SaaS Is Dead. The Workflow Is the Product.

We built a website workflow that made us ask an uncomfortable question: why were we paying for a website builder in the first place?

Using Codex, we can change the copy, adjust a layout, add a page, preview the result, check what changed, and keep a version we can return to. We have effectively rebuilt the part of Wix we needed for ourselves. The same workflow lets us keep improving all our websites as the business changes, instead of treating each redesign as a new project.

That experience changed how we think about SaaS.

The old bargain was compelling. A vendor built software once and made it available to thousands of businesses. We paid a subscription and avoided building and maintaining it ourselves. For common needs, that remains a good deal.

But the bargain gets weaker when a business has to reshape its work around a generic tool. You pay for the platform, then spend time working around its limits. With a website builder, even the site's future can be tied to the builder. Wix says a Wix site must run on its own infrastructure^[1]. That is a product choice Wix can make. It is also a constraint its customers inherit.

AI changes what we can choose to own. Codex can work in a codebase, make changes and run checks^[2]. For us, that means our websites can be source we control and improve, with an agent helping us do the work. They are AI native because change is part of the everyday workflow, rather than an occasional handoff or a feature confined to a fixed editor.

They are also designed to be vendor agnostic. We use Codex today, but our website source does not have to live inside Codex. We can change models, collaborators, and hosting as our needs change. Moving a live site can still take work, especially where outside services are involved. Owning the source gives us more choices when that day comes.

We do not think every company should build its own payroll system. We do think the default answer to "we need software for this" should stop being "which subscription do we buy?"

Start with the work. What is distinctive about it? What changes often? Where does the team keep compensating for a tool that almost fits? Then decide what to buy, what to build, and who will maintain the result.

This is the business model we are developing at Thinkforge Labs: strategy and implementation as a productized service.

Strategy means finding the valuable problem and making the choices a tool cannot make for you. Implementation means building the site, workflow, or system, getting it into use, and learning from what happens. Productized means the work has a clear path and a defined outcome. Each engagement still starts with the reality of the business in front of us.

Our website workflow is proof of the idea. We needed the freedom to update our presence continuously, so we built a way to do it. The value was not simply that Codex could write code. It was that we could connect a business decision to a working change without losing control of what we built.

That is what we mean when we say SaaS is dead. Subscriptions will survive. The assumption that every business must rent its way of working from someone else is the part we think is ending.