The Site Builder Is Dying. What Comes Next?

"You can build a website without writing code."

For a long time, that was the site builder's big promise.

But site builders that keep selling "not writing code" as the value itself will, I believe, eventually lose their reason to exist.

The reason is simple.

It's becoming easier to have AI write the code than to learn a dedicated tool and assemble a site in it.

There was a time when building a website meant writing code.

So NoCode site builders offered a GUI you could operate instead. Users could build sites by dragging elements and adjusting properties, without understanding HTML or CSS.

That was a big step forward.

But AI is changing what "not writing code" even means.

The NoCode of the future is not operating a complex GUI instead of writing code.

It's telling AI your intent, and having the code and software you need generated for you.

The premise of the NoCode site builder is starting to crumble

Traditional NoCode site builders have offered a GUI as an alternative to writing code.

But the better AI gets at generating code, the weaker that premise becomes.

To use a new tool, you have to learn its proprietary layout system, its panels, its components, its responsive settings, its concept of a CMS.

It's becoming easier to simply describe what you want to AI and have it generated on the spot.

On top of that, traditional site builders carry a major constraint.

In most cases, the only thing you can make with them is a website.

Meanwhile, AI-powered development environments can now build not just landing pages but internal tools, applications, databases, APIs, and business workflows.

In a world where AI can build any kind of software, the cost of learning a complex tool that only builds sites keeps rising in relative terms.

The closer we get to an era where you can truly build software without writing code, the less "build a website without code" is worth on its own.

Traditional SaaS and AI-native products

This shift is not just a site-builder problem.

All of traditional SaaS is facing the same question.

Until now, SaaS has provided interfaces for making specific work more efficient.

Design tools have their own way of working for design, spreadsheets for spreadsheets, web production for web production.

Users have learned a different piece of software for each purpose.

But AI is about to change this structure.

Instead of learning each operation one by one, users state a goal and AI executes across multiple kinds of work.

"Build this page."

"Research the competitors."

"Come up with three ways to pitch this."

"Match the design to the brand."

"Watch the numbers after launch and improve it."

These tasks are starting to become one continuous experience, rather than separate tools and separate screens.

Whether a product can adapt to this change will be the dividing line between existing SaaS and the new AI-native products.

Traditional products are adding AI features too.

Meanwhile, the newly emerging AI-native services are designed around AI from the start.

The difference between the two is not simply whether AI features exist.

It's whether the center of the product is manual operation by the user, or collaboration with AI.

When the underlying model evolves, an AI-native product's own capability leaps forward with it.

Things you couldn't build before suddenly become buildable with a model update.

Work that used to be complex gets done in a few exchanges.

This speed of evolution is nothing like traditional software development.

So for existing site builders, the real competition isn't just other site builders.

Every product that can build software with AI becomes a competitor.

From tools that connect people to code, to tools that connect people to AI

The role traditional site builders have played is connecting humans to code.

They let people who couldn't write code build for the web through a visual interface.

What's needed next is an interface that connects humans to AI.

AI itself will eventually become common infrastructure available to almost everyone.

Access to models alone won't be much of a long-term differentiator.

What will matter is what you hand to the AI, how you convey intent, and how you understand what it generates — how you compare it, revise it, and carry it into operation.

Natural language alone won't solve everything.

To direct design with words, users themselves have to be able to verbalize what they want to express.

Tell it "make it look better," and "better" means something different to everyone.

Layout, spacing, information hierarchy, brand feel, the choice of photos — many things are faster to judge visually.

So the GUI doesn't become unnecessary.

The GUI's role changes.

Until now, the GUI existed so humans could assemble the finished thing piece by piece.

From here, the GUI exists to convey intent and context to AI — to understand the generated options, judge them, and correct course.

It's a shift from an interface for humans to do the work, to an interface for humans and AI to work together.

Building for the web stops being about making pages

There's one more premise about site builders that needs revisiting.

Is web production really work that ends when a page goes live?

In reality, plenty of work comes after the site is built.

Researching the market and customers.

Thinking through the messaging.

Writing content.

Preparing images.

Keeping the brand consistent.

Making sure search engines — and AI — can discover the information.

Checking the response after launch.

Updating as things change.

Improving what isn't working.

Until now, this work has been fragmented across multiple tools and multiple people.

The site builder makes the page, a document tool holds the copy, an analytics tool shows the numbers, a project management tool carries the update requests.

AI has the potential to connect these fragments.

AI researches the market and competitors, and drafts content from what it finds.

It lays out multiple pitches and designs, and a human compares and chooses.

The chosen direction gets implemented as a site.

After launch, it proposes updates and improvements as it watches the response and the business evolve.

In other words, the site builder of the future may no longer be a tool for building sites.

It becomes an environment for running a business on the web — where research, planning, production, publishing, operation, and improvement flow as one.

What the canvas makes possible

When thinking about human-AI collaboration, canvas-style interfaces are interesting.

In chat, information flows in time order.

One answer follows another, and earlier output scrolls away off the top of the screen.

That suits conversation, but not necessarily comparing multiple ideas or taking in a complex context at once.

On a canvas, you can lay out generated content, research, and design directions in space.

Compare several options.

Keep related information close together.

Remove what you don't need.

Pick the part that interests you and have AI develop it further from there.

You keep the context in a form a human can grasp at a glance, while continuing to work with AI.

This is a little different from the idea of AI automatically finishing everything.

AI generates many possibilities; a human looks at them and gives direction.

Human judgment and AI's generative capability, combined in a visual space.

Interfaces like this may have an important role in the next generation of web production environments.

The website itself is changing too

What's changing isn't just how websites get made.

The role of the website itself is starting to change.

Until now, the assumption was that humans arrive from a search engine, read the pages, and take in the information.

Increasingly, AI will do that on our behalf — finding, organizing, and answering with information from the web.

More people will come to know products and services through conversations with AI, without ever visiting the company's website.

When that happens, a website becomes not only pages for humans to look at, but an information foundation for AI to understand correctly.

Making well-designed pages won't be enough.

The information has to be structured.

Kept up to date.

The meaning of the brand and the product has to come through clearly.

Reusable across multiple channels.

Easy for both humans and AI to understand.

What the site builder deals with will expand — from the website as a visual artifact to the business's information itself.

Defining web production for a new era

What the AI-era site builder needs is not to make the traditional building experience slightly more efficient.

It's to redefine the entire process of web production with AI as the premise.

Not hiding code, but creating the right points of contact between the user and AI.

Not finishing pages, but creating a loop where research, production, operation, and improvement circulate.

Not leaving everything to AI, but creating an environment where human intent and judgment carry further than they ever have.

The site builders we've had opened up web production to people without the ability to write code.

The next site builder has to open up the power of creation and operation to people who hold the power of AI but can't yet wield it well.

From the era of connecting people to code, to the era of connecting people to AI.

For the site builder to survive, it has to stop being a site builder.