Framer, Webflow, HubSpot, or Claude? Building a site in the AI age.

A year ago, building this site would have meant a six-week project, a designer, a developer, and a stack of compromises in between. We built it in days. Not because the tools got cheaper — because the bottleneck moved.

For most of the web's history, the hard part of building a site was the building. Translating a design into working pages took skill, time, and a specialist. So the tools competed on that: how fast can we get you from blank canvas to launched site?

That race is basically over. Framer, Webflow, HubSpot, and a model like Claude can each get you to a polished, responsive, live website. The execution gap between them has narrowed to almost nothing. Which means the interesting question isn't "which one is best" — it's "what are you actually optimizing for?"

Four tools, four trade-offs.

We use all four, depending on the job. They're not really competitors so much as different answers to the question of who owns the result.

Framer is the fastest path to something beautiful. If your priority is design polish and speed, and you're comfortable living inside its ecosystem, nothing beats it for a marketing site. The trade-off is ownership — you're renting the platform, and the day your needs outgrow it, you feel the walls.

Webflow sits one layer deeper. More control, a real CMS, closer to the underlying HTML and CSS — at the cost of a steeper learning curve. It's the choice when a site needs to scale into something content-heavy and structured, and you want a team to be able to manage it without touching code.

HubSpot is the most powerful of the four — and the one that asks the most of you up front. Because the site lives inside the same platform as your CRM, marketing, and automation, a page isn't just a page: it can personalize, capture, and trigger workflows the moment someone lands on it. That reach comes with a learning curve, and it's more than a simple brochure site needs if the content rarely changes. But when your site is genuinely content-driven — a real publishing and marketing operation with forms, campaigns, and a team behind it — nothing else connects the content to the rest of the business as completely.

Hand-built with an AI model like Claude is the other end entirely. You own every line. There's no platform tax, no export friction, no ceiling — and the thing that used to make this option slow, the actual writing of the code, is now the fast part. The trade-off moves to judgment: the model will build whatever you ask, including the wrong thing, confidently.

"The tools all clear the same bar now. What separates the results is taste, judgment, and knowing what you're building — none of which come in the box."

That last point is the one most people miss. When the cost of producing something drops to near zero, the value of deciding what to produce goes up. AI didn't remove the need for craft. It moved it earlier — out of the execution and into the brief.

What we actually chose, and why.

We hand-built this site. Not because it's the right answer for everyone — for a lot of our clients, Webflow or HubSpot's CMS is exactly right — but because of what we were optimizing for: total control of the experience, no platform we'd have to migrate off later, and a codebase we could evolve as fast as the business does.

The decision came down to three questions we'd ask any client weighing the same choice:

1. Who needs to edit it? If non-technical people update it weekly, you want a CMS and a visual editor. If it changes rarely and deliberately, owning the code is freeing, not limiting.
2. How long do you intend to live with it? Platforms are a fast start and a slow exit. The more permanent the asset, the more ownership matters.
3. What's the real constraint — time, money, or control? Be honest about which one actually binds. Most teams optimize for the one that feels urgent, not the one that matters in two years.

The skill that still matters.

Here's the part that doesn't show up in any tool comparison: a model will happily build you a beautiful site that's wrong. Wrong structure, wrong hierarchy, wrong thing emphasized. It has no opinion about your business. You do — or you'd better.

So the work shifted, but it didn't disappear. Less time typing, far more time deciding. Knowing what to leave out. Recognizing when "technically done" isn't "actually good." That's the same judgment we've spent twenty-five years building in golf technology, and it's the part AI can't hand you.

Closing thought.

Pick the tool that fits the job, not the one winning the argument online. Framer if you want fast and beautiful. Webflow if a team needs to own it without code. HubSpot if the site is a content-and-marketing engine and you're living inside its platform anyway. Hand-built if control and longevity are worth the responsibility. But spend your real energy upstream — on the thinking — because in the AI age that's the only part still scarce.

Gareth Londt — Founder & CEO