to codify our approach and formalize the AI solutions we’re bringing to market. Some we have been delivering to customers for the last two years. Others are new, built in anticipation of strong demand.

Here’s the part people underestimate: Claude had been a strategic partner for months before this project, so it knew about our pivot to offering AI. That accumulated history is what streamlined the process. Starting cold with a blank prompt would have been a slower exercise.

Evolve the theme, don’t torch it

With a go-to-market strategy in hand, we turned to Claude Design to brainstorm themes. We launched a design language a couple of years ago that still works, and the goal was to update it for the new offerings rather than replace it.

We cycled through a number of UI directions quickly and built marketing copy for each solution, new and existing, as we went. Design and messaging developed together instead of in sequence. That alone removes a week of back and forth.

Then Claude Code turned static files into a real site

Static HTML and CSS is a mockup, not a website. Claude Code did the WordPress work.

One practical note on the setup, since people ask: Claude and Claude Design are browser-based applications, while Claude Code launches through the command line. It’s also available now as a plug-in for VSCode, like Cursor. We kept Claude and Claude Design open in their own Chrome tabs, side by side, with Code running in the terminal.

A new domain for what’s coming

We also added aptaria.ai as a new domain. Over the coming months it will carry landing pages related to our AI services. Consider it a placeholder with intent.

The unglamorous part: DNS

Our WordPress hosting and administration partner deserves real credit. We switched domain name reservation vendors and updated DNS entries along the way. That’s the kind of work AI doesn’t do for you, and a bad handoff there costs you a day of downtime. Experienced hands on the infrastructure side kept a fast project from tripping over plumbing.

Weeks, not months

I’ll be blunt about the productivity gain: it’s hard to overstate. Previous refresh cycles took months. This expansion and refresh took a couple of weeks, from developing the go-to-market strategy to being live.

The compression came from removing handoffs. No waiting on a copy draft to review a layout. No waiting on a design file to start development. When the same context flows through strategy, design, and code, the calendar stops being the constraint.

The site is not finished, on purpose

The AI economy is moving faster than any annual web refresh cycle can track. We plan to update often, with new content and new solutions as they mature. A site that takes two weeks to refresh can afford to change its mind.

Tell us what you think

Go look at the new aptaria.com and give us your honest reaction: what’s clear, what’s not, and what you’d want to see next.

If your own site, or your internal tooling, is stuck in a months-long cycle, the constraint probably isn’t your team’s capability. It’s the number of handoffs between the people who decide, the people who design, and the people who build. Collapse those, keep a human accountable at each gate, and see what your calendar looks like.